Don Lancaster's What's New?       2006 Archive Blog

Pick your blog year...
1997	1998	1999	2000	2001
2002	2003	2004	2005	2006
2007	2008	2009	2010	2011
2012	2013	2014	2015	2016
2017	2018	2019	2020	-----

( other blogs appear here. )  

Closed out the 2006 What's New blog and started the
2007 What's New one.

Turns out that PostScript gladly reports integers up to
two billion or so to full accuracy. So the trick to picking
up another two decimal points of floating point reported
precision is to normalize each real to an integer in the
ten to hundred million range, convert it to a string, and
then dink around with the string to replace the decimal
point and sign.

Here's a simple and preliminary example that is only
good for reals between 1.0000000 and 9.9999999...

          /unitsto8dp { dup 0 le /isneg exch store
          abs /val exch store /workstr val 10000000
          mul cvi 20 string cvs store isneg {(-)}{( )}
          ifelse workstr 0 1 getinterval mergestr
          (.) mergestr workstr 1 workstr length 1 sub
          getinterval mergestr 20 string cvs } store

This routine uses my mergestr routine from my
Gonzo Utilities...

        /mergestr {2 copy length exch length add
        string dup dup 4 3 roll 4 index length exch
        putinterval 3 1 roll exch 0 exch putinterval
                             } def

I'll try to work up a GuruGram on a more general
solution.

I feel it is a good rule to always typeset first and edit
last. The exact arrangement and appearance of your
work is equally important as the words themselves.

And yes, spelling counts.

December 30, 2006

Spam seems to be accelerating at an insane pace,
and it is not too difficult to predict an imminent and
total collapse of the web because of it.

At present, we receive something like 2000 spam
messages a day, most of which are trapped by our
ISP filtering. Compared to a few dozen worthwhile
messages.

The alarming thing is that the spam to useful ratio
is dramatically accelerating. Further, the spam feeds
on itself. Since everybody now recognizes the tricks
to bypass the spam filters ( .GIF files, misspellings,
random words, repeated contacts, etc etc...) many more
spam messages have to be sent out to get even
remotely near their previous return.

The solution is simple: No more email arriving with
postage due.

There is apparently now a New England NEEIC
alternative to the California PIER new energy grants.

Overheard some alternate energy enthusiasts who
were lavishly praising Sterling engines as the
ultimate solution to low delta-t energy recovery.

It quickly became obvious that they did not have
the faintest clue of the underlying thermodynamics
or economics.

To date, the Sterling engine has been one of the
largest and most monumental engineering ratholes
of all times. Here is why...

    Carnot Matters -- There is a fundamental and
     unavoidable law of thermodynamics that says
     the best possible efficiency of any heat engine
     is proportional to the absolute temperature
     delta fraction. Thus your best possible efficiency
     a 20 degree rise at 70 degree F room temperature
     would be 20/(459+70) = 3.8 percent. And no
     real world system can be even this good.

    Efficiency Matters --As efficiency goes down,
    the size and complexity of the energy recovery
    device will disproportionately increase in a
    hyperbolic or worse manner for a given set of
    recovery values. Which is why absolutely free
    pv solar panels of less than six percent efficiency
    are totally commercially useless.

    Amortization Matters -- If your energy recovery
    device is producing an average of two cents worth
    of electricity per day and your total cost of ownership
    is three cents per day, you have a gasoline destroying
    net energy sink. The longer you run it, the more
    gasoline you destroy

   Gotchas Matter -- A Sterling engine needs a
    special part called a regenerator. Regenerators
    have to be long and thin and short and fat.
    They also have to be very good conductors of
    heat and outstanding insulators. Extreme
    engineering compromise is needed and nobody
    has come up with a good regeneration solution
    to date.

Much more in our Energy Fundamentals tutorial.

Got some demo code tentatively working for
some ultrafast Magic Sinewave solutions given
a good initial guess. It seems hundreds or even
thousands of times faster than before.

It is based on factoring the equations into a
guess cosine and an error term that ( presently )
is slope estimated. Rearranging the terms
and substituting leaves n linear equations in
n unknowns. This is easily and quickly solvable
by Gauss Jordan elimination.

At present, the solution is "not quite" one step.
A few iterations are still required. But they
do quickly drive the distortion down to all
zeros out to fifteen decimal places, and the
amplitude to a similarly exact value.

It sure is satisfying to see all those zeros.
And nothing but zeros on the forced harmonic
terms. Often with five or fewer passes.

I suspect an exact one step solution exists.
Fully deterministic for a valid initial guess.

Instead of a slope approximation of minus the
sine, a trig identity of cos (a+b) = cos(a)cos(b) -
sin(a)sin(b) might prove of value. But it seems
like there might be a second factor or some
larger interfering error yet to be discovered.

Preliminary code can be made available to
Synergetics Partners and associates at present.

Some useful things the faster code may assist
with: Exploring sinewaves with many hundreds
and possibly thousands of harmonics zeroed.
This would dramatically simplify filtering,
offer wide frequency operation, and raising of
uncontrolled harmonics out of the audio range.
At the cost of reduced efficiency and fancier
chips/programming.

And discovering scalable code for variable n
that does not require extensive reworking for
each and every n to be analyzed.

And finding new classes of Magic Sinewave solutions
beyond the six or so that are presently known.
Something that would give a "suppressed carrier"
would certainly be handy. Possibly with two pulses
of the "wrong" polarity early in the mix. Or of
multi amplitude stepped supply solutions.

A Magic Sinewave intro appears here.

Can the math precision and accuracy of PostScript
be improved? Its out-of-the-box six decimal point
reported answers were not quite good enough for at
least some of my Magic Sinewave research.

A case can be made that more than six decimal
results would only lengthen font files without any
measurable result in most graphic applications.

But 31 bit plus sign math should be good for two
billion states, or over nine decimal points. Minus
where they stopped in any of their transcendental
approximation algorithms. And minus what their
floating point routines allow..

Here's a preliminary experiment: The square root
of two in PS returns 1.41421 . Rounding to four
decimal places and subtracting gives .0000135899.

Combining the two returns 1.4142 1358 99. Compared
against the "real" answer of 1.4142 1356 23. Which
is right on at eight decimal places, and not too bad
at nine.

A simple text reformatter that does the combinations
for you should be possible. And based on our Gonzo
mergestr convenience operator.

Again, you would have to compare exactly what you
wanted against the individual algorithms in use for the
fancier functions.

Adobe does state a precision of "approximately" eight
decimal digits. And apparently are using single precision
IEEE 754 floating point. Because of a "normalization"
scheme in this, they are able to make several bits serve
double uses, resulting in the claimed precision.

Possibly more later.

Do we have yet another bogus water powered car
fiasco coming down? The known ( and repeatedly
remeasured zillions of times daily ) energy density
of STP hydrogen is 2.7 watthours per STP liter
electrically recoverable or 3.3 watthours per liter
total energy.

One company appears to claim to be able to input
1.0 watthours of electricity to produce 3.3 watthours
of stored hydrogen energy. In an electrolysis related
process.

Although a stunning reversal of centuries of electrochem
and thermodynamic research is one dim possibility, I feel
that much more likely explanations are (A) the usual
ho-hum incompetent measurement of oddball waveshapes,
(B) having water vapor or other gas components in
the output stream, or (C) being yet another in a countless
stream of attempted ripoff scams.

Their 2.5x improvement claim is remarkably similar to
the difference you would expect between average and
rms currents in a typical pulse waveform situation.

As we've seen before, "perfect" and "free" electrolysis
is totally useless when powered from high value sources
such as pv solar, wind, or grid. Because of the staggering
loss of exergy. Thermodynamic fundamentals guarantee
that a kilowatt hour of electricity is ridiculously more
valuable than a kilowatt hour of unstored hydrogen gas.
The process is pretty much the same as 1:1 converting
US dollars into Mexican Pesos.

And that is before amortization and maintenance.

Further, if this is an onboard vehicle system, the fanbelt
alone guarantees that no more than homeopathic placebo
quantities of hydrogen can be produced. As we have seen
here in the November 19th entry.

More on energy fundamentals here, on bogus water
powered scams here, and on bashing pseudoscience
here.

Nissan may have just come up with a stunning hybrid
vehicle engineering breakthrough: A motor with two
rotors and one stator. Each rotor can run at different
speeds, and either can act as a generator.

Firstoff, this solves the differential problem in spades.
The ultimate in electronically controlled positraction.

Secondly, generation from an ICE and regenerative
braking from the wheels can be handled by the same unit.
At substantially reduced cost and weight and complexity.

( Please read yesteday's entry first )

Continuing our Gauss-Jordan tutorial, but this time
the Jordan part. When we last left off, we had a
( relabeled ) array of...

             [  1    c01  c02  c03  c04  j05 ]
             [  0     1    c12  c13  c14  j15 ]
             [  0     0      1   c23   c24  j25 ]
             [  0     0      0     1    c34   j35 ]
             [  0     0      0     0      1      z   ]

where cxx is the row and column coefficient for
the left side equation terms, and jxx is the
similar row and column coefficient for the right
side equation term.

The usual way to solve this is by back substitution.
Start off with y = j35 - z*c34 and so on. And then
work your way up a row at a time, making more
complex calculations until you have v through z
all solved.

The Jordan approach starts off the same way, but
it works one column at a time, greatly simplifying
computer programming. Especially if more than
one n x n equation set is to be accommodated.

The rule is that any constant can be subtracted
from one term in the left side of the equation if
the same constant is subtracted from the right
side of the equation.

Subtract z*c34 from row 4...

             [  1   c01 c02 c03 c04 k05 ]
             [  0    1   c12 c13 c14 k15 ]
             [  0     0    1   c23 c24 k25 ]
             [  0    0     0     1     0     y   ]
             [  0     0    0     0     1     z   ]

So far, this is the same as the usual back substitution.
We now can observe y by inspection The difference
with Jordan is to continue with columns instead of
rows. Modify the rows by subtracting z*c24, z*c14,
and z*c04 to get...

             [  1   c01 c02 c03  0  m05 ]
             [  0    1   c12 c13   0  m15 ]
             [  0     0    1   c23   0 m25 ]
             [  0     0    0     1     0    y   ]
             [  0     0    0     0     1    z   ]

Next, modify column three by subtracting
y*c23, y*c13, and y*c03. And then column
two by subtracting x*c12 and x*c02. And
finally column one by subtracting w*c01
to get...

             [  1    0    0    0    0    v  ]
             [  0    1    0    0    0    w ]
             [  0    0    1    0    0    x ]
             [  0    0    0    1    0    y ]
             [  0   0    0    0     1    z  ]

Your values v through z are now instantly
readable by inspection.

Once again, the Jordan method takes just as
many calculations as back substitution, but it
greatly simplifies computation in that loops do
not have any multiple calculations or complicated
cross-coefficients in them.

Finally figured out what the "Jordan" part of Gauss
Jordan elimination is all about. Turns out that while
there are just as many calculations that are just as
complex as plain old back substitution, those calcs
lend themselves to much simpler and more easily
automated computer loops.

Consider five linear equations in five unknowns...

    A0*v + B0*w + C0*x +D0*y + E0*z = K0
    A1*v + B1*w + C1*x +D1*y + E1*z = K1
    A2*v + B2*w + C2*x +D2*y + E2*z = K2
    A3*v + B3*w + C3*x +D3*y + E3*z = K3
    A4*v + B4*w + C4*x +D4*y + E4*z = K4

While all sorts of solution methods exist, we seek
one that is computationally efficient. If we dink
around with some manipulations ahead of time, we
can eventually end up with a solution that will be
obvious by inspection!

Arrange the coefficients into a group of arrays...

             [ A0 B0 C0 D0 E0 K0 ] 
             [ A1 B1 C1 D1 E1 K1 ] 
             [ A2 B2 C2 D2 E2 K2 ]
             [ A3 B3 C3 D3 E3 K3 ] 
             [ A4 B4 C4 D4 E4 K4 ] 
             [ A4 B4 C4 D4 E4 K4 ] 

The rules for our "Gauss" part of rearrangement
are that any row can be scaled by any constant term
by term without changing the results.

And that any row can be subtracted from any other
row term by term and substituted. Again without
changing the results.

In interests of sanity, let "~" be any coefficient
that resulted from previous manipulation. Scale the
top row by dividing by its initial value...

             [  1    ~    ~    ~    ~   ~   ]
             [ A1 B1 C1 D1 E1 K1 ] 
             [ A2 B2 C2 D2 E2 K2 ]
             [ A3 B3 C3 D3 E3 K3 ] 
             [ A4 B4 C4 D4 E4 K4 ] 

Scale the top row by A1 and subtract it from the
next row down and replacing...

             [  1    ~    ~    ~    ~   ~   ]
             [  0    ~    ~    ~    ~   ~   ]
             [ A2 B2 C2 D2 E2 K2 ]
             [ A3 B3 C3 D3 E3 K3 ] 
             [ A4 B4 C4 D4 E4 K4 ] 

Similarly, scale the top row by A2 subtract it from
the middle row. Then scale by A3 for row 3 and A4
for row4...

             [  1    ~    ~    ~    ~   ~   ]
             [  0    ~    ~    ~    ~   ~   ]
             [  0    ~    ~    ~    ~   ~   ]
             [  0    ~    ~    ~    ~   ~   ]
             [  0    ~    ~    ~    ~   ~   ]

Now, scale the second row down by its first nonzero
coefficient...

             [  1    ~    ~    ~    ~   ~   ]
             [  0    1    ~    ~    ~   ~   ]
             [  0    ~    ~    ~    ~   ~   ]
             [  0    ~    ~    ~    ~   ~   ]
             [  0    ~    ~    ~    ~   ~   ]

Next, force zeros in the second column the same we
we did with the first, but using the second row for
subtraction and substitution...

             [  1    ~    ~    ~    ~   ~   ]
             [  0    1    ~    ~    ~   ~   ]
             [  0    0    ~    ~    ~   ~   ]
             [  0    0    ~    ~    ~   ~   ]
             [  0    0    ~    ~    ~   ~   ]

Keep working your way through the array, this time
scaling the third row down by its first nonzero term and
then using scaled subtractions to zero out everything
below in the same column.

Eventually, you should end up with...

             [  1    ~    ~    ~    ~   ~   ]
             [  0    1    ~    ~    ~   ~   ]
             [  0    0    1    ~    ~   ~   ]
             [  0    0    0    1    ~   ~   ]
             [  0    0    0    0    1   ~   ]

This completes the Gauss part of the process.
The lower right squiggle will be z by inspection!

From here, we can use back substitution or the
Jordan scheme. More on Jordan tomorrow.

Computing power has gotten FUNDAMENTALLY
INSANE.

Just realized I was sitting here solving 14 linear
equations in 14 unknowns to 64 bit precision. And
worrying about how I was going to speed up the
algorithm to get under 120 milliseconds. And being
upset that 32-bit math, while useful, was not quite
good enough to do the job at hand.

That, of course, is while limping along on an ancient
( almost two years old! ) 750 MHz machine. Compared
to back in college where I would spend hours with a
K&E log log duplex decitrig slide rule along with the
Mathematical Tables from the Handbook of Chemistry
and Physics to try and solve a simple transmission line
problem. To three percent accuracy.

Just about anybody now has personal computing power
that is unimaginably beyond the best available to only
the biggest schools or corporations a very few years
ago.

Which tells us that these days, if you have a problem,
throw some math at it. Another ten million calculations
is simply not that big a deal anymore. Brute force
reigns supreme.

And no telling where it will lead.

Places where bunches of intense math have proved both
interesting and popular include my Fun with Fields and
my ongoing Magic Sinewave alternate energy research.

Other technological breakthroughs reviewed here.
And more math stuff here and here.

Outside of the thousand cases of White Tiger Energy
Drink, Arizona Auctions seem to be in the middle
of their usual year end lull.

The recent military surplus changes seem to be
pretty much implemented. Per this map, it seems
that mil surplus bargains have completely vanished
in places like Arizona or New Mexico. While very
much increasing in areas like West Florida and
Central Pennsylvania.

Part of their "efficiency improvement" involves
transshipping unsold surplus nearly a thousand miles
from Southern Arizona to Northern Utah.

Your tax dollars at work.

Not sure where this will leave Government Liquidation.
Dollar volume at several of their manned sites clearly
appears to have dropped precipitously.

An interesting story on America's highest value
farm crop. Now exceeding both wheat and corn
combined. Mostly due to the enormously lucrative
DHS and DEA subsidies and price supports. Plus,
of course, absolutely superb 100% tax credits. Can
ethanol production from it be all that far behind?

Do these farmers really need all that federal aid?

Found and verified an even better and much
faster route to Magic Sinewave solutions.

Factor the equations into present angle +
slope*xoffset. Regroup the present angles
to the right side, leaving 14 linear equations
in 14 unknowns. Solve by using Gaussian
Elimination and Back Substitution.

Note that the fundamental error can be included
in a "pseudo distortion" error array. This
eliminates the amplitude variation we got with
the present methods. Note also that the slope
of cos(5x) is -5sin(5x) and similar.

I've got this working beautifully in PostScript,
except it is maddeningly at the PS math 32-bit
limit. I find PostScript infinitely more intuitive
and friendly than JavaScript when it comes to
exploring math options.

At any rate, the results seem incredibly fast
converging, likely needing only a single pass.
At speeds hundreds or thousands of times
faster than before.

Further work probably depends upon your
funded support.

So, what are the other "secret stuff things
to do" in the Gila Valley?

Sadly, our two best wild hot springs are no
more. But here's my own selection of good
stuff that is still here...

   ~ MGIO observatory tours
   ~ Discovery park simulator and telescope
   ~ Mount Graham aerial tramway
   ~ Gila Box Riparian Area
   ~ Hot well dunes recreation area
   ~ Eden hotel & hot spring ( restricted )
   ~ Arivaipa Canyon
   ~ The legendary 4WD "Rug Road"
   ~ Bonita Creek
   ~ Frey Mesa Falls
   ~ Morenci Mine tour
   ~ Eurofresh tours
   ~ Black Hills Rockhounding
   ~ Bear Springs ( restricted )  
   ~ Mazuma mine scam ( restricted )
   ~ Ash Creek Flumes
   ~ Kennedy Peak
   ~ Bear Canyon
   ~ Needle's Eye
   ~ West End Mines
   ~ Oak Grove Canyon
   ~ Willcox to Bonita bicycle loop
   ~ Cluff Ranch Wildlife Area
   ~ Mcilheney water scam
   ~ Tollhouse Canyon
   ~ San Carlos Lake
   ~ Upper Lower Middle Box
   ~ Cedar Springs
   ~ Crystal Hills ( restricted )
   ~ Pima Gap
   ~ Round Mountain Rockhounding
   ~ Morenci Crystal Cave
   ~ Frey Lake
   ~ Grantham Cave
   ~ Fisherman's Point
   ~ Stockton Wash
   ~ Gila Box float trips
   ~ Buford Canyon
   ~ EAC Library
   ~ Secret hidden springs with fish
   ~ Mesquite Bosque
   ~ Spring Canyon
   ~ El Capitan Canyon
   ~ York Valley bicycle loop
   ~ Roper and Gillard hot springs
   ~ Back Country Byway
   ~ Arizona Eastern Railroad ridealong?
   ~ Mt. Graham Ice Caves
   ~ Red Knolls pseudokarst
   ~ Safford Valley Grids
   ~ Eagle Creek Bat Cave ( restricted )
   ~ Guthrie Peak
   ~ Cliffton Historic District
   ~ Johnny Creek Loop ( restricted )
   ~ Obscure Gila River access points
   ~ Roper Lake
   ~ Hidden streams near Treasure Park
   ~ Lower Marijilda Crossing
   ~ Round the Mountain tinajas
   ~ High Creek Road
   ~ McEwen ruin
   ~ The "lost" CCC camp
   ~ San Simon dam
   ~ Marijilda Ruin
   ~ Lebanon Reservoirs
   ~ Pima Museum
   ~ Table Mountain Mines
   ~ Abandoned old highway 70 bicycling
   ~ Engle Orchard
   ~ Deadman Ditch
   ~ Old Safford Bridges
   ~ Big Lue Mountains
   ~ Prehistoric agave roasting rings
   ~ Bramaham Cave ( permit required )
   ~ Allen Reservoir
   ~ Bear Flat
   ~ Riggs Lake
   ~ Whitlock Cinega Hot Lake
   ~ Webb Peak Lookout
   ~ San Jose hot well
   ~ EAC Anthropological Exhibits
   ~ Zeolite beds
   ~ Carter Sawmill
   ~ Shingle Mill Canyon
   ~ C119 at Pima International Airport
   ~ Goat Hill ruin
   ~ H X Dam
   ~ Taylor Pass
   ~ Greasewood Range
   ~ Dutch Henry Trail
   ~ Pima Badlands
   ~ Muleshoe Preserve
   ~ Arivaipa Ghost Town
   ~ Old Morenci Trail ( restricted )
   ~ Turtle Mountain
   ~ Flying Butress Dam
   ~ Dankworth Ponds
   ~ Guthrie
   ~ Paddy's River
   ~ U of A agricultural Research Station
   ~ Apache Box Buddhist Retreat   
   ~ Mt. Graham sawmill
   ~ Bear Basin
   ~ Toppy's Cave
   ~ Wood Canyon
   ~ Deadman Falls
   ~ Grahm County Historical Society Museum
   ~ Santa Teresa Rock Climbing
   ~ West Peak
   ~ Amerind Foundation
   ~ Copper Canyon
   ~ Old Marble Quarry
   ~ China Peak Observatory
   ~ Redfield Canyon
   ~ Power's Garden
   ~ Fishhooks Wilderness
   ~ Whitlock Mountain Ski Condos
   ~ Day Mine Road   

Are we there yet?

Here in the Gila Valley, it is not all that unusual
to come across prehistoric potsherds dating from
the twelfth or thirteenth century. Potsherds have
a unique property in that the pots are extremely
delicate and break easily, but the broken pieces
are virtually indestructible. And keep forever.

Many potsherds can be dated and time stamped.
Relatively through style, art, temper, thickness,
tradeware routes, rim finish, color, slip, history,
and such. And absolutely datable by way of
thermoluminescence or paleomagnetism. And
association datable by using tree rings or C14
techniques. And, of course, stratigraphy.

With one exception, the Gila Valley is somewhat
of an archaeological backwater. Only a few dozen
mid sized sites are known, and most of them are
more "interesting" rather than spectacular. Most
artifacts seem to be more of the "steal the plans"
variety or imported from elsewhere.

That exception is collectively called the Safford
Valley Grids. These are rectangular groupings of
"fields" formed by removing rocks from their
centers and carefully and completely bordering.
There are many thousands of these, typically
grouped by the dozens or hundreds. But sometimes
occurring in singles. Often 15 by 20 feet or so.

These show up quite clearly on Google Maps.

They are invariably sited on mesa top like
benches of very carefully selected Gradual
slope and aspect and soils. Typical absence
of other artifacts and their arrangement
strongly suggests some agricultural use,
Such as dry farming or hand irrigating.

Obviously, countless hours of engineering and
use and maintenance went into their use.

A mystery to me is why they would piss around
with infertile and rocky benches of poor soils,
limited rainfall, and sparse vegetation. When there
was a perfectly good river and easily irrigated
fertile bottomlands a few hundred yards away.

One offered explanation is that malaria or other
fatal diseases frequented the "vapors" of the
bottomlands. Another is the seasonal variability
and violent floods of the Gila River itself.

The definitive reference on these is The Safford
Valley Grids, available as U of A Press Anthro
Papers #70. Many of the sites are easily visited
by a mile or two dayhike. Most are on public lands.

One of the least expected sources for genuine
new energy developments is Keelynet.

Among all of the pseudoscience dreaming
and perpetual motion scams, enough real
developments accidentally sneak through to
make this site worth a daily visit.

As mentioned before, at one time I was very
big on Book-on-Demand publishing. But
these days, I strongly feel that ALL dead tree
books are clearly doomed. Given the near
certainty of an upcoming decent reader, the
advantages of eBooks utterly overwhelm.

There are several interesting directions in
BOD that seem worth investigating, though.

The first of these is Lulu. Which is basically
an online vanity publisher. Who can handle
book production and marketing for you in
any quantity at rates that are not totally
outrageous.

The second is good old Gigabooks. Who
show you how to use classic hand bookbinding
techniques for your own low end, low cost,
( but highly labor intensive ) publishing.

The key ingredient for BOD that never showed
up and likely never will is a $500 book finishing
machine that trims and binds on a desktop. With
perfect bound hard and softbound quality and
appearance comparable to bookstore standards.

Meanwhile, high end solutions abound that make
absolutely no economic sense to me. Such as
the Espresso Machine from OnLine Books. And
costing a mere $100,000.

Let's see. Say you finance one of these at 10
percent for 5 years. Your monthly payment will
be $2124.50. Which means if you sell a thousand
books per month, the machine alone will cost you
$2.13 per book. Assuming, no scrap, free labor,
no down time, no royalties, no rent, and absolutely
free materials.

And assuming the book world will not be totally
and radically different within 5 years.

A thousand books per month, consistently day in
and day out is a huge number. Also, a thousand
books per month is thirty books per day. Or roughly
four books per hour. If it takes more than fifteen
minutes per book total production, the machine will
be unable to exceed a thousand books per month.

Thus clearly boxing itself into a financial corner.

The latest issue of Science magazine gives
further credibility that ethanol from corn
under US farm conditions is simply a fifteen
billion dollar vote buying scam.

Your tax dollars at work.

Tilman, Hill, and Lehman, vol 314, 8 December
2006 pp 1598-1600. "Carbon-Negative Biofuels
from Low-Input High-Diversity Grassland
Biomass".

In which they find out that plain old wild grass in
mixed species on largely unattended marginal
non-croplands completely and utterly blows US
ethanol from corn away. While also providing
carbon sequestering in the roots of the perennial
crops. With insanely lower new energy inputs.

More in our Energy Fundamentals tutorial.

Added additional links to our Arizona Auction
Resources page.

Your own custom regional auction finder can be
created for you per these details.

I can do this for you a lot faster, cheaper, and better
than you can. But if you must make up your own, here
is how to go about it:

Start the listing with the National Auctioneers
Association and your local equivalent to the
Arizona Auctioneers Association. Then compile
a list of local auctioneers from their members
listings that have web sites. Visit the sites to
make sure they are currently active. Eliminate
any realtor sites that do not also have estate sales.
Or any superspecialized services not of interest.

Then compile a list of national auctioneers that
may be active in your areas. This is similar to
the middle column of our typical listings. Next,
compile a list of links to all of your state's
universities, community colleges, schools,
cities, towns and counties. email any likely
candidates to find out when and how their surplus
property is disposed of.

Then add the other web auction finder resources.
Most of these could not find a pig in a disphan
when it comes to Arizona auctions. Typically there
will be 65 to 150 Arizona auctions announced at
any one time. See how many you can find how
fast without my help.

Then go on to column three. Where you get a
national newspaper list, a state newspaper list,
and links to all of the Craig's List resources
nearby.

Next, you go through each and every newspaper
for your state, finding which ones have active
classified sections and which of those have enough
auction listings to be worthwhile. Some newspapers
(but only those that wish to survive for a few more
weeks) will also have web access to their display
ads as well. Be sure to seek these out and link them.

You might also want to create a list of all of the major
companies and employers in your area. Again to
seek out which have surplus property routinely
disposed of.

Finally, you recheck each and every listing, asking
all auction sites to put you on their email lists.

Google just released the beta of their new Patents
Search page. As usual, it looks like this will
completely blow away the earlier patent resources.

Winners, of course, are in the marketplace and
losers and failures are in the patent repositories.

For most individuals and small scale startups
most of the time, any involvement whatsoever with
patents and patenting are virtually certain to result
in a net loss of time, energy, money, and sanity.

Per this tutorial. Or in the very few instances
when a patent may make sense, this one.

More on our Patents library page. More on
product development on our Blatant Opportunist
and GuruGram library pages.

What you may be used to and works really well in
one computer language may end up with lots of
rude surprises in another.

I'm very much a manic fan of PostScript, and use
it for just about all of my general purpose computing
needs. But our Magic Sinewave work demanded
some 64 bit math and web page interactivity that is
better met by JavaScript. So, I have been using
both of these recently, even letting PostScript write
the JavaScript code for me.

I lost a bunch of time yesterday over JavaScript
apparently being unable to find its own variables.
Turns out this is a normal and expected "feature"
instead of a bug.

In PostScript, any variable inside any proc will
be freely available to any other proc. Unless you
go out of your way to isolate it with saves and
restores. There is rarely any need to predeclare
your variables in PostScript. In fact, this is
pretty much unheard of.

But with JavaScript, a variable used inside a
function is ONLY available to that function on a
local basis. Other portions of the code will be
unable to reach that variable and generate a
"not found" error message.

Thus, for most JavaScript uses most of the
time, predeclaring your variables at the beginning
of your code is a must. Using the var directive.

A second major surprise is that PostScript's math
functions are directly in degrees, while JavaScript
uses radians. The adjustment is to multiply all
JavaScript degree values by pi/180.

I sorely miss an equivalent to {} forall. And JS
is much fussier over syntax. Its parenthesis can
have multiple meanings as well. And the dup ==
debugger in PS seems more useful than the alert
in JS. Especially since I've found no way to cut
and paste out of a JS alert. Except by hand.

More on PostScript in our Beginner Stuff and
our Gonzo Utilities.

The holiest mantra of card carrying members of the
Church of the Latter Day Crackpots ( for whom, of
course, Nikoli Tesla is the patron saint ) is "But you
have not done the experiment!"

Well, firstoff, extraordinary claims ALWAYS demand
extraordinary proof. Secondly, it's ALWAYS up to the
claimant to provide their proof, not up to others to
disprove them.

But most importantly, research time and research
dollars are valuable. The foremost rule of legit
research is to not waste time or dollars on anything
that has a negligible probability of success.

There ALWAYS has to be some basis for optimism
or some ability to rip off federal funds for any valid
research to proceed.

Obviously the same bad beginner's mistakes that
everyone else makes by an obvious con artist fraud
attempt are rarely enough justification for serious
commitment of new research dollars.

Two cases in point: The Meyer Water Powered
Car fiasco: In reality, the "experiment" is rerun
millions of times daily. There are commercially
successful devices including EDM machining
and Qprox products that simply would not work
if Meyer was correct. Not to mention that there
is a whole EIS field ( short for Electrochemical
Impedance Spectroscopy that routinely has
never shown even the tiniest shred of evidence
of any Meyer validity.

And those old Bearden Claims have newly been
trashed in the latest issue of Skeptical Inquirer.
For Bearden to be correct, an utterly incredible
number of people would have to have been dead
wrong for a ridiculous time in countless different
ways.

As we have seen before, finding an unlimited
source of free energy would be one of the most
heinous crimes imaginable against humanity.
One that would make Hitler look like Mother
Teresa. And one that would rapidly convert
the planet into a cinder.

More on pseudoscience bashing here.

Continuing yesterday's discussion of technological
developments we can expect to happen soon...

    INCREMENTAL BATTERY IMPROVEMENTS -
    Electric vehicles and truly useful hybrids certainly
    could use better battery technology. As could laptops
    and PDA's for longer battery life. A major technology
    breakthrough is unlikely, but small ongoing advances
    seem the norm. And we already have NiMH cells
    whose AA size capacity today exceeds yesterday's
    sub-C's. NiCad technology recently discovered how
    to use three outer electrons rather than just two for
    a potential 50% increase. Use matched versions and
    better manufacturing are routinely appearing on
    product shelves. And a major upgrade in lithium
    energy density is in the works.

   MAJOR WEB CRASH-- Yesterday, I got 2,350
   spam emails ( well filtered by our ISP ), 50 readable
   messages of which twelve were remotely useful.
   And the signal to noise ratio continues to deteriorate
   at a dramatic rate. If this is typical of others in the
   web, it is reasonable to predict that the web is near
   certain to totally choke on its own vomit within a
   very few weeks. The fundamental problem is that
   email arrives postage due. This feature must be
   eliminated if the web is to survive.

  CHEAP SANTA CLAUS MACHINES - "Printers"
   that print objects instead of messages are the crux
   of the upcoming Santa Claus Machine revolution.
   One intriguing low end approach is the RepRap
   system that is eminently hackable. Here's a typical
   service house, an example of architectural uses,
   a good Gateway Link site, and a good summary .

  CARBON NEUTRAL FUELS - We have already
   seen below why the Hydrogen Economy flat out
   ain't gonna happen. By far the best means of
   storing hydrogen is to bond it to carbon in a
   convenient room temperature liquid. Iso-octane
   and Heptane have proven themselves to be
   especially adept at this. Advantages of carbon
   neutral approaches to alternate energy (as
   opposed to carbon free ) are that (1) A significant
   portion of the stored energy is provided by
   the carbon component; (2) carbon appears to
   be essential for the creation of room temperature
   liquids of acceptable energy density and safety,
   and (3) energy losses from reformation can be
   eliminated, and (4) delivery infrastructure is
   entirely or at least largely in place.

   CLASS "D" AUDIO AMPS - I first ran a
   tutorial on these in the February 1966 issue
   of Electronics World and had a somewhat
   more recent summary appear here. While
   these have been "somewhat slow" out of the
   starting gate, they are now coming on like
   Gangbusters. Offering you high power, low
   distortion, extreme efficiencies, no crossover
   issues, and -- recently -- minimal output filtering.
   Leading suppliers today include Analog Devices,
   Maxim, National and Texas Instruments.

Here's some predictions of some technological developments
we can expect to happen soon...

    BIG LED'S - White light emitting diodes are now
    available in one and five watt levels approaching a
    dollar each at ever increasing efficiencies that can
    completely blow away fluorescents, let alone older
    incandescents. Expect a major move from specialty
    to routine lighting shortly. LED Journal is one good
    info source.

    EBOOK READERS - There is not the slightest
    doubt that the utter demise of dead tree books is
    imminent. An eBook reader or its replacement
    can be expected "real soon now" that offers
    better legibility, better convenience, and better
    currency that traditional publications. Aided and
    driven by a student revolt against backpacks. And
    accepting standard .PDF and other files with DRM
    a minor sideshow at best.

    TOOTH DECAY GONE - A cheap once-a-week
     mouthwash that completely eliminates dental
     cavities sounds too good to be true. But it does
     seem to be completely legit research proceeding
     at a faster than expected pace Per this original
     research, this activity today, and these links for
     more info.

    NET ENERGY SOLAR CELLS - As we have seen
    in our Energy Fundamentals tutorial, not one net
    watthour of conventional silicon pv electricity has
    ever been produced. Nor is any ever likely. Today
    the price of a synchronous inverter alone often will
    consume more than the value of the electricity run
    through it in its amortization. And that is before the
    recent price runup and unavailability of silicon. But
    new developments in SIGS solar and Quantum dots
    may make all of conventional silicon ancient history.

    $500 TRUE HDTV -- Flat panel displays have
    finally turned the corner. Besides routine and
    cheap 24 inch computer monitors, we can expect
    a less than $500 and 32 inch HDTV with a full
    1050 interlaced scan. Only it would be ludicrous
    to still call it a television set, because almost
    all viewed material will be computer, DVD, cable,
    or direct web download. Expect CRT's, plasma, and
    rear projection to die. And LCD's to dominate the
    smaller screens, while DLP mirror technology will
    handle the high end. Meanwhile, expect theaters
    to continue their death wish by refusing to upgrade
    to all digital screens.

Added some new links and corrected some older ones on
our Arizona Auction Links directory.

A photo and imaging tip left over from way on back
in my Popular Electronics days: There's this cheap
spray-on glop called Krylon Dulling Spray. Which
can dramatically improve anything bright on an
object that threatens to white burn. But the stuff is
somewhat difficult to remove and should not be used
on very high value or high tech items.

Additional photo and other tips here and here.

We have been having some email problems over the
last few days. Be sure to recontact us if anything
important slipped through the cracks.

Our email service does have an outstanding spam
filter. BUT -- yesterday we have 1240 spam messages,
54 real messages and 12 useful ones. If this is typical
of others in the web, it is reasonable to predict that
the web will totally choke on its own vomit within a
very few weeks.

Here is my proposed solution: The crux of the issue
is that email comes postage due. Eliminate this
feature. Make every email cost ten cents to send.
A recipient will have three options (A) accept the
source verified email at no cost to the sender, (B.)
accept the email and pocket nine cents, or else
(C.) return the email to the sender with another ten
cents due.

The average user would gladly pay a few dollars
a year to completely get rid of spam. And just
might get filthy rich if spam continues. And
the penny differential should pay for the service.

New ideas are just like pancakes or children.
You should always throw the first one away.

Magic Sinewaves are a newly discovered approach
to energy efficiency and power quality

Latest GuruGram #72 is on Faster Magic Sinewave
Zero Solutions.

Gonzo sourcecode for GG72 is available here..

Latest GuruGram #71 is on Enhancing your
eBay Tactical Skills V.

Gonzo sourcecode for GG71 is available here.
Earlier tutorials in the series can be found
here, here, here, and here.

Too good to be true department? I might have
found a way to make Magic Sinewave solutions
more deterministic very near their zeros.

If so, this should dramatically speed up analysis
time for very high zero harmonic solutions.

Here is the reasoning and some very preliminary
math. The full set of equations to be solved appears
here on page four. One seven pulse harmonic equation
to be zeroed might be something like...

      cos (5*p1s) - cos (5*p1e) +.... +
                           cos(5*p7s) - cos(5*p7e) = 0

At present, when we are very near a zero, we use
Newton's Method or "shake the box" to improve
our answers. We make a small change in the p1s
edge and see if things get better. We keep making
small changes till a minimum is found. The "small
change" is then reduced and the process repeated
first for p1s and then for other pulse edges till an
acceptable number of decimal place gives a desired
accuracy.

At present, dozens to hundreds of complex iterations
are used. The method works very well, but it is slow
if more than a few dozen harmonics are to be zeroed.

Instead, if we are only changing p1s, everything else
is a constant ( this has already been used to speed up
the existing code ). We will also assume that the
new p1s cosine can be linearly approximated because
it is so close to the old one.

Now, if we are only dealing with the fifth harmonic,
we have a constant, an error, and a p1s edge angle.
We can exactly take out all of the error by finding
a new p1s offset of -b/m. This follows from the plain
old y = mx + b equation of a straight line. b is our
error and m is our slope. Our slope is known and
will be the negative harmonic sine of angle p1s.

Sadly, this will be likely to make the other harmonics
worse, so only a portion of the offset will be useful.
The problem instead is to take a pile of y = mx + b
equations and simultaneously try to reduce their rms
error. Reducing the square of the errors is simpler
and works as well. We thus seek a minimum of...

       z = (m3x + b3)^2 + (m5x + b5)^2 + ...

A minimum (or maximum) results by finding the
slope and setting it to zero. Which seems to tell us
after expanding, differentiating, and regrouping
that...

angle correction =
         -(m3b3 + m5b5 + ...)/( m3^2 + m5^2 + ... )/

Which implies that the best angle correction can
be directly calculated once rather than finding it
through a many step iterative process.

Much more on this as it develops.

Your comments welcome.

November 30, 2006

Apparently Dr. HTML has gone dark. This
was a superb free web site testing service.

Please let me know if you find any other useful
similar sites.

Latest GuruGram #70 is on Enhancing your
eBay Tactical Skills IV.

Gonzo sourcecode for GG70 is available here.
Earlier tutorials in the series can be found
here, here, and here.

We are discontinuing sales of our slot machine
displays. Despite careful packaging, too many
are having shipping problems. Refurb also takes
me longer than the value they are returning. But
basically, I am tired of pissing around with them.

These remain perfect student PIC projects. And
( lacking any and all provision for coin or token
mechanisms ) they are completely legal in all
jurisdictions.

The remaining 45 or so are now available at $2.99
each. Strictly FOB Thatcher AZ. We will not
assist you in shipping in any manner beyond
free loading.

All are individually boxed. Some extra bulbs and
other parts are also available.

The Government Liquidation saga continues.

At one time, U.S. military surplus was sold
directly through their own DRMS service.

Where high risk and gross inconvenience got
combined with negligible competition and
unbelievably outstanding bargains. And for
which we did extremely well with everything
from nuclear holocaust fashion accessories to
tinfoil hat liners to water soluble swimsuits.

Sadly, someone discovered that it was costing the
feds $1.65 in admin costs for each dollar in sales,
and that simply dumping the stuff outside the
main gate with a "FREE" sign on it would be far
cheaper. Instead, the feds elected to privatize
mil surplus sales.

The main beneficiary of which was Government
Liquidation, a for-profit firm in Scottsdale, AZ.

Things started out really great. But over the years,
the Government Liquidation closing bid prices of
test equipment went up so high we could no longer
afford to bid on them. Minimum bids were raised
to $50 on all items, some of which would have
been risky at $2.50.

At least some of their site managers became
legendary in their customer rudeness, inflexibility,
and intentional hassle creation. Maddingly
infuriating bid extensions meant that your
optimum bid window was a few milliseconds wide
precisely 42.7765 minutes into their closing hour.

Many military bases dramatically tightened
security and entry hassles. At least one base
would not make a phone call for you over a
200 foot distance.

As a result, we have personally been forced to
scale way back in our mil surplus involvement.
Substituting industrial distress auctions that
seem infinitely superior on most counts.

And, just as the mil surplus scene could not
possibly get any worse --- it did.

It seems another someone discovered that the
DRMS is still grossly inefficient. Similar to the
British Sailing Ship Bureaucracy whose size and
costs peaked many decades after the last sailing
ship was removed from service.

Apparently many surplus warehousing centers are
being closed or downgraded. One consequence
appears to be that certain Government Liquidation
sites now list only a few worthless scrap items
instead of hundreds of potentially useful ones.
As in down to useless dregs. Arizona and New
Mexico seem particularly impacted.

As near as I can tell, DM Tucson surplus material
will be transferred to Northern Utah to save on
shipping costs. A mere 954 miles. Southern New
Mexico stuff, of course, goes to Central Colorado.

Obviously, there is definitely a major glitch today
in mil surplus availability. Whether it is temporary
or permanent remains to be seen. As will what role,
if any, that Government Liquidation will play in
future surplus opportunities.

Or your own role, depending on where you live.
The places to be appear to be Eglin FL and
Mechanicsburg PA..

Some more notes on wind power: Turns out the
Aermotor farm windmill folks are still in business
ad still grinding out classic inefficient windmills.

Pricing varies with tower height and diameter,
but around $12,000.00 is ballpark for the mill
and the tower combined. Naturally that is for
pumping only. A generator and such costs a lot
extra. As does shipping and installation labor.

Their recommendation is to have the blade bottom
an absolute minimum of fifteen feet above any
building or terrain, with an unobstructed distance
of 400 feet in any direction.

Normal recommended minimum spacing for more
modern wind farms is four blade diameters center
to center and at least twelve blade diameters back
to back.

The efficiency peak for most windmill technologies
has a rather narrow tip speed to wind velocity
ratio range. Thus fancy electronic regulation would
be a must if efficiency is to be optimized

Yes, there are vertical axis windmills that offer
interesting engineering tradeoffs. Some of which
might behave better near a building roof. These
the Darrius and Savonius Rotors. Both are low
efficiency and do not appear to be of much current
interest in "real" wind farms.

I get the impression that these "gee whiz" approaches
had their chance and blew it. Besides low efficiency,
the Darrius Rotor will not self-start and the Savonius
Rotor speed range and efficiency is much lower.

Few people are aware how severe a restriction the cube
of wind speed is for efficient operation. You normally
would not want to live or work in any building where
the average wind speed was high enough to make for
practical energy production.

Even if you put windmills on the roof of such a building,
planting trees or earth berming to shield the wind would
save bunches more on heating and cooling costs than
any windmill scheme could ever hope to compete with.

Besides being ridiculously cheaper.

Decided to double check some fundamentals of
wind energy over yesterday's post. Mukund Patel's
newly revised Wind and Solar Power Systems
would seem a pricey but good place to start.

Where we find that recoverable wind energy is
nominally proportional to the square of the blade
diameter and the cube of the wind speed. We also
see that the maximum theoretically recoverable
wind energy is 59 percent and corresponds to an
exit wind velocity of one third the input speed.

Real world recoverable efficiencies include fast
two blade modern systems at 45% and the old
farm Aeromotors at 30%. But peaks are only
over a very narrow range of blade tip speed vs
wind speed.

Sure enough, working anywhere remotely near
the ground or a building roof is studiously
avoided. Towers at least four to five blade
heights are recommended for smaller diameter
systems.

Which makes this "what were they thinking?
building having many small windmills directly
on the front edge of its roof be what those French
Veterinarians would call a "four paw".

The number of physical, electrical, economic,
and psychological fundamental errors here totally
boggles the mind. I'd give the blades three weeks
flat before the differential speed across them totally
wipes them out Naturally, their employees will
be long since gone because of the psyc stress.

When properly full burden accounted, there is no
way in hell that this absurdity can ever become a net
energy source. And thus will forever remain a gasoline
destroying net energy sink.

Surely they did not do this on purpose. If they did,
I've got some ski condo timeshares in Arizona's
Whitlock mountains I'd like to sell them.

There's a lively discussion going on over in the
sci.electronics.design newsgroup over an individual
who thinks that wind microarrays could form a major
alternate energy solution.

To me, this looks like the usual case of someone
who does not have the faintest clue what they are
getting into, have not done their homework, and,
for that matter, don't even know what homework is.

Naturally, they started off running to a patent attorney
over a concept that ( if workable ) would be totally
obvious to a practitioner in the field. Besides having
centuries of prior art. And thus patently worthless.

Ideas succeed only when they reach the working model
beta test stage. Once long ago and far away, unproven
ideas used to sell for as much as a dime a dozen. But
these days, ten cents a bale in ten bale lots is wildly
beyond unrealistically optimistic.

Fundamentals of why patents and patenting can be
ludicrously absurd for most individuals and small scale
startups appear here, while some guidelines of when
patents actually may be marginally useful appear here.

The bottom line is that many practitioners in the field
clearly agree that wind energy does not scale downward
worth a damn.

Several obvious things would appear to go against
very small windmill microarrays. The first is that
rule #001-A of wind engineering is to have a laminar
flow that is free of turbulence or areal gradients. This
is one of the major reasons why useful windmills are
placed on high towers.

Smaller windmills placed on building roofs have an
obvious and grievous flaw: The wind velocity at the
bottom blade tips is near zero, owing to the boundary
conditions of the air layer in contact with the flat roof.
Giving us a second reason why useful windmills are
placed on high towers. The third, of course, is that
wind is much stronger at significant unrestricted
altitudes above ground level.

There also can be very serious unresolved infrasonic
problems that cause extreme psychological stress
in people living in buildings with windmills attached.

But the main problem is that the economics suck.

Let's look at one possible set of numbers: Four meters
per second average wind speed is way high for my area
and many parts of the country. Which has an energy
capability of 39 watts per square meter.

Best possible theoretical recovery would be 59 percent,
and a smaller wind device more likely 20 percent. So,
about eight watts per square meter effective recovery
for a wind device of one square meter area, or about
a 44 inch or so diameter.

Total production in 24 hours would be 192 watthours
per day. Or around two cents of electricity per day at
ten cents per kilowatt hour avoided cost.

From http://www.hsh.com/calc-amort.html we see that
$45.23 invested at ten years at ten percent amortizes
out to two cents per day.

Thus there would be NO NET GAIN whatsoever if the
wind generator cost more than $45.23. Assuming zero
labor and installation and maint costs.

To be a reasonably worthwhile endeavor, the 44 inch
windmill thus must sell for a lot less than ten dollars.
Including both its regulator and synchronous inverter.

Smaller wind machines, of course, would have even more
absurd economics. There is also a tendency for wind
machines to not do well in their hub area. Smaller machines
might have a larger hub to useful blade ratio, further
diminishing their meager capabilities.

Now, yes, these figures are conservative. But the point
remains that no calculations whatsoever were apparently
done. And it seems unlikely to me that any could be done
for most areas of the country that would be in any manner
better than conventional solutions.

At least one eBay seller is completely honest when
they stated " Buyer is responsible for all fright
arrangements".

Which should become a classic right up there with
the long ago offered swash stickers.

Photo postprocessing can get downright obsessive
at times. Ferinstance, what can you do when an
otherwise useful image has some bad burns in it?

Such as these "before" and "after" images that
you can click expand upon...

     

The first thing is to decide whether reshooting or
extensive rework is worth the time and effort.

Next try some "around the edges" easy stuff to
see if somewhat reducing the hot spots makes
enough of a difference.

Then try crowding. In which you make the near
problem areas more contrasty, sharper, and
more uniform. Being very careful not to get too
contrasty or too obvious.

Followed by attempting direct repairs from the
outside working in. Then focus on the worst of
the "beyond salvage" areas.

From the good parts of the photo, extract some
pasteable objects. In this case, grab two or three
capacitors to a work area. White outline them and
correct any overlapping wires or defects. Then make
-90, +90, and +180 degree rotated copies of these to
form a catalog of useful shapes.

Paste the most obvious shape into the worst spot.
Being careful to de emphasize it by having part of
it under the label or having wires cross it in an
appropriate manner. Extend or modify the existing
wires so they center on the capacitor ends.

Continue working on smaller and smaller details,
again raising contrast somewhat, darkening, and
making things match and lineup. Be sure to avoid
any gaffes at this point, such as wires that do not
go anywhere or capacitors that are not connected.

Very small problem areas can simply be made
muddy enough or obscure enough that the viewer's
attention is drawn elsewhere.

Consulting and postproc services available. More
photo techniques on our Auction Help library page.

Order mixups are a bad scene all the way around.
It pays to set up all sorts of double checking by
at least two people to avoid problems. Your 30:1
Buy/Sell Ratio does give you a first order of
defense against an occasional mixup.

Writing the full customer name on the box where
the label will eventually go is a good idea. We
just had a mixup involving two orders whose first
name both were "Carl".

Placing a copy of the sender and intended
address inside the box on pricier items is also
a very good idea.

Trying to get items back or exchanging them is
usually bad news. IF one of the two recipients
has not yet received their half of the mixup, you
sometimes can ask them to refuse their delivery.

If you send something to the wrong person, they
are in every manner entitled to treat the item as
a gift. And are under no obligation whatsoever to
do anything that in any manner would correct the
problem. Especially on their own cost or initiative.

If the items are fairly inexpensive and you have
lots in stock, comping a new shipment makes by
far the most sense. If you are out of stock, a
prompt refund of all costs is a very good idea.

Chances are the customer will be back when
they get to keep some unordered free stuff.

Telling both customers the email address only
of the other might get they to exchange parcels
after your total refund. But be sure you are out
of the loop. Because neither is under any obligation
whatsoever to take any action or trust the other.

Even on a costly item, issuing an immediate
refund to the real buyer and prepaying return
shipping to the wrong one is a very good idea.
With, of course, the fastest possible emails that
explain what is coming down.

Should you actually get the item back, and should
it pass a very careful inspection, you can always
try re offering it to the original buyer. Whoever
sent the thing back to you deserves something
extra, such as a ten dollar gift certificate.

But only on receipt, of course.

Above all, find out why the mistake was made and
take steps to correct similar future problems.

More on similar topics in our Auction Help page.

How to spot an extroverted engineer: They
stare at your shoes, rather than their own.

Duh. Always try the obvious.

Only a third of our electronics store inventory
has been listed so far on eBay, and I have
recently gotten very lax about new listings.

One lame excuse has been that eBay just
introduced a fairly wonderful new way of
listing that was ridiculously faster and easier
than before. Then discontinued it because
of apparent quirks and bugs.

It finally occurred to me yesterday to use
old Adobe GoLive for my eBay listings.

This is ridiculously faster, gives you local
storage, makes links trivial, gives better
textwidth control, easy bolding, color
addition, use of tables, etc etc...

To interact, you copy the HTML code off
the GoLive screen and paste it into the
old eBay listing code. Being sure to only
have the description part and no headers.

Naturally, excess HTML is bad news on any
eBay listing. But forced paragraph breaks and
line lengths, URL links, bolding, an extra
image, or even limited additional color can
more than prove of value.

Sadly, Adobe does not seem to be further
developing GoLive. Apparently favoring
Dreamweaver instead.

More on our Auction Help library page.

Apparently, a web "sniffer" to monitor your
outgoing traffic is an incredibly complex and
subtle piece of gear. Because they seem to be
outrageously expensive and hard to get.

A freeware program called SNORT comes close.
But it is for the Linux crowd and there are major
Win XP use problems.

A simple "traffic monitor" utility that would
give you a thermometer display of your previous
day's activity and a URL list would go a long way
towards instantly spotting a spam source or
someone else taking control of your computer.

A warning that you sent 175,000 emails yesterday
might give you a clue that something might have
been slightly amiss. And ISP use of similar monitors
could dramatically reduce spam at its source.

email me if you know more about this than I do.

Factoring is an incredibly powerful math tool.
And I continue to be amazed by the "cubeless"
cubic technique that dramatically can speed up
cubic spline calculations.

Ferinstance At^3 + Bt^2 + Ct + D can be
reversed and factored down once into
D + t(C + Bt + At^2) and factored again
into D + t(C + t(B + At)) and computed
using nothing but multiplies and adds.

Rather than needing explicit calculation
of a square and a cube.

Similarly, yesterday's improved slope calc
is of form Ax + Bx^3 + Cx^5 + Dx^7.

Factor it once to get an intermediate result
of x(A + Bx^2 + Cx^4 + Dx^6) and again
for x(A + x^2(B + cx^2 + Dx^4) and again
for x(A + x^2(B + x^2(C + Dx^2).

While this does require calculating a square
(with the dup mul in PostScript), it eliminates
any direct need for third and fifth and seventh
order calculations.

Similar revelations in our Math Stuff library.

Did a little more work on the Spline-fitting-to catenary
we just looked at as GuruGram 69. Some rough and
preliminary additional analysis can be found as
PLOTERC1.PSL in our PostScript library.

First, the generalized equation for the slope of a
catenary as a function of a is ...

 dy/dx = x/a + x^3/3!a^3 + x^5/5!a^5 + x^7/7!a^7 + ...

which can be crudely but obviously coded as...

     /findslope { /xx exch store xx aa div
     xx dup mul xx mul aa dup mul aa mul
     div 6 div add xx dup mul dup mul xx
     mul aa dup mul dup mul aa mul div
     120 div add xx dup mul dup mul xx
     dup mul xx mul mul aa dup mul dup
     mul aa dup mul aa mul mul div 5040
     div add 1 atan  } store

This has the advantage of working for
any a that you predefine as /aa 0.5 store .
Making your code a lot more general.

This can be simplified and sped up by
letting w = x/a, reducing the equation
to...

dy/dx = w + w^3/3! + w^5/5! + w^7/7! +...

Finding the actual rms errors is not that
big a deal if you are willing to work in
constant t space rather than constant x
space. Doing so greatly simplifies things
and remains a very good approximation.

The process is to find the A-H spline constants.
then for incremental t, find the equivalent x
and the equivalent y. The real y is also found
as cosh (equivalent x/a). The difference is the
instantaneous error. And the square root of the
summed squared distance divided by the number
of samples will be very close to your rms error.

Finding A-H is trivial from routines found in
the cubic spline library...

/findAH {

                /A x3 x2 3 mul sub x1 3 mul add
                   x0 sub store

                  /E y3 y2 3 mul sub y1 3 mul add y0
                  sub store

                  /B x2 3 mul x1 6 mul sub x0 3 mul
                  add store

                  /F y2 3 mul y1 6 mul sub y0 3 mul
                  add  store

                  /C x1 3 mul x0 3 mul sub store

                 /G y1 3 mul y0 3 mul sub store

                 /D x0 store /H y0 store

          } store

Doing so fairly quickly reveals an amazingly
good solution at an initial tension of 0.841 and
a final tension of 1.642. With an average error
of something like 0.0002.

Consulting services available.

Refurb Log:

While engineering notebooks are pretty much
ancient history, it is still a very good idea to start
a textfile log on each and every repair item.

Such a log should show all disassembly details
as they occur, connector polarities and wire colors,
and, above all, your logic thinking at each and
every part of the refurb.

Picked up an apparently mint Wavetek 1081 at
an auction that had not been powered in decades.

Initial powering seemed to work fine, then went
to bizarre display readings and strange front
panel responses after a few seconds. Which
might be consistent with a blown electrolytic.

The project was valuable enough that a service
log was started and the CD Wavetek Manuals were
immediately ordered on eBay for $9. After receipt
of the CD, oversize schematics were carefully
printed out and taped back together.

A careful visual inspection revealed nothing unexpected
in the way of loose connections, bad odors, or any
apparently stressed parts.

The back panel connector to the companion 1076
was disconnected without change. The manual
shows a convenient 2x5 terminal block with all
of the supply voltages on the top board. Checking
this immediately revealed the -18 volt line was
really at -0.6 volts. And a power-off Ohms check
revealed 0.4 Ohms to ground. A value far too low.

A blown regulator could be expected at this point,
but Wavetek's magic "superregulting" power supply
circuit had a fairly high impedance in the regulator
ground line Which preculded a direct internal short.

Also, the 7819 regulator does have internal short
circuit protection, and the -0.6 volts sounded just
about right. Thus the regulator could be moved
way down on the suspects list.

Removing the other boards did not remove the
short, leaving the main problem on the power supply
board itself. The manual proved very valuable in
dealing with the oddbal front panel knobs. To
remove, you nonobviously pry the knob cap off
with your thumbnail. And then use a large (!)
screwdriver to release the internal clamp.

"Obviously", that big old output capacitor was
bad. But stupidly cutting it did not remove the
short. Reminding us of the key rule of NEVER
assume ANYTHING that you have not proven!

And its iatragenesis co-rule of NEVER hurt the
patient!

Turns out there was also a bunch of low level
marker circuitry also on the power supply board.
Fortunately, removing the most expensive and
impossible to replace Harris IC from its socket
did not alter the short.

After carefully studying the entire circuit and
every -18 volt path, an obscure 10 microfarad
bypass capacitor C148 was found. Sure enough,
unsoldering one end of the cap (which is what
should have been done with the previous one!)
left the power supply impedance up at a few K
and a near dead short on the capacitor itself.

At this point, there may have been collateral
damage and you could still screw up by wrong
assembly or a connctor orientation mixup or
a bent pin. And additional capacitors could
easily blow on repowering.

But the basic problem had been corrected.

(To be continued, one way or another)

Latest GuruGram #69 is on Cubic Spline
Approximations to a Catenary Curve.

Gonzo sourcecode for GG69 is available here.

Just picked up a HP Officejet Pro DTN inkjet
printer and am initially quite impressed with it.

Around $200 with rebate. Fast 37 pages per minute.
Separate cartridges for each color, so no throwing
away the magenta and yellow when you run out of
cyan. Instant automatic networking by plugging
it into a spare router slot. Optional full wireless.

But best of all, includes a simple snap-on duplexer
for automatic two sided printing. What happens
is the the back prints upside down and the page
is dried for a few seconds, then pulled back and
rotated 180 and the front prints normally.

Print quality is more or less "laser comparable".
Ink cartridges may be hard to find locally but
readily available online. Ink quality and endurance
has apparently been dramatically improved.

Thanks to Acrobat distiller, there no longer is
any need for PostScript built into any printer.
Just distill and print your .PDF file on any
modern printer.

Still need some advice and probably more than
some assistance in converting thousands of
Apple II 3-1/2 inch disk based textfiles and other
files into a format usable for our website archive.

Present thinking is to go serial, then USB, then
to a USB drive. email me with your suggestions.

Added several recent files to the menu in
our GuruGram library.

Latest GuruGram #68 is on Fundamental Factors
Underlying Technical Innovation.

Gonzo sourcecode for GG68 is available here.

Added several new links to our Auction Help
library page. Your own custom auction finder
can be created for you per these details.

Sometimes a sneaky workaround can be
better than either a sledgehammer cure or
no fix at all.

Our Log File reader custom software presently
has a hard limit of 65,535 log entry lines being
processed at once. The reason being that there
is a 65K limit on the size of any one PostScript
array.

Splitting and recombining the file is an obvious
solution for really popular websites. But that
would need a major overhaul of the code.
Another possibility is to scan and group URL
sources by user. But that would take thousands
of extra passes through the entire long log file.

A solution I am exploring is to go into a sampling
mode above 65K. You measure the total lines in
the log file. If they exceed 65K, you throw only
enough of them away at random to fit your max
permitted legal size.

Ferinstance, say your log file is 83K lines long.
You then pick a random number from 0 to 82.
If that number is less than 64, that line is used.
If not, that line is ignored.

You also print out the downsampling factor when
and where it actually is needed. And with that
many samples, your log info should still be much
more than statistically significant.

All done with trivial mods to the existing code.

What happens when you mix two signals together?

Because of two wildly different definitions of the
word "mix", untold confusion has resulted among
radio and audio engineers for many generations.

Mixing audio is done linearly. Meaning that you
simply add the two channels together. More often
than not, you will make one channel higher in gain
than the other seeking some sort of balance. And
you should end up with exactly the same frequencies
that you started with. Anything else may create
intermodulation distortion or other undesirables.

Mixing radio is done nonlinearly. Meaning that
you must in some manner multiply the two channels
together. With the intention of creating sum and
difference frequencies. One of which is normally
used and the other rejected by filtering.

The key to radio (and tv and wireless, etc... )
mixing is this ( or a related ) trig identity...

    sin(u) sin(v) = 1/2 [ cos(u-v) - cos (u+v) ]

Thus, some sort of a multiplication is needed
if you are to get sum and difference frequencies
as useful outputs. Our trig identity above uses a
true four quadrant multiplier. These were at
one time quite expensive, but routinely see
wide use today.

A multiplication can result from any sort of
nonlinearity, intentional or nonintentional.

Suppose a diode or a comm channel or
whatever has a nonlinearity expressed as
a Taylor series something like...

    f(x) = ax + bx^2 + cx^3 + ...

The "a" here on our linear term is simply
our linear channel gain. Very often the cubic
and higher terms will be very small and can
be ignored. Leaving our square term to
create our main nonlinearity.

Note that...

     [ sin(u) + sin (v) ]^2 =
     sin(u)^2 + sin(v)^2 + 2sin(u)sin(v)

and that our final term is a multiplication.
This term results in our sum and difference
frequencies.

The loss of a traditional rf mixer depends upon
how strong the amplitude of the square term
in the nonlinearity is.

Mixers (espeically those that are not true
multipliers or doubly balanced can suffer
from various aliasing artifacts. the worst of
these is the image at B + 2(A-B) = 2A-B.
Other problems can be input feedthru, half
frequency artifacts from self-squaring, and
similar unwanted crossmod spurs.

Got to rewatch the Tommy rock opera by
The Who recently. Which raised several
obvious questions...

    Was there some sort of symbolism in
     the hang gliding over the last supper
     scene?

   Could the Acid Queen sequence in
   some manner have been drug related?

   Could the scene where Mommy humps
   the Oscar Meyer Weinermobile be
   considered pornographic?

All in all, the final ten minutes include
some of the most spectacular and the
most stridently motivating music ever
recorded anywhere by anybody.

Our major GuruGram on the key secrets driving
recent technological innovation seems to now be
stabilizing. And there's either a light at the end of
the tunnel or else a train heading towards us.

So far, we have got ( with news dates below ) ...

1. Decoupling. (aug 18)
2. Accurate replication. (sept 25)
3. Elimination of the gatekeepers. (oct 23)
4. Computing power insanely beyond awesome (oct 25)
5. Nonlinearizing the tyranny of time (oct 27)
6. 24/7 instant gratification via time compression. (oct 30)
7. Teeny Nano New Nu. (oct 31)
8. Virtualosity through disembodiment. (nov 1)
9. Programmability (nov 3)

And these now seem likely...

10. Devaluation
11. Indirect payback.
12. Unexpected value added
13. Hybridization
14. Increasing competitiveness
15. Lesser trends wrapup?

But I have a hollow feeling that I've missed
an obvious biggie or two.

What would you add to the list?
Please email me with your suggestions.

Continuing our analysis of the key secrets driving
recent technological innovation, number nine on
our list would be programmability.

With programmability, a general purpose device
is "taught" to meet the needs of a specific user
while minimizing or eliminating entirely the need
for specialized, low volume, dedicated hardware.

While laptops and PC's are obvious examples,
the smaller embedded microprocessors such as
the PIC overwhelmingly dominate in the nunber
of units sold and their extensive use range.

A secret of programmability is often to repeatably
combine very simple steps of moving, adding, testing,
and performing fundamental logic operations. Many
individual steps repeated one at a time can go into
a programmable algorithm that gets a larger job done.

One key concept that makes a programmable
computer a computer is its ability to test. And
then, based on the results of that test, alter its
future course of action. A test might view a
single bit flag and do nothing if cleared or branch
somewhere else if set. Typical flags are based
on a zero result, a negative number, or a set carry.

Another key concept of many programmable
devices is the ability of something external to
interrupt the normal action and temporarily go
on to perform a special task. A programmable
device that can do many things at once is said
to be capable of multitasking. Which often is
nothing but the creative use of interrupts.

Small programmable modules called subroutines
greatly simplify and organize the program code.
At the same time, they allow their own reuse at
many different points in the program sequence.
A fancier type of subroutine with well defined
inputs and outputs might be called an object.

My own very favorite programming tool is
called table lookup. In which you simply find
a previously known answer instead of going to
a lot of trouble to calculate it. My Magic
Sinewaves extensively use table lookups.

Some programmable devices combine their
instructions and data into one area, while
others keep the two as separate as possible.
Leading to differing Harvard and Princeton
computer architectures. Each with its own
proponents and unique capabilities.

While programming can be done at the individual
bit level, fancy tools have evolved using assembly
language and various higher level languages.
Two differing routes towared programmability
involve interpreted code where everything gets
done as it comes up in sequence. Or compiled
code that does whole tasks but is more complex
to initially create.

More examples on programming here.

Here are the arguments against the hydrogen
economy:

    1. Terrestral hydrogen is only an energy
         carrier or transfer media and not a fuel.

   2. Terrestral hydrogen creation is inefficient
        as considerably more energy of usually
        much higher quality has to be input than
        is eventually returnable.

   3. No large terrestral source of hydrogen gas
        is known. Water, of course, is a hydrogen
        sink and, by fundamental chemical energitics,
        is the worst possible feedstock.

   4. The CONTAINED energy density of
       hydrogen by weight is a lot less than gasoline.
       And drops dramatically as the tank is emptied.
       The energy density of hydrogen gas by volume
       is a ludicrous joke.

  5. Virtually all bulk hydrogen is produced by methane
       reformation. And thus is EXTREMELY oil
       dependent.

  6. Hydrogen has the widest explosive range known,
       the least spark energy required for ignition, and
       has no known colorants or odorants. Its flame is
       often invisible or nearly so.

  7. There is more hydrogen in a gallon of gasoline
       than there is in a gallon of liquid hydrogen.

  8. No effective vehicle compatible means of hydrogen
      storage is known that is remotely as cheap, safe,
      dense, and convenient as carbon bonded hydrides.

  9. No infrastructure exists for gaseous hydrogen
      distribution. Pipelines in particular raise major
      density and embrittlement issues.

 10. Electrolysis from high value sources such as
        grid, wind, or pv is totally useless as a hydrogen
        source because of the staggering loss of exergy.
        There ALWAYS will be more intelligent things
        to do with the electricity.

  11. Improper burning of hydrogen produces highly
        polluting nitrous oxides.

  12. Terrestral hydrogen is basically a POLLUTION
        AMPLIFIER that INCREASES the pollution of
        its underlying sources. It is utterly ludicrous to
       claim that hydrogen is in any manner, way,
       shape, or form "nonpolluting".

  13. Hydrogen rots most metals through embrittlement.

  14. "Carbon Neutral" solutions would appear better
        than "Carbon Free" because (A) A significant
         measure of the energy of most fuels is in its carbon
         fraction, (B) Carbon appears to be essential for
         convenient and safe room temperature liquids,
         and (C) Reformation is not required or else
         is simpler, cheaper, and wastes less energy.

  15. An optimal hydrogen storage solution exists by
        carbon bonding as in heptane or iso-octane. Both
        of these room temperature liquids ain't broke.

More info here, here, here, and here.
An energy efficiency solution here.
Research and consulting services here.

Continuing our analysis of the key secrets driving
recent technological innovation, number eight on
our list would be Disembodied virtualosity.

It used to be that ideas, instruction, entertainment,
and escape were rigidly attached to their physical
distribution media. We had these funny pagey
things made out of dead trees we called "books".
Songs were molded into plastic disks with groves
or pits in them called "records" or "CD's".

Movies arrived on long pieces of tape called
"VHS" or on plastic disks with smaller pits called
"DVD's". More timely info or escape were on
floppier and even funnier pagey things we called
"magazines" Or once ( once very long ago and
far away ) "Newspapers".

And messages, of course, were ink placed on
paper and stuffed into envelopes. You than paid
the federal government to delay the delivery
of your message for up to a week or two.

The focus was usually on acquiring the physical
media rather than its actual content. Naturally,
when and where possible, manufacturers would
obsolete one media format and come out with a
new one. Requiring all previously purchased
content to be bought anew. Thus eight track no
longer reins supreme.

At the same time, an elaborate "priesthood"
developed around physical media creation
and distribution. With the result that the original
creative sources (authors, composers, scholars,
etc.. ) of the IP intellectual property only
received the tiniest fraction of the final selling
price of the physical media being distributed.
And even that tiniest fraction was often stolen
outright through contractual ripoffs.

In each case, the content was physically locked to
its medium of distribution. These days, though, most
information is simply disembodied bits floating around
in cyberspace. And deliverable at most any speed in
most any format, and freely accessible without physical
"container"constraints.

And, most importantly, readily convertible into any
chosen temporary physical format. Present or future.

Threre is no particular difference between the
cyberspace bits used to convey a movie or a song
or a math textbook or a power bill or a crawling-on-
the-grass gothic novel. And no particular reason to
relegate any wanted combination of bits to any
specific "hard copy" distribution media or format.

Established media houses who feel that
"business as usual" can in any way, shape, or
form continue are awaiting a rude awakening.
Especially if they feel that suing their best
customers is a viable policy. Or feel that IP
sources do not deserve and must receive the
lion's share of the final selling price.

Virtual reigns supreme.

A related concept are the virtual sets now
common in tv and movie production. In which
live actors can be dropped into any imagined
environment. And done so from any camea
angle or zoom. Going far beyond traditional
blue screen or green screen video source
switching.

Such movies as Toy Story, Robots, or Cars
even got shot completely on location in
cyberspace. And bits and pieces of CGI
virtual animation show up everywhere.

Yet another virtual concept involves exactly
how and where and by whom things are going
to get designed.

At one time, electronic circuits were first checked
by breadboarding and mechanical assemblies by
physical prototyping. Current practice uses virtual
emulation, sinulation, and modeling instead.  

And a reasonable prediction would be that the next
big things along these lines would be the widespread
use and dramatic price reductions of new rapid
prototyping systems or Santa Claus Machines.

Continuing our analysis of the key secrets driving
recent technological innovation, number seven on
our list would be Teeny nano new nu.

Things are getting smaller.

Many electronic components are now so small
you can't even see them and don't dare sneeze.
Which is bad for individual experimenters and
students, but otherwise has dramatically reduced
the size, weight, and cost of emerging electronics.

One early example was the Newtek Calibar. This
pen sized device completely replaced an entire
television studio full of test gear. While providing
cleaner and better waveforms to boot.

But you have to draw a distinction between "small"
and "really small." The latter also being called
nanotechnology. Great heaping bunches of very
interesting things happen when sizes approach the
molecular level.

Relative surface area goes up as volume goes
down. And all the rules of friction, stiction, and
attraction/repulsion dramatically change. For
instance, the horsepower per pound of a
nanoturbine is ridiculously higher than that of
its full size bretheren.

By going to nano pore sizes in Super capacitors,
their energy density can be significantly increased.
A new technology called Quantum Dots promises
to lead to pv solar panels that may someday become
true net energy sources that are both renewable and
sustainable.

Many unusual chemical and electronic and
medical properties emerge with strange shapes
at the nano level. Most notable of which are
Buckyballs and hexagonal carbon tubes.

And quantum computing itself on the nanoscale
promises to do complex things in interesting ways.

Chemists and biochemists in particular are excited
about building entire instruments at integrated
circuit scale. For chromotography, DNA analysis,
and even direct neural interconnects.

A cheap once-a-week mouthwash that completely
eliminates dental cavities sounds too good to be
true. But it apparently is completely legit research
proceeding at a faster than expected pace.

Per this original research, this activity today, and
these links for more info.

Continuing our analysis of the key secrets driving
recent technological innovation, number six on our
list would be 24/7 instant gratification via time
compression.

The industrial trade journals that are in the process
of self-destructing give us a classic example here. The
"standard" way of selling an indutrial product used to
be to take out a $15,000 ad in a trade journal, wait a
few weeks for it to be published, wait a few more weeks
for a bingo card recipient lead list, wait a few more weeks
to mail our printed literature or alert your reps, and then
wait a few more weeks for an order or two to dribble in.

These days, any more aware industrial supplier has an online
24/7 store. One that, of course, welcomes single and small
quantity orders. From anybody, anyplace, anytime. And
makes all of their pricing conspicuously obvious.

They also have full product info available, free downloads
of any software needed to run their systems, and totally
free repair and service manuals. Including continuing
full support on older and obsolete products. All created at
the tiniest fraction of traditional data books and print
service costs. And with zero waste or obsolescence.

At least this is true of roughly half of the industrial
suppliers online today. The rest will either quickly
follow suit or will shortly render themselves into
noncompetitive nonentities.

The outcome is not the least in doubt.

Meanwhile, there are new industrial data agglomerators
who gather in industry wide data for your instant and
fully objective availability. One major example of which
is the Data Sheet Archive.

Topo Maps are another obvious example. A complete
collection of any larger area was outrageously expensive,
easily harmed, and hard to maintain. For just a hiking
trip, a visit to an outdoors store might have been needed.

Today, you might go to Topozone and print out the latest
info on exactly the size and scale you need. Without any
infuriating borders or page crossings, even. But chances
are you'll be attracted to such newer and more flexible
services that combine map info with aerial photography
and other services. Such as Google Maps, Google Earth,
or TerraServer.

email is perhaps the most dramatic example of 24/7
instant gratification.

A Google Search, of course, is a nearly perfect example.
What used to be a frustrating trip to a library and a
possible six week wait for a no longer available
Interlibrary Loan that might or might not happen can
now usually be resolved in a few seconds.

Even at 2 AM on a Sunday morning.

A lively discussion of the new CIGS pv
solar breakthrough is going on now at
Slashdot.

Although I find it more than curious that
while many billions of dollars are being
poured into the fusion engineering rathole,
and while California is blowing nine billion
dollars by paying people to put gasoline
destroying net energy sinks on their roofs,
that a big deal is made at throwing a few
million of pocket change at what has extreme
promise to becoming a major and near in
solution to genuine renewable and sustainable
low pollution net energy.

More in our Energy Fundamentals tutorial.

Made another major auction find: A pair of
near-mint KIM-1 microprocessor trainers!
These 6502 devices predated the Apple and
introduced a generation to microcomputer
fundamentals. Besides being eminently
collectible, these have provenance as I actually
used them in my early microcomputer classes.

They are factory authorized KIM-1 clones
absolutely identical to the originals. And
INCLUDE genuine MOS TECHNOLOGY
wall chart schematics. These have been
slightly modified ( convenient power cables,
a ground clip, and a few output test points
for scope access ), but are eaily and fully
restorable to original condition.

Also have a bunch of related SYM-1 micro
trainers awaiting some checkout and refurb.
These are more of a back burner project.

email me if you have any interest in these.

Continuing our analysis of the key secrets driving
recent technological development, number five on our
list would be Nonlinearizing the tyranny of time.

"The moving finger writes, and having writ, moves on."
ain't necessarily so any more. At one time, type was
painstakingly set one line at a time. Typewriters worked
with one character at a time. Movies were edited by cutting
and splicing with scissors and glue above the cutting room
floor. Video similarly was A-B roll edited only in real time.

In book publishing, it was an absolute no-no for an author
to try and tell the publisher how the text and words were
to be arranged. And woe be to the author who dared try
to --gasp -- change even one word after typesetting. For
it was outrageously expensive and time consuming to
do so.

These days, of course, we routinely typeset first and
edit last. With zero cost penalties and great heaping
bunches of benefits. Because we can now do nonlinear
editing.

The original breakthrough in this ability happened when
word processors became screen oriented rather than
line oriented. You could now edit and change anything
you could see. Such techniques as cut-and-paste and
spell checking and advance outlining became routine.

And a whole new level of capabilities got added when
layout programs became document oriented rather
than screen oriented.

Missed a paragraph? No problem. Just let it ripple
on through the whole story. Regardless of how many
pages are impacted. Reposition a figure so its text
is relevant? No biggie. Tivial, even.

The similar breakthrough in video and movies came
about with the Video Toaster and related software.
That let you store your entire movie or show in
instantly accessible form. You could now easily
combine old and new material in any order regardless
of its time sequence. And, of course, CGI and sound
synchronization manipulated it in previously undreampt
of ways.

Thus shattering the tyranny of time.

Just picked up an exceptionally clean Wavetek
1081 + 1076 combo at an auction. This is the
system for broadband swept cable system
measurements. Usable to 1000 MHz.

It literally sat unused in a climate control cabinet
for most of its life and probably has less than three
hours of actual use.

We have yet to check it out and pick up the
readily available documentation for it.
Please email me if you have any interest in
this outstanding instrument pair.

Continuing our analysis of the key secrets driving
recent technological development, number four on our
list would be Computing power insanely beyond awesome.

Back in the olden days, adding one vacuum tube to a
system involved a major upheaval in size, cost, and
power consumption. But these days, adding another
half million transistors to say, an electric can opener,
is no big deal at all. Chances are they are already
sitting underused in the microprocessor of choice.

Engineering and other math analysis at one time had
to be super efficient because of the value of the user's
time. But throwing another ten million calculations at
something is now often trivial.

Thus, many brute force solution methods now become
eminently practical. And driven by the utterly unbelieveable
and unprecedented drop in the cost of memory and the
availability of raw computing power.

Two examples of brute force math techniques that have
only recently become feasible include our Magic Sinewaves
and our updated Fun With Fields tutorials. As have the
recent stunning advances in CGI computer animation.

Jeff Goddell's new Big Coal: The Dirty Secret Behind
America's Energy Future is a very good read.

Where we find that despite wildly overoptimistic
claims of recoverable coal reserves, coal remains
by far the cheapest and simplest means of generating
electricity in the US. With zero foreign dependence.

At least when such minor externalities as its
catastrophic global warming are ignored. And
by using "business as usual" accounting.

I find it interesting to compare the two power
plants in Springerville. To the west is a pv
plant whose only tiny little problem is that it
remains a huge energy and dollar sink. But
silently producing little polution noiselessly with
few employees and virtually no environmental
hasles. And incrementally expandable.

Whose expansion has been dramatically scaled
back since the economics today make no sense
whatsoever. Because there is no way in hell that
conventional silicon pv panels can ever become
a net energy source. When properly full burden
accounted at the system level.

To the east is a coal plant being significantly
expanded along with some patchwork pollution
add on repairs. Despite their source mine being
expected to run out of coal within three years.

Right now, coal remains king. But I expect pv
to eventually displace much of it, especially for
peaking uses. First because coal will become
more and more expensive as externalities such
as huge carbon dioxide loads are addressed.

And secondly because of new fundamental
breakthroughs in pv involving CIGS and
Quantum Dots and roll-to-roll room temperature
and pressure processing that hold great promise
for pv eventually becoming a useful net energy
source.

One that quite likely has the future potential to
become truely renewable and sustainable.

More in our Energy Fundamentals tutorial.

Continuing our analysis of the key secrets driving
recent technological development, number three on our
list would be Elimination of the gatekeepers.

Also known as disintermediation.

A gatekeeper was a profit stealing and time wasting
obstructionist between you and something you needed
or wanted.

The internet has not only eliminated gatekeepers,
but also the gate itself and the fence the gate was
at one time supposed to go into.

Obvious examples are used car dealers, manufacturer's
representatives, book publishers, music companies,
scholarly journal houses, and, ( sadly ) librarians.

Anyone reasonably swift can now use the online Blue
Book to find exact vehicle values. And can buy or sell
the cars themselves via Craig's List, eBay, or any of
a number of online Auto Trader publications. To the
point of which it is fundamentally insane to buy off a
used car lot.

I consider manufacturer's representatives to have
been the vilest of the vile, several evolutionary steps
below, say, lawers. These epsilon minuses made it
virtually impossible to find out how much something
actually cost. Or to get hassle free technical info in
a timely manner. Replaced, of course, by 24/7 online
stores and totally free instant technical information.

At one time book publishers offered extensive services
not available elsewhere. Such as art departments,
type setting, competent editing, or economics of scale.
With the possible exception of promotion and marketing,
there is nothing a modern book publisher does that you
cannot do faster, cheaper, and better by yourself. Or by
cash-and-carry targeted web alternatives. And their
competitors.

Similarly, the only purpose I can see for record companies
is to steal royalties off their talent. They accomplish
absolutely nothing useful otherwise and clearly no
longer serve any worthwhile societal purpose. Anybody
that wants to can now pick up their own recording
studio for pocket change.

Publishers of scientific journals are a good example
of gatekeepers in the process of self destruction. An
individual today has the choice of paying a journal
many hundreds of dollars to publish their paper after
a long delay and then making the issues so expensive
that their own school libary cannot afford a copy. Or
of instantly and freely web publishing to a worlwide
audience. To survive, the journal publishers must
realize that copies of every paper more than three
years out of date should be available free and without
ANY restrictions, not even a registration. Except
perhaps for a sane monthly access limit.

Librarians were classic gatekeepers. Many of whom
were control freaks. But the notions that the gatekeeper,
the infoseeker, and the infosource had to be in the same
place at the same time, or that the info somhow had
to be "returned", or that access was restricted to
anything less than 24/7 are clearly ludicrous today.
As to online access, Burger King ( and similar WiFi
locations ) offer better service and more info access
convenience than do most libraries.

With far fewer food and drink restrictions.

We will shortly have the last cabinets of our "2SX"
semiconductor assortments up on eBay. These
Pacific Rim replacement transistors have become
extremely hard to find.

We also have several dozen larger quantities of
specific "2SX" part numbers individually antistatic
bagged, often in groups of fifty. Many obsolete
semi houses have a $250 line item minimum on these
items, when and if they can be found at all.

With the new ROHS regulations, the price of these
can shortly be expected to go through the roof. We
are unlikely to find any more beyond our present
stock.

What is a "reasonable" amount to spend at a
live auction if you are serious about significant
eBay selling?

Firstoff, if you do not make a $25 mistake at an
auction, you are not bidding aggressively enough.
But otherwise, you should strictly "bottom feed",
continually offering only lowball bids. Preferably
on "contents of cabinet", "contents of room",
high count inventory, or on "pass - combine it with"
poisioned lots clearly requiring triage.

Distant auctions may involve additional fixed
expenses that can get waaay out of hand beyond
125 miles. But you can sometimes combine
other activities such as exploring restrauants
or visiting friends or camping or just traveling
somewhere new to offset part of the cost.

To justify an auction, there has to be at least one
"gottahave" item locally, five of them for an
hour's drive, and twenty of them for a major event.

Which is why live or online previews are crucial.

As is continuous and extreme research to pin down
the more obscure auctions unlikely to be well
attended or overbid.

If you are careful about religiously seeking out
a 30:1 sell buy ratio, then spending $200 at a local
auction for a possible $6,000 return makes sense
if you attend at least one auction every two weeks.

On the other hand, spending $2000 at an industrial
distress auction for a possible $60,000 return need
only be done twice or so per year.

The best solution, of course, is a mix of routine
low return auctions and the occasional one where
everything goes just right and you hit it big.

I'd be hesitant to spend more than $12,000 at any
one auction because that is a lot of money to risk
that could go wrong in many different ways. Not
the least of which is grossly overvaluing what you are
bidding on, its condition, and its marketability.

Plus, the chances are that similar auctions will happen
in the future that will let you widely spread your risk.

If you win more than five percent of your bids, you are
Paying waaay to much. Naturally, you make up for this
by bidding on twenty times the lots you could possibly use.

More on live auctions in this tutorial.

The evidence that all is not well with military surplus
continues to pile up. Government Liquidation offerings
in certain states ( especially Arizona and New Mexico )
are now down to useless dregs. The bottom of the barrel
is clearly being scraped. Current volumes would obviously
seem far below their operating costs.

When all this settles down, will their be a huge bubble
opportunity of way too much stuff dumped all at once?
Who, if anyone, will be offering surplus items? Or is this
the end of mil surplus bargains once and for all?

As mentioned before, we have backed way off on
mil surplus because of price runups, arrogant site
managers, infuriating bid extensions, and site
security issues. Our present focus is mostly on
industrial distress auctions.

Such as these. Your own custom eBay supply source
finder can be created for you from this service.

Continuing yesterday's guidelines on eBay item
storage: Several rooms of our house are dedicated
to shipping and immediate storage of smaller items.
A shop area is reserved for refurb.

A few huge items are occasionally stashed in the
driveway or beside our spa area. We try to avoid
these unless an exceptional opportunity arises.

A $45,000 doghouse in the back yard (actually a
photo processing lab bought from the feds for $175)
holds useful and convenient storage of popular and
mid-sized items.

But our main storage consists of five rental units a
block or two away. Right now, our focus is on making
our available storage much more efficient through
better inventory control and heavier triage.

There is room for a steel building in the back yard,
but the present dollar volume does not quite justify
it. And, as others continue to conclusively prove,
eBay does not scale worth a damn.

There is an optimal size for any eBay business.
Try to exceed it at your peril. Several of the largest
eBay sellers just got done by greed and excessive
fixed cost expansion.

Additional tutorials on our Auction Help library page.
Including this eBay selling tutorial and this eBay
buying tutorial.

There's a lively discussion on what you need for
eBay storage solutions over on the alt.marketing.
online.ebay newsgroup. I thought we might start
a list of guidelines that work for us here...

~ Avoid any and all long term future obligations.
     Rent storage by the month, not the year.

 ~ Immediately triage incoming items, converting
     lots of big items into fewer smaller ones.

  ~ Avoid placing anything in storge that has not
     already paid for itself. Seek out a 28 day cash
     out and a 15 month hang time.

   ~ Collect bargain shelving and oddball cabinets
      from local auctions. But never pay more than
      $2.50 each. Use these to organize and make
      your storage more efficient.

  ~ Seek out alternate storage solutions, such as
     a rentable house or excess garage space or
     a distress real estate location.

  ~ Always handle items with extreme care. Most
     especially when loading and unloading.

  ~ Agressively keep accurate inventory records.
     Never store anything that you do not know
     what it is or when you are going to sell it.

  ~ Have an Alvin Pile where you can immediately
     flush anything that is not working for you.
     Preferably getting paid in the process.

   ~Have different storage areas for immediate
     inventory, long term winners, and stuff that
     needs further processing.

   ~ Chances are a forklift can be rented from a
      nearby lumberyard or machine shop by the hour
      if you only occasionally need one.

   ~ Clutter will eat you alive the instant you turn
      your back on it. Continuously maintain neat and
      orderly work areas.

  ~  Be sure to use theTOP HALF of any storage area.
     This can cut your costs by two.

  ~ If you haven't dealt with it in fifteen months, you
     ain'g gonna. Flush it NOW.

  ~ Don't store anything that you are going to end
     up throwing away two years from now anyway.

  ~ If you touch an item that has been in storage
     deal with it NOW. Never set it aside.

   ~ Treat items that will need refurb or repair
      seperately from routinely sellable new items.

  ~ A handtruck and a dolly or two can be enormously
     useful. But avoid buying new or paying list price for
     them.

   ~ Metal storage buildings ($15K) or shipping
     cotainers ($1700) are not all that expensive and
     quickly can pay for retail rentals. Available land
     and zoning permitting. But long term sales volume
     has to be justifiable. Perhaps 100K in yearly sales to
     justify a 10K investment.

   ~ Avoid selling anything that you cannot hold
      extended at arm's length.

   ~ Shit floats to the top. If you find yourself forever
      moving an item to get at what is underneath,
      flush it.

   ~ The sooner you get rid of useless trash, the
      better.

  ~ Continuously review what you have stored where. 

We'll probably be adding to this list over the next
few days, so be sure to check back.

Instant and efficient hot water involve conflicting
goals. It takes a ridiculous amount of current to
rapidly heat water, which in turn puts outrageous
demands on needed wiring and peak loads.

As a result, many individuals will end up sorely
disappointed with hot-water-on-demand systems.
Especially direct resistance heated ones.

Secrets of efficient hot water are: (1) do not
let the water cool between heating and use;
(2) do not heat the water any more than its
needed temperature. (3) use low exergy energy
such as gas or a heat pump, (5) Place the heating
and use areas as close as possible. (6) Start with
room temperature water rather than winter ground
temperature water. (7) Preheat and prestore at
least some hot water for initial use. And (8) use
cogeneration where the heated water is an otherwise
wasted byproduct of HVAC systems.

Here's the fundamental math involved:

One BTU is the amount of energy needed to
raise the temperature of one pound of water by
one degree. One BTU is also 0.2931 or roughly
one third of a watt hour.

There are eight pounds of water per gallon,
so eight BTU per heated degree. Room
temperature water would need something
like a 50 degree rise, and outside supply
water as much as a 90 degree rise or higher.
Thus as much as 90 x 8 = 720 BTU's would
be needed per gallon of water. Or 210 watt
hours of energy.

If a gallon of water was to be heated in
an hour, then a reasonable 210 watts of
power would be needed. Heating the gallon
in a minute would need 60 times as much
or 12,600 watts. Or a 100 amp service at
110 volts. And faster heating rates would
quickly become utterly outrageous.

Typical home systems providing more useful
hot water rates typically demand exclusive
use of two 40 amp circuits. Whose provision
and upgrading costs have to be included in
your analysis.

Looking to the future, this might end up a killer
ap for supercapacitors. Where the energy
is accumulated at a sane rate and then
delivered at an insane one. Thus load levelling
and dramatically reducing peak powers and
installation costs.

Despite statements from others that "books will
last forever, I strongly feel that the utter and total
demise of books is suddenly going to happen a lot
faster than most everybody expects.

Driven by the near infinite superiority on all counts
of electronically distributed media. And aided by
such factors as the iPod phenomena, a student revolt
aginst backpacks, and routine acceptance of lousy
ergonometrics by cellphone users. Arrival of a "perfect"
book reader or its equivalent is only a matter of weeks
away at the very most.

All of which, I believe, slams the door completely
on Book-on-Demand publishing. Which I once felt
very strong about and was very much into.

Links to the remaining BOD faithful can be found
here. But it seems to me that the BOD window of
opportunity is long since gone. And, at best, was
simply an illusion.

The real killer of BOD, though was that nobody
was willing to step up with a suitable collate-jog-
bind-trim $600 device essential to BOD success.
Those few machines available typically cost several
hundreds of thousands of dollars, used old technology
designed and sold in old ways, and remained basically
unreliable, high maintenance, and hard to use.

Not to mention outrageously high per-book
amortization costs. Or their need to continuously
run at 100 percent capacity.

Possibly some really innovative solutions could
have happened, such as using a waterknife for
trimming and some sort of a pressure clamping
binding system. As might some sort of a hybrid
system where the cover had a chip in it with
fully searchible and linkable content.

Our new mail server seems to be up and running
and working well. Improvemnts include the spam
trapping of the 500 or so spams we get each day,
better accepted names alias control, and better
remote access.

We have a particular problem of needing two
workstations to independently access the same
messages without interaction. This is solved by
leaving messages on the host and cloning a
second mail file that autocopies the first.

Once again, some messages may have been
missed or lost, and our spam filter may be
too aggressive. So be sure to email us if
we seem slow responding to recent email.

October 14, 2006

As I mentioned before, I am very much a Netflix
fan. Just revisited the Ken Burns Civil War series.
Where we discover that McClellan was by far the best
general the South had. And that the South won the final
battle of the civil war!

Students these days will want to be sure not to miss the
final disk so they can find out how it all turns out. In the
seldom viewed alternate ending, Grant shows up drunk
at Appomattox during his surrender to Lee.

Comparing the Original to the New Hitchhikers Guide
vids is not even close. The original is infinitely better
on all counts. Especially the cheezy special effects.
Note that there is a separate bonus disk for the original
tv series.

One of my favorite vids of all time ( right up there with
Godzilla versus the Night Nurses ) was Putney Swope.
Which is not nearly as totally outrageous as it once was
but remains eminently viewable. Curiously, the new
comments correct a myth about the original: The
helicopter guy was a random Mel Brooks, but not the
Mel Brooks. And Putney himself could not remember
lines, so the black is voiceovered by Robert Downey.

Netflix has an amazing number of listings and lets you
reserve titles far ahead of their actual availability.
A few rare and ancient comedies apparently have not
yet made it to DVD. Two I'd like to resee are Make
Mine Mink and Our Man in Havana.

Two of my favorite available videos remain This is
Spinal Tap. Which is a rock documentary about
England's loudest band, once called the New Originals.
Besides being noted for their punctuality, their Heavy
Metal Memories album went pewter.

Plus, of course, The Adventures of Buckaroo Banzai.
Which is best categorized as a medical racing science
fiction horror rock documentary love story comedy
action adventure. And is unquestionably the finest
movie of its type ever made.

"No matter where you go, there you are".
Apparently they actually made this film on purpose.

Two other gems remain The Mouse that Roared
and Safety Last. The latter of which is of particular
interest to cavers and climbers. Where the only safety
net was a small matress on a smaller ledge high on a
building. When they tested the matress with a dummy,
the dummy bounced off and flipped seven stories down
into traffic. All stunts were real! And done with missing
fingers! But they did cheat a little on camera angles.
Be patient with this flick. It starts out rather dull and
tedious.

And, of course, it includes THE clock scene.

We are upgrading and moving to a new mail server.

As a result, there may have been some mail missed
or unanswered in the last few days. Please try
contacting us again at don@tinaja.com if you are
experiencing any problems.

Google Maps apparently has been quietly replacing
much of their older satellite imagry with newer and
higher resolution aerial photos. I've paricularly noticed
dramatic improvements in the area north of Pittsburgh,
PA.

I also happened to notice that the lumps of coal that I
found in the woods below my grade school many moons
ago are now the world's smallest strip mine.

What is really exciting is where Google Maps, Google
Earth, and others are going with online mapping. The biggie
cavers are waiting for is stereo to find sinkholes. Resolution
of a tenth meter should be well within reach.

As should variable spectra multithresholds in which you can
false color blend ultraviolet, blue, green, red, and infrared
spectra for such uses as crop health investigation and such.

Plus a variable mix of topo and geopolitical data. Such as
this crude demo I did a ways back of Pacman, Arizona.

Similar goings-on in our Fonts & Images library.

I've always been fascinated by math models and
statistical analysis. Even if the model does not
accurately describe the underlying situation.

Someone complained that they went to an auction
and did not get anything. What are the odds of this
happening?

Assume that there are 40 bidders and 1000 lots.
Assume further that the someone did a proper
preview, there are no shilling games, and that all
other bidders are equally motivated.

We could expect something like a Modified Raleigh
distribution of outcomes, with the someone scoring
on 25 lots being the peak. On any one lot, their odds
of a miss are 39/40 or 0.975.

On 40 lots, you would expect "one time constant" or
1/e odds of (0.975)^40 = 0.36233 to miss or about 2:1
odds of a hit.

On 1000 lots, your odds of not getting anything at your
auction would be (0.975)^1000 or 1.01071e-11.

Or one in 100 billion or so. aka "ain't gonna happen".

Similar topics in our Math Stuff and Tech Musings
and GuruGram libraries.

We are repeatedly asked about our NO FOREIGN 
SALES policy. First and foremost, that is what we
offer. If you do not like it, leave.

No exceptions. Ever.

Over the years we have carefully experimented with
foreign sales and have consistently and repeatedly
found that foreign sales are not remotely worth the
time and effort of seeking them out.

If you are a success with domestic sales, you do not
need any foreign ones. If you are not, then foreign
sales will not help you in any manner.

Some specifics...

~  Foreign sales invite ripoffs and scams.
~  Shipping costs are higher.
~  Fulfillment takes much longer.
~  Closure takes ridiculously longer.
~  Delivery times are a joke
~  Theft is endemic.
~  There are stiff fees for currency conversions.
~  Our suppliers contractually forbid such sales.
~  Shipping damage is more likely
~  Honoring guarantees becomes cost prohibitive.
~  Keytowing to bureaucracy NEVER pays.
~  Communications barriers create problems
~  Lying on green slips is an international crime.

And most of all...

~  Certain countries ( most especially Canada )
    are CERTAIN to cause you grief.

More in our Auction Help library page.

The plot thickens till it clots. Curiouser and
curiouser. One of the apparent reasons for the
Government Liquidation lack of useful stock
is that DRMS is going through a "cost cutting"
measure that seems to be equal measures of
tragedy and farce.

Emerging details can be found by searching on
MEO for "most efficient organization" and the
A76 Competition.

Apparently dozens of DRMS warehouses will
be or have been closed, including Holloman and
Kirtland. Davis-Monthan seems to remain
open.

The bottom line seems to be this: There is
now a major ongoing glitch in military surplus
availability that may or may not be terminal.
How much of what role Government Liquidation
will play in whatever emerges remains unclear.

There seems to be a bunch of "new" Arizona
auctioneers and auctioneer wannabees. I've
just added Amazing Auctioneers and a few others
to our Arizona Auction Resources listing. Along
with Proxibid.

Your own custom regional auction finder can
be created for you per these guidelines.

Additional live and eBay auction info here.

Refurb Log:

Had lots of unreasonable difficulty finding some
nuts for a lot of bulk miniature phone jacks picked
up during our recent electronic distributor buy.
And offered on our eBay store
.
Perhaps a general review of nutedness and boltedness
is in order.

There are two basic types of threads, English and
Metric. A comparison of which can be found here.

Classic electronics used English with sizes #4-40 and
#6-32 for most component mounting, with the oversize
binder head being very popular. Sturdier stuff used
#8-32 and #10-32. Note that #10-32 is an example of
a "fine" thread. Some hardware stores may still stock
"stove bolts" that have a less used #10-24 "standard"
thread. These nuts are sometimes square.

The typical "very small" threads were #2-56. Sources
of even smaller bolts and nuts include J.L. Morris and
various advertisers in Model Railroader magazine.

Good sources for regular sized nuts and bolts include
Small Parts and McMaster Carr. Your own local
hardware store, auto parts store, or Home Depot may
surprise you with their availablity. But the quality may
be low, the sizes larger, materials and finish choices
limited, and the costs outrageous.

Over on the metric side of the fence, threads are normally
expressed as #6-1.0 or similar. Meaning a six millimeter
diameter and a one millimeter thread pitch. A 6 mm
bolt or nut is very close to a quarter inch size, but, of
course, is not thread compatible. A "normal" or
"corase" threaded metric nut might be #6-1.0, while a
normal "fine" threaded one might be #6-0.75.

Similarly, a #M5- is a little bigger than an English #10,
a #M4- is slightly smaller than an English #8-32 and a
#M3.5 is pretty close to an English #6-32.

It ended up that the submin phono jacks used "none
of the above" threads, being an ultra fine #M6-0.5 for
the miniature jack and #M4-0.5 for the subminiature one.
Turns out these enormously difficult to find sizes are
stocked by Mouser as part numbers 48KN004 and 48KN006
on catlog page 1479.

There seems to be some further evidence bolstering
my feeling that all is not well with Government
Liquidation. As we've seen, Arizona and New Mexico
listings are dramatically down. Typically being less
than a dozen scrap items compared to the normal
hundreds of useful ones.

This GAO Report has some very negative things to
say. At present, they are complaining that they are
"only" getting twelve cents on the dollar. What they
missed out on is that before GL, they were literally
paying people to haul surplus away due to outrageous
admin costs. Something like $1.64 admin for each
$1.00 in sales.

And this new Internet Ripoff Report also seems to
suggest somewhat less than stellar vibes.

One of the persistent myths of eBay is that
"whatever works for Wal-Mart should also work
for eBay.". In reality, the success strategies are
pretty much exact opposites.

eBay is inherently a low volume, high fulfillment
cost, highly customized, unpredictable, extreme
value added, non real time, erratic, self-competitive,
and widly non-scalable sales venue.

Some of the eBay rules that work for me are
found in our eBay Selling Secrets tutorial.

First and foremost is maximizing your personal
value added by offering unique products that you
have specialized expertise in. And feel very strongly
about. Items that are not available elsewhere.

Second is a decent sell/buy ratio. Most eBay sellers
try to squeak by on ludicrously low margins. My
mininum recommendation is to always seek out a
30:1 or higher SBR.

Suitable supply sources are found here.

Third is minimizing your fulfillment costs by
absolutely no foreign sales combined with a
strict policy of VISA/MC/Paypal only.

Fourth is NEVER drop shipping and NEVER buying
wholesale. Nor selliing anything you do not personally
own and have posession of. Nor anyting that you cannot
hold extended at arm's length. Nor any item with excessive
scam potential, such as ANY consumer electronics.

Fifth is optimizing your time return by initially
listing at higher than expected prices. Always
aim for a 28 day payback and a 15 month hang
time.

Sixth is keeping absolute control of all of your
costs. If you are not including your pro rated
water bill in your cost accounting, chances are
overwhelming that you are running at a net loss.

Always know your Leavenworth Ratio, or how
much net net net you are getting per hour compared
to prison labor rates. Minimum wage, of course, is
far beyond the pale for the majority of eBay sellers.

Seventh is proper image post processing. A mininum
of two hours per image is recommended, because
few other eBay activities can generate such a high
return for time spent.

Much more on our Auction Help library page.

October 4, 2006

Added R&S Auctions to our Arizona Auctions
Resource listing. This was the last major holdout of
the larger AZ auctioneers to create a web presence.

Just completed a new Northern Ohio Auctions
Resource listing. Doing so turned into quite a trip.
Apparently every man, woman, child, and Holstein
cow in Ohio is an auctioneer.

And most of them seem to have bad hair. Not a problem
in Arizona where nearly all auctioneers ( male and female )
dress western. and all wear cowboy hats. One sight to
behold was Good Ole Boy Bruce Tingle dressed in his
classic western garb while modeling a mink stole.

Or then again, maybe not.

There really is an Ohio auctioneer named Shilling. And
another named Marks. The obvious thing for them to
do is to merge into the Shilling Marks Auction House.

More on which here.

Many auctioneers still do not even have email addresses
and consider the web to be a threat rather than an opportunity.
eBay, of course, is beyond the pale. The general quality of
many auctioneer websites remains mesmerizingly awful.

I purposely limited my Ohio listings to auctioneers that
had real and working web sites. I further required that
combined realtors and auction houses had to have a
strong presence on the "things and stuff" side rather
than on the "land and houses" side.

Some of the resources I rejected can be found on the
AuctionZip, National Auctioneers Association, and Ohio
Auctioneers Association web sites.

Additional auction help can be found here.
A Custom Auction finder can be created for you here.

One of the rules of thumb from back in my aerospace
days was that the hardest system improvement to
ever accomplish is three decibels.

Chances are the you can make minor performance
upgrades by twisting and tweaking. Or revolutionary
ones by resetting to zero and applying new technology
and concepts. But improving an existing system by
40 to 50 percent can be outrageously difficult.

The NIMH battery folks did just that a year or two
back when they discovered they could stabilize
reactions that used three outer shell electrons
for conduction rather than just two. As a result,
today's AA cells completely blow away yesterday"s
sub C units.

In the September 22nd, 2006 issue of Science Magazine,
such a "3 db improvement" to supercapacitors may just
have been made. It turns out that by going to extremely
small carbon pore sizes, there is an unexpected increase
in capacitance (and consequential energy density) by
fifty percent or so. Independent of area.

J. Chimola... Anomalous Increase in Carbon Capacitance
and Pore Sizes less than 1 Nanometer. Pages 1760-63.

Discovered an alternate free zip code finder that
looks quite useful.

As we'vealready seen in our Energy Fundamentals
and elsewhere, not one net watthour of conventional
silicon pv electricicity have ever been generated.
Nor is any ever likely using conventional silicon pv.

True renewability and sustainability will depend on
emerging CIGS and Quantum Dots technologies.

We can readily identify what has to happen for
pv solar to ever become a truly viable net energy
source: CIGS or better technology. Quantum dots.
Multiple workfunctions. Roll to roll processing.
Sane temperature and pressure manufacturing.
Minimum use of low cost and lower tech materials.
Nothing organic or polycrystalline. No concentration
or tracking. Unbox, plug-and-go fully integrated
systems sold out in Aisle 13 of Home Depot.

There's apparently a new candidate we might call
ZMT Technology, short for Zinc-Manganeese-Tellurium.
Like CIGS, it provides multiple workfunctions to raise
efficiency.

So far, it looks like unproven hype. Government labs
are always suspect, because they have accomplished
absolutely nothing in the last half century. But it just
may be a step towards resolving the multi workfunction
issues that are likely crucial to decent pv efficiency.

I am very much a Netflix fan. Most of our local
Mom & Pop video rental stores have gone belly
up. And 75 seconds in a Blockbuster store was
enough to convince me that those arrogant misfits
need staked to an anthill.

I rarely go to theaters, with Cars being my most
recent visit. I still marvel at how they absolutely
nailed old route 66. I've been to Radiator Springs!

Apparently others feel the same way about moviegoing
because one of our two multiplex theaters just folded
into becoming a truy bizrre church.

It seems to me the very survival of movie theaters
depends strongly on the instant availability of
direct digital projection. Yet the theater owners and
the movie distributors are pissing around arguing
who is going to pay for the upgrade.

In reality, if the movie distributors paid ALL of the
upgrade costs, they still would be ridiculously ahead
on print and shipping costs.

Not to mention being able to stay in business.

We now have outstanding but pricey home video
routinely available. Within a very few months,
further price drops and performance improvements
can be expected to completely and utterly blow
away the movie theater experience.

Caused by the anytime start - fast forward - pause -
view again - contol volume on one side compared
against the squawling baby -obnoxious cellphone user -
blabbing patrons - $12 popcorn - sticky footedness -
incompetent projectionists of conventional theaters.

September 29, 2006

Several alt.marketing.online.ebay posters have asked about Bruno.

Bruno is the AMOE attitude relateralization facillitator. One of
his specialties is reconfiguring top posters. In a related endeavor,
Bruno is also a product durability tester for a major New Jersey
baseball bat manufacturer.

Bruno also does trucking for Norfolk & Waay. Who, in a Kilgore
Trout sort of manner are NOT me and NOT my website. I have
only the vaguest clue who their webmaster is.

There seems to be quite a bit of confusion over
exactly what either regular or super capacitors can do
in the way of energy storage for vehicles and such.

The bottom lines are (1) while the power density of
the latest capacitor developments are becoming
impressive, their energy density remains uselessly
bad. And (2) A capacitor behaves wildly different than
a battery when it comes to charging or discharging of
its internal energy.

A pure capacitor obeys the law i = c (dv/dt) or its
integration of v = (1/c) integral i (dt). Capacitors
MUST be strictly current driven. Attempting to
instantaneously apply a voltage to a capacitor will
cause an infinite or near infinite current and a
possible explosion.

All a capacitor knows about is whether there is
an incoming or an outgoing current that will alter
its internal charge state. Terminal voltage is
the result of the current integrated charge state.

The energy storage of a capacitor is given by
w = (1/2) Cv^2. Thus the stored energy of a
capacitor goes up with the square of its
terminal voltage. The stored energy is measured
in Joules where a Joule is one watt second.

There are 3,600,000 Joules in a kiliowatt hour.
Meaning that today's capacitors cannot remotely
approach the energy demands of long distance
personal vehicle use. But they certainly are of
interest for a few seconds of passing power or
regenerative braking storage.

Another misconception is that capacitors and
batteries are in some manner similar. Most
batteries maintain a pretty much constant
voltage as electrochemical conversions take
place during charging and discharging.

Plain and simply, capacitors do not regulate
worth a damn.

With capacitors, instead of a constant terminal
voltage, the stored charge is proportional to the
square of the terminal voltge. Or the voltage
goes up as the square root of the stored charge.
A full or an empty capacitor will thus have wildly
different terminal voltages.

Any attempt to charge a capacitor from a resistive
source or an ordinary generator will result
in one half or more of the energy being lost inside
the generator's internal resistance. Efficient charging
and discharging of capacitors requires a switchmode
scheme where voltage is transformed to an inductor's
current and the inductor's current is then repeatedly
sent to the capacitor. One name for such a circuit is a
transimpedance converter.

Transimpedance converters must obviously work over a
very wide voltge range. A 2:1 range converter can only
use 75% of the possible stored charge. A 3:1 might
approach 89% of stored charge. And that would be for
ONE optimal input or output level. Any real world use
would call for dramatically wider control ranges.

Any capacitor (and especially a supercap) has a definite
maximum terminal voltage that must never be exceeded.
Traditionally, good engineering practice leaves at least
a 50% margin between normal system voltages and the
allowable maximum.

It is the greasy whistle that gets squeaked.

Some rumor mongering on the Arizona Auction Scene:

After forty years or so of threatening to retire, Busby Auction
might be getting serious about actually doing so. At any rate,
they are both away on a two year foreign mission. There are
several "second tier" Gila Valley local auctioneers, some
licensed and some not. I expect "real" auctioneer Bruce Tingle
will take up most of the slack from out of Willcox.

The once awful parking situation at Sierra Auction has
apparently gotten a lot worse. Irate business owners
nearby have been known to stick safety pins in tires and
pull similar nasties. Sierra has also recently expanded,
opening a Tucson yard with monthly sales.

Auction Advantage seem to have closed their Pomerane
auction house yard. They may be doing a cooperative
on-site venture with Troy Wilde out of Yuma as part of
Desert Auctions USA. There are at least seven auction
houses competing in the marginal Benson, AZ area.

Cunningham continues to emphasize real estate, but
is apparently still doing occasional industrial auctions
I have done quite well with their previous offerings, and
the deals that got away have been even more spectacular..
Any Cunningham auction house sales are done through
Auction & Appraise.

Government Liquidation seems to have sharply cut
back on their Arizona and New Mexico government
surplus auctions. A few scrap items seem to have
replaced their hundreds of normal listings. While they
have been highly useful to me in the past, I've pretty
much backed off on using them. Owing to prices ending
up insanely too high, their maddeningly infuriating
automatic bid extensions, hostile site managers, too
high opening bids, and increased base security hassles.

I've added some newer and more distant auctioneers
to our Arizona Auction Resources page. These include
Pot O Gold, Desert Auctions USA, Fusco and Battermans.

Plus a new ( but so far largely useless ) Classified Ad Network .
Note that auctions can show up either in the announcements
or for sale categories, so two searches may be needed.

A general tutorial on the Live Auction Scene appears here,
Additional resources are on our Auction Help library page.
Your own local custom auction finder can be created for you
per these guidelines.

What are the fundamental underlying "secret"
forces driving recent technological developments?

How can you identify and tap these for your own
uses? I'm probably going to work up a GuruGram
on this, but first thought I might develop some of
the concepts here to get your feedback.

We've already seen decoupling as a major factor
and possibly the single most important biggie.
Other tenative factors might be...

  1. Decoupling.
  2. Accurate replication.
  3. Bypassing the gatekeepers.
  4. Seamless pan and zoom
  5. Virtualosity.
  6. Nonlinearity.
  7. Superimposition
  8. Programmability
  9. 24/7 availability
10. Devaluation.
11. Nanoification.
12. Disembodiment.
13. Increasing competitiveness.
13. Indirect payback.

On to number two...

With accurate replication, the end user gains the
ability to make copies of items or IP that are as
good or better than those from traditional sources.

People once went to a commercial printer because they
could not produce their own copies that were as good
or as cheap by themselves.

Authors once used book publishers because typesetting and
artwork creation were unbearably difficult, time consuming,
and second rate using previously available tools.

Musicians once used music companies because today's
cheap quality mixing and editing was once impossibly
expensive. And that distribution required a physical media
presence.

"Perfect" copies of video or music by either the creator or
the end user are now trivial and clearly eliminate the surface
noise, flutter and wow of previous generations. Generation
loss is now gone.

One obvious consequence of cheap and fast perfect copies is
that anyone overcharging for their product will cause new
copies to be made instead by the end user. Thus, the price
gougers all loudly scream "piracy". Rather than realizing
that additional copies would likely not be paid for anyway
and probably increase the product awareness and demand.

Traditional copyright and IP protection has thus largely
become totally useless.

The biggie waiting in the wings, of course, is the Santa Claus
Machine. For instant replication of real world parts.

More to follow. Please email me with your input.

This MySpace theft of IP and bandwidth may be
becoming endemic. I've now got a second image
being stolen. Like the original, it makes no sense
whatsoever to use as teenybopper wallpaper. But
at least it is only a 33K JPEG file instead of a 9
Meg bitmap. It still uses up our ISP traffic and
fouls up our log stats.

I don't know what the solution is. MySpace refuses
to take any action without costly, draconian, and
ambiguous response of the rights owner. Clearly
they are trying to make the rights owner the cause
of the problem.

The obvious solution of replacing the image with
one of a suitably clad individual pioneering new
methods of animal husbandry might have the exact
opposite of the intended effect. And increase rather
than decrease any demand.

This also would piss off the legitimate original users
of the imagry.

Added two new links to our Arizona Auction Resources
page. A guided tour of Arizona Auctions appears here,
and a general tutorial on the live auction scene here.

Your own custom local auction finder can be created
for you following the details here.

Super multi-day mega auctions have their own sets
of major disadvantages...

~ They are nationally promoted, heavily advertised,
   and attract a huge pool of bidders.

~ Bidders may be bigger spenders as they may
   have committed to hundreds or thousands of
   dollars in airfare and hotel bills.

~ Pricing rarely drops below $25 per item.

~ Extreme focus is needed to target your items
   of interest. Energy levels MUST be carefully
   conserved.

~ Auctioneers may get ahead or behind, meaning
    that your lots may not be on the expected day.

~ Professional auctioneers may be enormously
    difficult to understand and may pull all sorts
    of secret insider stuff. You must listen very
    carefully and respond clearly and quickly.

 ~ Online bidders are at a huge disadvantage due
    to comm blowups, time delays, and any sudden
    auctioneer changes in lot combinations, groupings,
    and payment options.

 ~ Mega auctions in your area of interest are often
   few and far between.

~ Ordinary items in sane quantities (especially anything
    tool or machine shop related) are likely to sell for
    outrageously high prices.

~ A local auctioneer is much more likely to cut you a
   personal deal. Or at least favor taking your bid.
   Because they want you back.

~ Previewing and planning can be overwhelming. Especially
   with a sitdown or theater style auction.

~ Your own travel and trucking and rigging and casual
    help may become major and high risk considerations.

~ Some "professional" bidders delight in pissing contests
   and will not let you outbid them under any circumstances.

~ Many bidders will be spending company money rather
    than their own.

~ If you are attracted to the auction, chances are that many
   others also will be.

~ Quality of product may differ wildly depending on
    which area is being offered. New inventory versus
    repairs versus obsolete junk.

Plus these benefits...

~ If there are 9000 lots and only 1000 bidders,
    each bidder scores 9 lots on the average.

~ Enormously huge bargains can result IF you
   can triage obscure multi-pallet lots.

~ Careful and repeated inspection can reveal hidden
    values that the others may miss.

~ Many items will be high up on murky shelving
   that others may miss carefully inspecting.

~ Items are often grouped into "contents of room"
   or "contents of cabinet" to keep the total number
   of lots down. Knowing exactly what is in the room
   or cabinet can give you a tremendous edge.

~ Bunches of stuff will fall through the cracks
   as not everyone can possibly notice everything.

~ Some larger lots will run out of bidders before they
    run out of product. Meaning that the dregs may
    go for a pittance.

~ Only a few lots will be of specific interest. It does
    not matter how much the others go for.

~ Side deals can often be cut on "pass it" items
   gotten for $10 each. As can poisioned lots where
   a trash lot is combined with a good lot to drive
   away bidders.

~ Auctions typically run very late. If everybody
   else runs out of interest and energy before you
   do, outstanding opportunities can abound.

You can find these mega auctions on our Auction Help page.

A Custom Auction finder can be found for you here.

September 21, 2006

I recently turned down some consulting work to help
develop a flywheel-supercap-regenerative vehicle with
a 300 mile range. I see at least four major problems:

~ Lithium batteries have a much better energy
   density than flywheels. And much greater
   potential for future improvement. Without any
   of the gyro, safety, costs, and exotic materials
   of flywheels. Another problem is that huge
   and cost ineffective motors are needed if the
   windup time is to be significantly less than the
   discharge time.

~ Supercaps are beginning to show a promising
   and useful power density, but supercap energy
   density remains ludicrous. Supercap energy
   is normally measured in watt seconds. There
   are 3,600,000 watt seconds in a kilowatt hour.
   Another problem is that supercaps do not
   regulate worth a damn, making recovering
   more than a fraction of their potential storage
   a serious challenge.

~ The potential of regenerative braking is wildly
    overestimated. As five minutes driving any
    vehicle with a retarder should dramatically
    prove. A reasonable estimate is that adding
    regeneration may add nine percent of range
    for a careful driver. And a sloppy driver will
    reduce their range more than regeneration
    can recover. Yes, if you already have onboard
    generation and storage anyway, then regen
    certainly should be cheaply added. But it is
    otherwise just plain not that big a deal.

~ Any new high tech product development is best
   linited to ONE "iffy" technology. Stacking
   three of them on top of each other dramatically
   increases the probability of failure.

Much more in our Energy Fundamentals tutorial.

Consulting services available.

Just witnessed what may or may not be a major improvement
in firefighting. A company by the name of Hazard Control
Technologies is selling new buckets of F-500 glop that appear
to simultaneously replace wetting agents, penetrants,
BOTH class A and class B foams, halogen suppressents,
and Noflash all at once. Without most of the costs and
complexity and training and fussiness of CAFS.

A SIX TIMES suppression over ordinary water is claimed.
No aspiration or air injection is needed. Product can be
delivered with any eductor or simply dumped in the tank.

The trick is a custom crafted molecule called a Michelle
Encapsulator. Basically, there are hooks on both ends of
the molecule, hydrophobic on one end and hydrophillic
on the other. Which has the potential to tightly grab and
bind hydrocarbon and water molecules.

This is somewhat similar to the fallen-from-grace halogen
suppressants. In additional to blocking oxygen and reducing
temperature on the fire triangle, these also can directly
interfere with the chemistry of fire.

One of the bad features of classic foams is that they also
are good insulators, so they drastically limit how
fast temperatures can drop.The most impressive feature
of F-500 is the sudden temperature drop. You can touch a
pile of burning tires five seconds after suppression.

Smoke also vanishes instantly. Mix is the usual one percent
for class A, three percent for gasoline, and six percent
for alcohol or polar solvents. Ordinary nozzles are used.
Pump pressure at the eductor has to be somewhat higher
than normal. The delivered product appears as a milky liquid
rather than a foam. Glop is claimed to be envinormentally benign
with a long shelf and mix life.

Cost is apparently a $30 premium over a normal $110
5 gallon bucket of classic AFFF foam.

Not addressed was whether the glop is as good as
AFFF for dealing with killer bees. Which is around
one half or more of most rural department foam use.

Is this a fundamental breakthru in firefighting or some
PR hype piled on top of a minor improvement in a
wetting agent?

Time will tell if this glop lives up to its demo promises.

Playing with some of the Springerville pv Solar numbers
seems to lead to some interesting results.

I read somewhere that this is the largest pv facility in
the US, and possibly (depending on how you rate a
German unit) in the world. Its effective peak capacity is
listed as 4 Megawatts. At any rate, there are very few
pv sites even remotely comparable in size or scope.

Next door is a "regular" sized coal power plant having
two 400 megawatt units online, one nearing completion,
and one planned. 1600 megawatts total eventually.

But the "exchange rate" between pv facilities at six
hours per day and coal facilities at 24 hours per day is
perhaps 4:1. So the largest pv site in the US is something
like 1/400th or 0.25 percent of a typical coal generation
facility.

One quarter of one percent.

A comparable solar site to match coal generation would
probably be 400x44= 17,600 acres or (at 640 acres
per square mile) about 27.5 square miles. Using up most
of the desert between the plant and Springerville.

Which appears to be a much larger area than a coal
plant and its mine combined.

There are 4047 square meters in an acre and 44 acres
on the site for a total of 178,068 square meters. Nominal
peak incoming energy is 1000 watts per square meter.
Dividing the peak 4 megawatts by the acres gives a peak
power production of about 23 watts per square meter.

Yielding an energy conversion efficiency of 2.3 percent or
so. This figure is low because not all of the total area is
effectively utilized for actual cell generation. We'll
note in passing that the latest of conventional power plants
are approaching 62 percent thermal efficiency.

Now it sure is nice looking at all those panels silently
giving the illusion of producing electricity. With no
apparently obvious pollutants or noise. And not even
any obvious employees. In a facility that can be rubber
stamp approved in ten minutes flat without any objections
from anybody whatsoever. And also can be incremently
and quickly expanded on any scale at any time.

The only tiny oint in the flyment, of course, is that pv
solar today is outrageously expensive and can in no
manner compete with traditional generation. It is
in no manner renewable nor sustainable when properly
full burden accounted. Such costs should truly reflect
ALL credits and subsidies, both obvious and hidden.

The entire site, of course, clearly remains a net energy
sink.

This picture may dramatically change when combined
breakthroughs in CIGS technology and Quantum
Dots attempt to bring us much closer to eventual
renewability and sustainability.

More in our Energy Fundamentals tutorial and in
our preveious What's New entries below.

Details on the Tucson Electric Power photovoltaic
Springerville solar website appear in a no longer
available website.

Its exact location was not obvious to me, and one
of the natives insisted that the pv panels were
really the Greer Snow Fences! Turns out the
facility is five miles due east of Richville.

You can start with Google Maps, and search on
Springerville, AZ, select hybrid, and go 12 miles due
North and then magnify. The panels become totally
obvious and leap out at you. ( The Greer Snow fences
are barely obvious 20 miles west of Springerville on
Route 260 .)

By the most astounding coincidence, you must drive
right by their pv solar panels on your only way to their
two landfills and three (soon to be four) coal fired plants.

Conspicuously absent in the website is their
fully burdened peaking avoided cost. But some
guesses could be made that suggest to me that
the site is not yet remotely cost competitive.

A reasonable cost for afternoon oil fired peaking
might be six cents per kilowatt hour. A recent
pv production day was 20,000 kilowatt hours or
$1200 if competitive. It is not clear whether the
20,000 kWh is at the panel terminals or net actual
grid delivery minus operating costs and conversion
efficiency. Nor whether this adjusts for days of
available sunshine.

At any rate, this site tells us that $1200 amortizes
something like a 2.7 million dollar investment at
ten percent over ten years. My best guess is that
the facility and its overhead and its amortization
cost a lot more than this. I'd estimate 46 cents per
kilowatt hour as their true burdened pv avoided cost.

There are something like 35,000 panels installed.
If these cost $100 each and everything else was free,
you would be way over competitive peaking costs.

Naturally, a R&D facility would not normally be
expected to be cost competitive short term.

Clearly, they are not aggressively expanding their
site for routine avoided cost peaking production.
In fact, some previously planned expansions have
apparently beed deferred. Possibly because of the
recent runup in pricing of conventional panels.

But the coal plant next door is being aggressively
expanded. Even with a somewhat questionalble future
coal supply.

More in our Energy Fundamentals tutorial.

What can you do when an image you are post processing
has areas burned way too far into the white? Nikons in
particular seem to have linearity problems at the extreme
white end of their dynamic range.

You can check these guidelines against the original photo,
the retouched bitmap, and the final .jpg...

~ Try reshooting the image without flash using
    available light from a skylight and bounce card
    or outdoors in the shade or under a light tent.

~ Try "multiple exposures" to dink with the gamma
    and brightness of problem areas using Imageview32,
    PhotoShop, or whatever. Trim to problem subjects
    and paste.

~ Try using our Dodge and Burn tools.

~ Optimize the appearance of the rest of the image by
    Swings and Tilts distortion correction and background
    knockout.

~ Try to make foreground items perfect by adding
   sharp outlines and sliding acceptable texture areas
   into overwhite ones.

~ Unify and darken the outline of all overwhite objects.

~ Subdivide all white items by top, sides, edge features.

~ Add credible line detail to burned areas.

~ Convert overly white areas into pastels of several
   different shades.

~ Try to neutralize large panel areas into a uniform
   and nondescript mottled gray.

~ Replace any burned white wire with darker colors,
    especially any striped ones.

~ Make up imaginary darker details when totally absent.

~ Pay special attention to areas where background panels
   or whatever shows through. Expecially with bunched wiring.

~ Simply remove the worst examples of anything lots of
   are present.

~ Take out excess detail, especially shiny and "sugary"
   boltheads.

~ Pay particular attention to darkening and unifying edges.

~ If all else fails, make an area so nondescript that the
   viewer looks elsewhere.

~ Remove "sugar" or "speckle" by cutting and pasting
   small nearby sample area.

~ Try repeating good objects by pasting them over any
    burned ones. Scaling or rotation may be needed.

~ "Chase an edge" by repeating a good sample over
    and over again into problem areas.

~ On burned out lettering, darken and expand into
    credible hieroglyphics . These should improve
    automatically on final size reduction.

~ Try to remove ALL remaining overly white pixels.

~ Hope that the final size and resolution reduction to
   .JPG will cover any problems remaining.

Seminars and custom consulting available.

Human cannoball during his circuis exit interview:
"Sorry to se you go. Hard to find a man of your calibre."

The hardware designer situation is even more bleak than
the a-beginner-can't-compute-on-a-computer situation we
looked at yesterday.

Kids became Silicon Valley superstars by first taking
apart old radios back when the Weller Soldering Gun
reigned supreme.

High schools and even middle schools had outstanding
electronics courses. All of which went away with the
aerospace dryup when the evil empire self-destructed.

Most community colleges cancelled their electronics
technology programs a decade ago. Often giving the
money to the football team instead.

There was ham radio where kids actually did stuff like
inventing parametric amplifiers and single sideband
and bouncing radar off the moon. Instead of being a
laughable geriatric parody of its one time greatness.

There were great electronics magazines, the foremost
and finest of which was the original Popular Electronics.
All long gone, with the possible niche exception of
Circuit Cellar.

And my, oh my, the kits. You could make stuff cheaply
that worked, saved you money, and outperformed the
real stuff on the market. Heathkit, of course, was the
biggie.

All long gone.

Meanwhile, a college student who works for us did not know
the name of the direction to the left when facing north.

May you live in interesting times.

Jeff Dunteman occasionally has topics I find of interest
in his ongoing blog. A recent entry made the observation
that today's crop of computers have no way whatsoever
for a rank beginner to simply jump in and start computing.

Even a three year old could proudly create a SYNTAX
ERROR on an Apple II. Several flavors of BASIC were
available for immediate use by casual beginners.

More importantly, the next step beyond BASIC was the
ability to list and inspect and disassemble programs. Tasks
no longer available with a few keystrokes and impossibly
more difficult with complex operating systems, multitasking,
compilation, and relocatable code modules.

Yes, you can go out and buy a copy of BASIC but nobody
seems to do so. Even more sadly, kids programming systems
like LOGO seem to be falling by the wayside. And general
purpose programming languages like PostScript seem to have
little beginner interest. Despite their ability to instantly give
an incredible range of wildly useful results.

We got to where we are by three year olds who could create
their own syntax errors and went on to bigger and better
things. It appears we are in the process of throwing the
babies away and drinking the washwater.

Picked up a super premium quality Pritchard 1880-C
combined dual grating Monochromator and precision
Photometer. These are apparently an optical spectrum
analyzer as they let you dial your own single color of
light to exceptionally narrow bandwidth. Either manual
or under swept motor control.

Apparently the intended use ( nobody else could afford
one! ) was for military evaluation of flight instrument
interaction with low level night vision devices.

This should have all sorts of interesting optical research
applications as we can offer it at a tiny fraction of its
original estimated cost of $85,000.00. So far we have
not tested it, but it appears to be in good condition and
should be capable of being brought up to original factory
specs. Includes precision lens and one extra filter.

email me if you have any interest in this unique optical
opportunity,

September 12, 2006

Had to once again lock out KNOCKOUT.BMP. You
can email me for your personal access.

MySpace users continue to insist on using this huge
and laughingly unsuitable raw bitmap as wildly inappropriate
(and stolen) wallpaper. Causing theft of ISP services
and ongoing problems with our log files.

Meanwhile, the clueless morons at MySpace themselves
make reporting ISP infractions a draconian and impossibly
dificult process. Involving sending sensitive information
over nonexistant channels. It is quite obvious that MySpace
has no intention whatsoever of respecting the intellectual
property of others.

Please, if you are a legit user of this sampler of image
backgrounds, make your own copy on your own website.

There's a small auction opportunity upcoming
that I just do not see how I can participate in. But if
you happen to have an 18 wheeler, some time, and live
in the site neighborhood, please email me or else call
(928) 428-4073 immediately.

This tiny auction squeaks by with barely over 9000 lots
and many of those are only a scant five to twenty pallets
each.

The term "adequate" comes to mind. They also have
an electric drill for sale.

One As we've seen in our Energy Fundamentals tutorial,
not one net watthour of conventional silicon pv electricity
has ever been produced. Nor is any ever likely owing to
the intractability and inefficiency and unsuitability of
conventionally processed silicon as a photovoltaic converter.

In fact, the historic price reductions of conventional silicon
pv panels have now been reversed, and are in a sharply
upward trend, owing to the disappearance of their former
scrap silicon supply.

To me, it makes no sense whatsoever today to try and use a
conventional silicon pv panel for home power generation.
When properly fully burdened accounted for on a total
cost of ownership basis, all you have done is placed an
outrageously expensive gasoline destroying net energy
sink on your roof. The longer you run it, the more gasoline
you destroy.

For today's panels are in no way either renewable or
sustainable. Even if your panel generates ten cents worth
of electricity per day at an amortization cost of ten cents,
per day, all you have is an elaborate "paint it green"
transfer mechanism of conventional energy. Only when
fully burdened costs drop significantly below conventional
does the possibility of renewability or sustainability even
arise.

At present, the amortization cost of the synchronous inverter
alone can consume more than 100 percent of the value of
the electricity run through it. And, of course, no other pv
storage system is remotely cost competitive with grid buyback.

The panels do see effective use for (1) solar calculators
where users gleefully pay $500 per kilowatt hour for their
electricity, (2) genuinely remote low power uses such as
mountaintop repeaters and distant livestock watering tanks,
(3) scammers stealing subsidy dollars and similar ripoffs,
and (4) scammers stealing government research funds.

Proof of this is that there is not one power utility
anyplace ever that is using conventional silicon pv
technology to provide avoided cost competitve peaking
power on a routine basis. All utility use of conventional
silicon pv to date has been has heavily subsidized R&D
or "Hey, we are green!" PR flack.

Fortunately, there are a pair of recent developments in
photovoltaic technology that might be able to get us as
far as ONE THIRD of the way towards genuine renewability
and sustainability. Possibly within a decade or two.

The first of these is called CIGS technology, short for
copper - indium - gallium - selenide. These reduce the
thickness needed for solar conversion by two orders
of magnitude ( 90:1 or so ) and can process in huge
sheets without the need for high temperatures or high
vacuum. And do so with comparable stability and
efficiencies of conventional panels.

Be sure to search on both "CIGS" and "solar" on
Google, or you obviously will get zillions of false hits.

The second major development involves quantum
dots, per this tutorial. In a conventional silicon pv
cell, a certain amount of iucoming energy is needed
to knock one electron loose. Any energy above is
converted into heat and is the cause of inherently
low conventional silicon pv solar efficiency.

With quantum dots, two or more electrons (up
to seven in theory!) can be converted from the
excess energy not used by the first conversion.
Which promises a possible doubling or tripling
of cell efficiencies.

Much more on our Its a Gas and GuruGram
library pages.

One of the projects that has fallen by the wayside
from way back in our PostScript Desktop Publishing
class was a resume that answered the question
"what if ONE single individual was responsible for
each and EVERY major industrial disaster over the
past few decades?"

This would be an obvious candidate for an onion
satire. Couched in words like "proven ability to
provide a commanding media presence", a
deeper investigation into his employment history
would include being an ethics officer for Enron,
honcho of the FEMA rapid intervention team,
did tanker safety training for Exxon, fire prevention
for Yellowstone Park, system monitoring supervisor
for Cherynoble, was a Union Carbide tank inspector in
India, ad director for New Coke, etc... Possibly even
starting out as an Edsel intern trainee.

What would your input be for several employers
of his and the related job titles? email me.

One trick that can increase the apparent resolution
of a digital camera: Crop very tightly to subject and
then paste into a larger background.

One way to do this is to load the camera image into
Paint, copy to clipboard, select Attributes to expand
image size, and then repaste original back.

Background knockout then proceeds either manually,
automatically, or with optonal Auto Vingetting.

I feel there are many very good reasons why you need
at least a five megapixel camera for your eBay photography.

These include...

~ Cropping throws away many if not most pixels.
    Even a modest one-eighth crop throws away
    nearly half (28/64ths) your available pixels.

~ Bayer filtering requires four camera pixels for
   one final pixel. As in red-green-green-blue.

~ Essential post processing steps such as rotation
   or distortion correction are best done at least 2X
   final size.

~ Lettering retouch and rework demands extreme
   resolution during correctrion

~ Processing algorithms in smaller cameras have
   ranged from terrible to ludicrous.

Naturally, if you really want resolution, you should use a
scanner instead. Or both at once. A 1200x1200 pixel per
inch scanner is equivalent to a 134 Megapixel or so camera.

Detailed free image post prep tutorials appear on our
Auction Help and Fonts & Images and GuruGram
library pages.

Cubic spline fits to a catenary raise a few interesting
gotchas, but appear quickly solvable. Here's the results
of a preliminary look...

If the catenary is symmetric, it is, well, symmetric.
This means it is made totally out of even functions
and the only even function in a cubic spline is the
second order one.

The cubic part never gets into play.

Thus, at least a fourth order polynomial is required
to accurately approximate a fully symmetric catenary.
All a cubic spline can do is use its second order term
which is a parabola. There is something around a
five percent worst case error with a unity normalized
a=1 catenary.

Which is the normal difference between comparable
y=x^2 parabolas and y = (a cosh(x/a))/2 catenaries.

But by going to a half-symmetric catenary, use of two
cubic spline fits of NUBEZ4PTS1.PDF at a=1 gives us what
appears to be a nearly perfect fit on the first pass, even
when magnified by 6400 percent. The slope error also
appears to be nearly zero at the center joint of the two
splines as well. As is the slope discontinuity.

I suspect this is easily proven if you dig deeply enough
into the underlying math. A catenary can be shown to
be the path of the focus of a rotating parabola.

The approximation remains quite good for a values
in excess of 0.5. Lower a values cause significant
errors. a=0.4 has around a five percent error, and
anything less is pretty much useless. These can be
minimized by adding additional four point fit splines.

I'll try to make a new GuruGram out of this if I get
the chance.

Termite walks into a bar, climbs up on stool.
"Say, is the bartender here?"

Picked up an older Adept 310 (aka 10000-310)
robotic controller. This is a huge 300 pound box
that includes all the high level drivers.

I'll probably offer it as is a time or two and then
strip it for replacement parts and such. The
whole unit is strictly FOB local pickup only.

Please email me if you have any interest one way
or another. This was a $6800 unit. 1600 hours.

Two random observations: Now that gasoline prices
are sky high, a lot of people leave their engines
idling at the post office. Very rare before.

There's also a distress auction of the Hip Chick
dress store. I don't suppose the choice of name
had anything to do with it. "Up Chuck" would have
been an infinitely better choice.

I still have not exactly fully isolated a bizarre
bug that seems to hang Thunderbird on one but
not both of our machines. I've decided to dig
down further just in case this malware can in
fact generally cause grief to Thunderbird.

The symptoms are as follows: A rare one of the
apparently "usual" obvious eBay phish scam
attempts will either hang Thunderbird outright
on initial loading or else cause an excessively
long load time. I have yet to separate whether
this is a blowout or a very long delay.

Since loading does not complete, any reset or
further attempted use of Thunderbird does the
same thing. I have access to our email sources
through Putty. By deleting the malware at its
source, Thunderbird will later load properly.

However, clicking on the remaining malware
message will once again hang Thunderbird.

The sledgehammer repair is to click on the
previous message, then shift click on the post
message, then delete all three. If previous and
post messages are important, they should first
be copied or saved.

The eBay phish message pretty much looks
like any other of its ilk. It is not unusually
long or otherwise attracts attention to itself.
Only one phish in a hundred or so seems to
present this particular problem.

The obvious thing to do is isolate the code and
then shorten it till only the malware portion
remains.

Your inputs on this are certainly welcome.

A reader requested a way to fit a single Bezier
Cubic Spline to a catenary. We already looked
at some tools to do this in NUBEZ4PTS1.PDF of
GuruGram #59.

A catenary is formed by hanging a uniformly
weighted chain. Per this source or this one. The
catenary is "almost" a parabola, but not quite
and has a form of y = a cosh (x/a).

True catenaries are relatively rare, as taking,
say, a suspension bridge cable. The cable is
pretty much a catenary when hung, but as soon
as you add the uniform bridge load to it, it becomes
more of an ordinary parabola.

We can easily and exactly fit a cubic spline to
a parabola. Just use the two end points and two
points a third of the way as input that you send
to NUBEZ4PTS1.PDF.

I strongly suspect a very close or possibly true
approximation to a catenary can be done by
similarly fitting the end and third position points.
Particularly when doing the fancier iterations
method of NUBEZ4PTS1.PDF. Whose acutual
working code can be found as IMBZ4P01.PSL

It would be reasonable to expect that a minimum
length (as found by the iterative method) Cubic
Spline is in fact a catenary. I'll try to prove this
one way or the other if I get a chance.

An older method of using two splines to fit a
1/x hyperbola appeared in HYSPLINE.PDF.