Don Lancaster's What's New? 2010 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	-----	-----	-----

Latest GuruGram #117 is on "Level II" Enhancement of
Pre-Cyber eBook Conversions

Its sourcecode can be found here and a use example here.
More GuruGrams here and free eBooks here.

Closed out the 2010 Archive and started the 2011 one.

Here's a useful technique to ease and speedup conversion
of bitmaps into their PostScript vector equivalents:

Temporarily scale the bitmap to exactly ten times the
number of pixels of vector width. Then flip it vertically
so that the numbers increase from the top.

Locations of all the PostScript vector elements can be
read by using the pointer for locations, and the box
select for widths. Mentally divide by ten to get your
new PostScript coordinates.

I'll try to work up a demo on this whenever.

Revised and updated our Auction Help library.
Mostly by removing some broken links out of the Arizona
section.

A review of Arizona Auction Resources can be found here.
You own custom auction finder can be created for you
per these details.

There are both powerful advantages and disadvantages to
doing a "Level II" ebook remastering by combining Acrobat
10 with my Gonzo Utilities.

As this new demo clearly shows, Level II is markedly superior
to results you would normally get by using Acrobat 10 alone.
A comparable Acrobat 10 only result appears here.

Files are much shorter and text quality is much higher. Far
fewer fonts are normally required. Full text searching is
provided and maintained through later rework. Image
appearance is much better, and, despite high compression,
can be arbitrarily magnified to any desired resolution without
degradation. Value-added features such as full color, web
links, hyphen elimination, emphasis substitution and such
are easily included. Fill justification is easily preserved
with only the most minimum of text rekeying. Halftone
images are greatly improved and can be made text searchible.

Turning to the disadvantages, if there is some historic, legal,
or IP reason an exact replica of a precyber source is needed,
then the method is largely inappropriate.

Secondly, it is tricky to do more than four or five "easy" Level II
rework pages per hour. Some pages ( especially electronic
schematics ) may take much longer. Thus the project has to be
reasonably profitable to justify the time and effort. Similar,
problems arise iif many new ebooks are needed in a short time.

Finally, the uniquely high quality of the Gonzo Utilities is based
on them being proudly non-WYSIWIG. If you are not into
"bare metal" machine language programming, then they might
not be at all suitable for you.

We do offer custom Level II rework for both current and
pre-cyber ebook projects, as well as training seminars. You
can email me for details or call (928) 428-4073.

There's some interesting and sneaky tricks you can pull if
you add some unprinting spaces to the justification routines
in our Gonzo Utilities.

The rule is that all leading line spaces will be deleted. If
in the rare instance you want to add leading spaces, preface
them with a |j or |k. Or both if your layout is extremely
critical.

If you want to force a fill justify on the last line of a page,
add as many spaces as you need, and follow them, again, by
a |j or |k. Naturally, this takes a last line that is "nearly full"
to start with.

If a command ends up at the end of one line and the action
is supposted to begin on the next one, the command may be
ignored. This can be a problem with web links or bolded
emphasis. The cure is to add one or more spaces to the
end of the previous line. This will force both the command
and the command action to be together as they belong at
the beginning of a line.

Beyond this tutorial, the best way to study Gonzo is to
view most any of the .PSL files in our GuruGram Library.

Once again expanded and updated our Gila Valley Day Hikes
library page.

Here's a list of links to our Wesrch papers. They sort of
form a "the best of" for our website...

Machine Language Programming Cookbook Classic Sampler
Restoring faded or scuffed text for web use
TV Typewriter eBook   
An "Un-Halftoning" scheme to improve eBook Images
Remastering a Technical Book for Web Distribution
The Incredible Secret Money Machine 
Allen Reservoir Failure Docs  
RTL Cookbook Classic Reprint 
IC67 metal locator classic reprint
Hanging Canal Slide Show  

Remastering video for web distribution   
Stability Issues in Gauss-Jordan Solutions   
Gauss-Jordan Solution of nxn linear equations
Staliiking the Wild Paradigm   
Prehistoric Hanging Canals of the Safford Basin 
Recent Developments in Magic Sinewaves  
Enhancing your eBay skills VIII
Website Link Checking Tools  
Secrets of Recent Technical Innovation  
Lessons Learned During a uv Lamp Debugging  

Some Possible Book Scanning "Gutter Math" 
Utilities for HTML & XHTML Revalidation
eBay buying secrets
eBay selling secrets  
Pseudoscience Bashing Secrets  
Isopod Energy Monitor   
Enhancing your eBay Skills V  
Build this TV Typewriter
The next big things   
Elegant Simplicity    

Enhancing your eBay Skills VI  
Cubic Spline Mininum Point Distance  
pv photovoltaic panel intro & summary   
Energy Fundamentals Intro & Summary
Real Time Acrobat PDF Animation  
A Solid State 3 Channel Color Organ   
When to Patent   
Exploring the .BMP file format  
150 Gila Valley Day Hikes  
A Gonzo PostScript Powerpoint Emulator

 Enhancing your eBay tactical skills VII  
 Synthesis of Digital Power Sinewaves 
Graham Tram Plan and Profile
Some fifth generation Magic Sinewaves  
Drawing a Bezier cubic spline through 4 data points
.BMP Bitmap Circular Lettering 
An expanded ultra fast magic sinewave calculator  
How to trash a vehicle hydrogen electrolysis   
A Partial History of the Gila Lumber and Milling Company
Some bitmap perspective lettering algorithms & utilities

How to bash pseudoscience
An Improved Bitmap Typewriter  
Using Distiller as a PostScript Computer  
Some Architect's Perspective Algorithms and Utilities 
Successful eBay Buying Strategies
Why Electrolysis Ain't Gonna Happen
The math behind Bezier cubic splines
Some Image Post Processing Utilities  
The Case Against Patents
Some eBay Selling Strategies

A Digital Airbrushing Algorithm
Don't Get Sick!  
Some More Energy Fundamentals I  
How to scam a student paper   
Some Inverse Graphic Transforms  
Nonlinear Graphics Transforms   
Three Phase Magic Sinewaves   
Bitmap to Acrobat PDF Image Conversions  
The way things were -- an unauthorized autobiography
A review of some pixel image interpolation algorithms

Some possible false color and rainbow improvements
An Executive Guide to Magic Sinewaves
The worst of Marcia Swampfelder
Acrobat PDF Post Document Editing Tools
A new method of solving electromagnetic fields
A Newbie's Intro to the Web
Gonzo PostScript Tutorial and Directory
An Ultra-fast Magic Sinewave Calculator
Some Energy Fundamentals
Secrets of Technical Innovation

Here is a brief example of the benefits of combining  
Acrobat X with my Gonzo Utilities for "Level II"
eBook rework. Full 16K sourcecode is found here.

The improvements in file size, text quality, image quality,
new value added features, and preservation of searchibility
can be stunning. Especially when starting with a pre-cyber
textbook.

The original Acrobat-only remastering can be compared here,
starting with page 9.

The process begins by scanning the original book pages to
Paint, and transferring these to Axcrobat X for a Level I
eBook remastering. . The text only output is then fed to
the Gonzo Utilities where a "standard font" conversion
can be done much faster, simpler, less infuriating, and more
conviently than redoing a fill justification from within Acrobat X.

Note that no text rekeying is required! When needed at all,
text mods are usually quite minimal.

The benefits of this route is that all text is now unicoded,
and very few fonts will need embedded. Full searchibility
is preserved, even after repeated PDF-->PostScript-->PDF
improvement steps. Text appearance over ClearScan is
generally significantly improved. And Gonzo's three stage
progressive microjustify does a better job than the
original phototypesetting could manage in the print book.

Since exising fonts can be shared, additional text-only
pages can add as little as 2K to your final file sizes.

Value added features are easily included. One example is
to replace italic fonts with tinted and bolded text.

Background speckle vanishes without a trace.

Bitmaps can be improved or outright replaced by
short and compact "raw PostScript" routines. The
original gray halftone boxes did not look all that great and
got considerably worse if you attempted Acrobat's
reduced file option. These got replaced with some
raw PostScript routines that offer color tints, smaller
file sizes, and much better appearance. They also
add full text sarchibility, which was not available in
the Acrobat original conversions.

The original B/W "pen and ink" drawings were converted
to full color. Mostly in a fast process involving the bucket in
Paint. With care, full color can often end up smaller than
the untined original.

A dilemma exists with Acrobat X's reduced font size
option. In that it largely prevents decent magnification
of any image. An interesting value added workaround is shown
in our example. In which any size reduced original can be
click expanded into a web delivered master image of
arbitrary quality and appearance. All done with only
a few dozen bytes of additional .PDF code.

Consulting and conversion services available.

I was asked if there was not a recent drop in eBay
sales.

It is just that usual little dip between the fall slump
and the winter slack period.

More eBay resources here.

Bingo! And Eureka! and Hoo Boy! I may have found
solutions to many of the limits of Acrobat X. For files
that are much smaller, fonts that are better looking,
and new value added. While fully preserving searchibility.

Just use ClearScan to output text to my Gonzo Utilities.
The fill justification font replacement in Gonzo is ridiculously
easier to use than the klutzy and infuriating text replacement
routines in Acrobat X. But you will still have to decide
exactly how much time and effort is justified for what
degree of improvement you want to achieve.

Switching to a standard font eliminates all the hassles
of non unicode and losing the ability to search after a
PDF-->PostScript-->PDF rework. Better yet, if you
are using ClearScan fonts on a page-by-page basis, there
may be thousands of these in a 300 page precyber
eBook with manditory embedding. Replacable by three or
four fonts total that may not even need embedding. For
a ridiculous reduction in final eBook file size.

Text quality generally goes up somewhat. The Gonzo
Utilities normally are a very good match that go beyond
the original fill justification. And, while seldom needed,
can force a rare mismatch with either an extra space
or negative kerning.

You can decide how many of which type of images
to improve. Ferinstance. an image might be colorized
and offered in the small size Acrobat "do not exapand"
extreme compression. While clicking on it can give
you an arbitrarily larger and better direct web access
for only a few dozen bytes increase in .PDF file size.

Fax style gray boxes can be replaced with raw PostScript.
For benefits of dramatically smaller file sizes, much better
looking uniform color tints, and new full searchibility.

Other images can be left alone, retouched, or replaced
with raw PostScript for similar benefits. Dramatically
improving most Pre-cyber book remastering.

I'll try to work up a demo on all this in a few days.
The Gonzo Utilities are, of course, proudly non-WYSIWIG,
so they clearly are not for everybody.

Found a few hundred more feet of the Mud Springs
Canal, acutally inside the "black hole of Central". The
full story is eventually likely to unfold as follows:

The Mud Springs canal is 7 miles long and a feeder to
a 3 mile long Jernigan Canal. It likely sources from far
up Ash Creek. Its early portions remaiin unverified, But
a rancher's PVC pipe verifies the route credibility and
there is no sane alternative to "Where else could it
have come from?"

A case can be made that this was possibly the first of
at least a dozen prehistoric canals. Owing to its entire
route being viewable at once from certain points. And
to the absence of certain fancier features, unusual
terrain, and nonobvious routes of the others.

There is a short stub canal that apparently dumps into
a wash midway at the Mud Springs Expressway crossing.
A possible explanation is that this is for mud control or
dealing with flood overflows.

Further downstream is a large and failed flood control dam.

Smaller and different from Allen Reservoir. The dam completely
runs roughshod over the canal without any accomodation of
any type. Which suggests extremely strong evidence that the
canal is in fact prehistoric.

A very well definied "hanging" portion follows, going "up"
a wash wall to continue. Just past where the canal once
again reaches the bajada is a strange circular structure.
Very pithouse looking, but probably centuries newer. And
only half a meter from the canal. Until proven otherwise,
this is the "troll house".

Somewhat further southeast, the canal appears to branch to
source the Jernigan Canal. Evidence in this area is somewhat
weak and consists primarily of long linear arrays of dead flowers
combined with the same linear array studiously avoided by the Creosote
Bushes. But all three ends connect to more credible evidence.

A small spur in this area leads to a small and obviously Anglo
tank. This is presumed to be a "steal the plans" and "dig
out an old ditch" similar to Hawk Hollow Tank.

The two canals parallel each other for a surprisingly long time
over very minor spacing. Presumably this was needed as a
"setup" to maintain slope on the two routes. The canals
never seem to get more than 800 feet apart. Sometimes
much closer.

The west spur vanishes into the Black Hole of Central to
emerge as the well defined Jernigan Hanging Canal. The
east spur also vanishes into the Black Hole of Central to
emerge as a well defined continuance that appears to
be headed for river bottomland near the Union Canal.
The final thousand feet is indistinct and largely trashed.

Slopes and terrain inside the Black Hole are eminently
credible. Possible explanations for the absence of hard
evidence are incomplete exploration, sheet flooding, or
stream piracy.

Meanwhile, the Allen Canal enters the Black Hole of Central
from the east. Where it goes and whether it merges with the
Mud Spring canal has yet to be determined. The only other
route would be unlikely as it is heading towards a cineaga.

Kiddies, we are dealing with utterly spectacular world class
stone age engineering here. Orders of magnitude beyond beyond.

Yet, with one or two grateful exceptions, I cannot seem to interest
the powers that be in the Southwestern Archaeology hierarchy.
Just because I skipped one too many compulsory faculty teas
half a century ago.

At one time, it was outrageously expensive to add color
on top of black and white printing. These days, adding color
to a web ebook reprint is almost trivial. And possibly
second only to live links as potential value added.

Consider these before and after examples. The color version
was done in something like twenty minutes using the bucket in
Paint. We've saved these as .PDF files. Amazingly, the color
version is slightly smaller than the B/W one at just under 200K.
If you do not need click-expand, the Reduced File Size option in
Acrobat can get you down under 24K. And that is with all PDF
overhead .

Some tips: It is common in black and white "pen and ink" drawings
to have white breaks between foreground and background.
These should be eliminated during colorization. One of the ruder
surprises of the paint bucket is that it will "break out" and trash
everything with even one pixel of true white on its border. Always
build "fences" in black or the pour color before using the bucket!

Also, do any resizing or rotations after colorization. Otherwise,
light gray "sugar" is likely to appear. The bucket demands
pure white in its pour area.

Be sure to move the Print Production option into the
Tools menu on Acrobat X. The Preflight Report
Generator is exceptionally useful. It shows you
each and every font and each and every image
in your file, along with great heaping bunches
of other stuff.

Usually after you have done a preliminary upload
of an eBook, you will want to improve the viewer
experience and reduce the file size. How much
time and energy you are willing to spend on
this task very much depends on the potential
value and income return on the posting.

While replacing bitmaps with raw PostScript is
the obvious way to do both, at present I do not
know how to do so without trashing the PDF
search ability. Clues appear in the toUnicode
objects and in Adobe Tech Note 5411.

While Acrobat X Pro's ClearScan does an amazingly
good job at replacing fill justified fonts, it seems to
degrade halftone images, especially gray fills.
These can get unacceptably bad if you use them
with Acrobat's Reduced Size PDF option.

I'm working on a mod to the Bitmap Typewriter
that should allow easy replacement of a gray
halftone with a true gray. This should also allow
text searching, and might provide for color tints
with only a modest file size increase.

Another glaringly obvious problem: New ClearScan
font glyphs are created for each and every page.
Which means that an eBook might have several
hundred "nearly identical" embedded fonts. The
workaround involves substituting one full and
complete and properly uniCoded font on each
and every page.

I'm not yet sure how well fill justification can be
preserved with font substitution.

The Preflight Report may reveal images that are supposed
to be text. These can be gotten rid of by retouching
the original scanned page bitmaps in Paint.

I had a kangaroo joke, but I forgot the paunchline.

Newly uploaded a Machine Language Programming Part I
to our rapidly expanding eBook Library.

This was excerpted from Chapters 6 and 7 of the original
Micro Cookbook volume I. Free download.

Amazingly, it largely remains useful and relevant to this
day. Especially for beginning PIC Programmers.

Other recently released eBooks include the Incredible Secret
Money Machine, the TV Typewriter Cookbook, the IC67
Metal Locator Thesis, and the RTL Cookbook.

Exporting from Acrobat X to a bitmap can be an extremely
useful tool. You can adjust and edit the bitmap or even change it
to very compact raw PostScript code.

But doing so with the default settings may lead to ratty text, even
after a ClearScan pass during the .PDF reconversion.

The workaround is to get into Preferences-->Convert from PDF
and change either .PNG or .TIFF resolution from Define
Automatically to 600 DPI or higher.

This may lead to enormous file sizes that may need later
adjustment.

If you manipulate bitmap BMP files with Paint or
Imageview32 before sending them to Acrobat, you
may end up with figures that are wildly the wrong size.

The usual cause of the problem is a resolution forced to 96
DPI or missing entirely from the .BMP header.

A simple workaround is to get into Irfanview, then select
Information, and change the header .BMP resolution to whatever
is needed for proper scaling. Make sure this is the last thing
you do before sending the file to Acrobat!

Note that this often will give you much higher final quality than
a simple size reduction . The more bitmap pixels Acrobat has to
work with, the better the final quality.

I've been exploring the new Acrobat X. One of its new
features is an option called "Save as Reduced Size PDF".
An average size reduction of 30 percent is claimed.

Aparrently part ( or possibly all ) of what it does is apply
some sort of extreme compression to any non-text bitmaps
remaining on the page. I found some page files going from 70K
clear down to 29K. That's the entire .PDF file, not just the
image!

It apparently can do this without degrading "average"
bitmaps. But I found this file reduction scheme unacceptably
degrades some B/W halftones. Especially ( particularly in the
case of a light gray fill ) halftones that are already marginal.

One workaround seems to be to apply file reduction selectively
only to those pages that give acceptable results. A second
is to get rid of the halftones, replacing them with grayscale or
full RGB images. Or raw PostScript artwork.

Picked up a Garmin eTrex H personal GPS navigator. It is
proving extremely useful in the Prehistoric Canal
explorations.

Ferinstance, I can get the GPS coordinates of a place I
want to go off of Acme Mapper. And then go to it
with the eTrex. Or record where I was, again to
Acme Mapper. Or set a waypoint for my 4Runner
in heavy brush and be able to return right to it.

Or purposely cover a loop route to explore twice
the terrain without worrying about exactly where you
are at all times. There also seems to be a highly
subjective "focus factor" that makes it easier to
cover more ground faster with less energy.

List price is around $100. Accuracy is typically
about ten feet. But you can dramatically improve this
by going to differential GPS. Either with a second
unit or a web or commercial service. Differential GPS
works by tracking the reported position of a known
fixed position over time. Correnctions are useful over
tens or even hundreds of miles.

The five buttons make entering alphanumeric data a pain,
but are otherwise intuitive and ergometrically useful.
While the buttons are intentionally hard to press, I
feel they overdid this "feature".

Also nice is reporting such stuff as distance and speed
traveled. As with most GPS, the altitude resolution is
rather crude and largely useless. Hundreds of waypoints
can be saved and stored for later analysis.

More on similar projects here.

It is the tiny details that sometimes can get to you.

Each tiny speckle on a scanned B/W Acrobat original can cost you
over 200 bytes in final file size. Owing to the object calling a
CCITT Fax Decoding. A dozen of these per page on a 200
page book adds up to half a meg of extra .PDF final file space!

You can spot these by viewing the .PDF file in a text editor
and searching for CCITT. Rows and rows of these means
you have a problem. These also will clearly appear when
you do a preflight report.

One workaround is to scan first to Paint and carefully
remove all the obvious glitches. Plus, of course, keeping
your scanner glass squeaky clean and unscratched at all times.

Here's a sneaky way to find any hidden glitches: Get
into Acrobat and edit objects. On a temp work file, delete any
known images, then any known text. Only the glitches should
be left. Remove them by retouching the original bitmap.

Found what likely is another half mile of the Mud
Springs Canal. Closing on the Jernigan Site and
with highly favorable topography.

But the intermediate terrain is proving enormously
difficult to yield further evidence.

More here and here .

Expanded and updated our Prehistoric Hanging Canals
of the Safford Basin presentation to include the latest
Mud Springs Bajada developments.

Renesas is giving awy free RX62N single board computers.

All you have to do is convince them you have a reasonable
ap in their present contest.

The USB powered card has some very impressive features,
including an on-board accellerometer, temperature sensor,
speaker, microphone, and three phase motor controller
circuitry.

Not your average KIM-1 knockoff, fer sure.

The plot thickens till it clots.

There is a 3000 foot or so square patch here I think I'll call the
Black Hole of Central.

Two prehistoric canals clearly go in, and two clearly come out.
But there is not the slightest trace so far of anything inside.

Possible explanations......

1. I could not find a pig in a dishpan.
2. I'm looking for the wrong things in the wrong place.
3. Massive sheet flooding in the past.
4. Stream piracy.
5. Modern trashing.
6. There never was anything there.

I guess I'll try full transects over the non-obvious portions next.

What is really frustrating is that straight line extrapolations
suggest the two canals cross at right angles!

Your participation welcome. Your Draganfly can be delivered
to 3860 West First Street, Thatcher AZ, 85552.

Latest GuruGram #116 is on Restoring Faded or Scuffed
Text for Web Distribution,

Its companion source code is found here, and additional
GuruGrams here.  

Two useful tools for eBay image prep and other postproc
are the open lasso and the octant remap.

Replaing the background on any sloped line or curve can
be painful. Instead, use the lasso to grab a portion of the
new background, being careful that its "open" slope
between its start and end nearly matches the slope
you are trying to match.

If you are working on edging something circular, get a 45 degree
portion of it exactly the way you want. Then flip and rotate
to build the full circle.

This should be nearly eight times faster than rebuilding the
entire circular outline. If you are careful with your flip and
rotate sequence, each new paste will exactly match the
mirror edge of the previous one.

Our eBay listings can be found here.

The TVT Cookbook is now available as a free eBook.

Several minor improvements remain to be done.

What did improve dramatically was the latest version
of the free Acrobat Reader.

Nav is much friendlier, screens cleaner, downloads seem
faster, and its behavior inside a browser much improved.

Finally got my copy of Acrobat 10 Pro. And I must say
that I am utterly and completely underwhelmed.

Yeah, the organization is a little better and it is
somewhat more convenient to use. And some of the
final file sizes can sometimes be significantly smaller.
ClearScan may be doing an even better job than before.

The loss of the text search ability in going from PDF-->
PostScript-->PDF still seems to have no obvious
solution or workaround.

Note that XP Service Pack 3 is required for this revision.

Did you know that Paul McCartney had a group
before Wings?

I'm thinking about adding several new pages to our
Gurur's Lair. Perhaps one that is eBook specific and
one on Gila Valley History & Prehistory.

eBooks currently available include RTL Cookbook,
the Incredible Secret Money Machine, TV Typewriter
Cookbook, and the Metal Locator thesis. With
parts of the Micro Cookbook scheduled next.

Some of the Gila Valley stuff would include the
Hanging Canals paper, the Hanging Canals slide show,
a Neely paper, the Allen Failure Docs, the Tramway
story, the Tramway slide show, and, of course, our
Gila Valley Day Hikes.

Let me know what else you want to see.

Repaired their broken Political Cartoons link on our home page.

A reminder: If you have a web page that others are likely to link
you should never delete any link!

Always replace it with a referral instead.

Here's a method of restoring faded, scuffed, or partially
missing pages such as back covers that can be most
useful when remastering classic books for ebook distribution.

It takes no special skills or software, but is rather labor
and time intensive.

First .BMP scan the cover or whatever in RGB to the highest possible
resolution, preferably 600 DPI. Then get into Paint.

Find the best lowercase "a" on the page and improve it as
best you can. Also upgrade its background. Save a copy of this
letter to a separate alphabet stash.

Copy the letter to your clipboard and use it to replace each and
every other "a" on the entire page. Then repeat for lowercase
"b", and so on through the entire alphabet. If several fonts are in use,
repeat for each and every font.

If the lettering is really bad, symmetry can help. As can building
new letters out of portions of older ones. Sometimes letter groups
or whole words can be used to make the process slightly more
efficient.

Plain old cut and past techniques can then be used to repair any
background between the letters.

Do the usual gamma, brightness, contrast, and sharpness adjustments in
ImageView32 or Irfanview. Then send your bitmap to Acrobat 10 for
.PDF conversion.

You may or may not want to try ClearScan. This may degrade the fonts on
subtle colors or strange art or few characters. But will give you full text
searching. ClearScan works amazingly well in black and white, but tends to
choke on pastels or anything mottled.

Going from .PDF ---> PostScript ---> PDf while doing
a book remastering has all sorts of stunning potential
for image quality, file sizes, and value added features.
But doing so seems to trash any search ability.

Some partial answers seem to be evolving. It turns out
there will be one or more ToUnicode objects in an Acrobat 9
generated PDF file. These directly relate standard and non-
standard glyph characters.

A ToUnicode sample might look like this...

/CIDInit /ProcSet findresource begin
12 dict begin
begincmap
/CIDSystemInfo
3 dict dup begin
/Registry (Adobe) def
/Ordering (SI-Fd188) def
/Supplement 0 def
end def
/CMapName /Adobe-SI-Fd188-0 def
/CMapType 2 def
1 begincodespacerange
<0000> <FFFF> % character space definition
endcodespacerange
33 beginbfchar
<0000> <FFFD> % unknown replacement character
<0001> <0026> % Ampersand
<0002> <002E> % Full stop
<0003> <0043> % capital C
<0004> <0053> % capital S
<0005> <0054> % capital T
<0006> <0061> % small a
<0007> <0062> % small b
<0008> <0063> % small c
<0009> <0064> % small d
<000A> <0065> % small e
<000B> <0066> % small f
<000C> <0067> % small g
<000D> <0068> % small h
<000E> <0069> % small i
<000F> <006B> % small k
<0010> <006C> % small l
<0011> <006D> % small m
<0012> <006E> % small n
<0013> <006F> % small o
<0014> <0070> % small p
<0015> <0072> % small r
<0016> <0073> % small s
<0017> <0074> % small t
<0018> <0076> % small v
<0019> <0077> % small w
<001A> <0079> % small y
<001B> <007A> % small z
<001C> <0020> % space
<001D> <0009> % horizontal tab
<001E> <000A> % linefeed
<001F> <00A0> % nobreak space
<0020> <00AD> % soft hyphen
endbfchar
endcmap
CMapName currentdict /CMap defineresource pop
end
end%

This is a plain old ASII remapping textfile, once the
Flate Encoding has been removed. The first column
is the ClearScan glyphs, while the second one is
the Unicode ( normal alphabet, sort of ) equivalents.

Acrobat or Acrobat Reader will normally use this file
object during searching.

Meanwhile, unless it is told otherwise, Distiller will
NOT convert special glyphs to plaintext. And thus
will block searching any later PDF conversion..

A partial workaround is found here. I'm still not clear
on how to deal with more than one glyph family per page.

The project continues...

Updated and expanded our Gila Valley Day Hikes
library page.

Found part of why searching is lost in going from PDF to
PostScript back to PDF. The key secrets are detailed in
Adobe Ap Note 5411.

Turns out that a ToUnicode file has to be preplaced into
Distiller for mappings to be recognized. This would seem to
get tricky fast if there are more than one ClearScan font
sets in a document. Or even on a page.

The ToUnicode file is a simple ASCII character by character
remapping.

Not sure exactly where to go from here. Still need Acrobat 10
for evaluation.

One of the foremost hallmarks of a good employer is
their not drawing that much of a distinction between
casual dress days and the clothing optional ones.

A process we might call "un-halftoning" can dramatically
improve most halftones used in book remastering.

Rescan the book pages to the highest possible resolution
with full grayscale or full RGB .

This should produce very large oversize halftones. Reduce
the images down to their normal size using ImageView32
or Irfanview.

Poof. All the halftone stuff should vanish as resizing does
its thing with bicubic interpolation. Instead of a black ink
halftone approximation to an image, we have a full gray
image approximation to a halftone!

Here are before and after examples. After can be
further improved with some retouching and making
the legs more similar.

I'm wondering if the real underlying cause of all of the
perpetual motion ludicrosities of the past few years is
a total failure to understand Fourier Series.

Just about all of the scams and stupidities in one way
or another involved "pulsed dc".

The only tiny problem with this is that there is no
such thing as pulsed dc! You instead have a waveform
with various ac and dc terms. More often than not,
each term can be separately considered and evaluated.

Ferinstance, only the dc term in a pulsed waveform
( or ac terms that rectify to a dc component ) plays any
role whatsoever in electrolysis. As guaranteed by
Faraday's Law.

Any low duty cycle waveform is exceptionally easy to
mis-measure, owing to its strong difference between its
average and rms values. Invariably, the wrong measurement
will be low. Often wildly so.

More here and here.

One of the principle indicator species of overgrazing
is ----> cows.

What recession? Copper is near $4.01 and Cotton is near $1.19.
Historic highs.

Few locals seem to be complaining. Except for the builders
of totally tanked real estate.

Even there, outstanding auction buys exist .

Latest GuruGram #114 is on Remastering a Technical
Book for Web Distribution.

Its sourcecode can be found here. And two examples of
eBooks in process here and here.

I still do not have conclusive proof that the prehistoric
Mud Springs canal exists. IF it exists, it would be
one of the most significant of Mt Graham water
exploitations. Routing the highest flow rate stream
of Ash Creek "up" and over the largest and most obvious
bajada.

For a projected length of five miles! A case can be
made that this might have been the earliest of the
canals. Based on the entire route being visible from
its starting point. Which would vastly simplify its
engineering.

The evidence so far: There is an obvious and
( largely ) unarguable half mile solid chunk of the
canal halfway along the route. It even includes
a diversionary canal stub presumably used for
flood control or mud removal.

The Jernigan Canal provides an obvious teminus
for the Mud Springs System. It too is well defined
and unambiguous, but only a quarter mile long. But
any intermediate evidence has yet to be found.
And another canal is within 800 feet but defies
any and all interconnect attempts.

The only rational source for such a large canal
system would seem to be Ash Creek. And a
rancher has laid a PVC pipe along the route,
both proving its feasibility and raising "stole
the plans" issues.

But once again, any early or diversionary
evidence has yet to be found. The topography
is certainly feasible. 4WD access is rough.
Miles are foot only.

Field mice are very much needed on this project.
email me if you have any interest.

And your contribution of a Draganfly can be shipped
directly to 3860 West First Street, Thatcher, AZ.

A Flate Compression viewer can be found here.
With its tutorial here. This can be useful in
analyzing .PDF documents.

The returned results might be either 8-bit or else
7-bit data. Others may require additional terms
or interpretation in the DecodeParams dictionary.

That's not a bug -- its a feature! Going from Acrobat PDF
to PostScript back to PDF can give you all sorts of file
size and image quality improvements. But I was mystified
as to why any search ability was lost in the process.

It turns out that ClearScan glyph search linking is intimately
linked to ClearScan itself. And that ClearScan demands a
bitmap input and will not work in the presence of other fonts.

Sadly, I do not yet have a reasonable workaround. Font
substitution does not seem to make any difference. And
outputting as a bitmap and reClearScanning would negate
any PostScript benefits.

I have a hunch this might be an unintended side effect of
trying to protect the IP rights of the original fonts.

It will still be several days before I can determine if the
Acrobat 10 behavior will be any different.

Many hints as to how ClearScan operates can be found
by exporting the .PDF file to PostScript and viewing the
results in an editor or word processor.

Apparently each encountered glyph is numbered with
a sequential name, such as /000, /001, /002 etc. These
names are directly used in a xshow loop.

Even if font substitution is in use, the names remain
the same. Thus the ClearScan text imaging is fundamentally
different from the usual alphabet based xshow loops.

But surely somewhere there has to be a table matching
glyphs to the font alphabet. Finding and understanding this
process remains a major mystery.

Found a rather strange local cienega. One that gives
credence to the theory that there were a lot more
springs in the area and they were a lot larger than present.

This one is strange. It has reeds and breifly flowing
water. The water does not appear bitter or warm.

There are no obvious signs of anything prehistoric,
although it is only a thousand feet from a fairly major
habitation site. Curiously, there are no signs of modern
development as well. Such as fences, traps, dams,
or diversions. It is "just there".

An interesting question is whether the Allen Resevoir
was once part of a fairly major cienega. The Golf Course and
Cluff poinds may also have had similar origins.

"Truth is stranger than fiction" gets even more bizarre
when you wander into Western New Mexico.

The Brushy Mountain Radar Station is South of Mule
Creek at the end of a secret mountain laboratory road
whose access is easily controlled. It started life as a
cold war facility and presumably still sees use for drug
interdiction activity.

Little known is that it includes oversize kitchen and
dormitory facilities. It was supposedly used as a remote
retreat by both the Kennedy and Johnson presidental
administrations. Since then it occasionally has seen such
mundane uses as BLM team building excercises.

BUT and IF you needed to stash some super secret stuff
( such as, say, some extraterrestrial aliens ), the facility is
virtually ready to go.

Even stranger is Terrortown, once known as Playas. This
began as a company town for a long defunct smelter
and was bought by a New Mexico school and funded by
the Department of Homeland Security for a training
facility for counterterrorism and urban hostage situations.

More here, here, and here.

Truth can certainly be stranger than fiction. As a bunch
of Gila Valley happenings over the years can attest.

One of the most blatant scams was the McEniry Tunnel,
a scheme to tunnel all the way through Mt. Graham.
The gold and silver could simply be scraped off the ceiling
into ore cars, greatly simplifying extraction. At the same time,
zillions of acre feet of water could be recovered, along with
great heaping bunches of electricity. Plus lots of timber.

The entire prospectus, of course, was an outright lie. The
Grahams are precambrian intrusives with virtually zero
mineral content whatsoever. The site today is a plain old
short mining tunnel.

BTW, all the locals have their own favorite spelling and
pronounciation of "McEniry". And love to argue about it.
I'll stick with his signature on the above prospectus. This
is also sometimes called the "Triumph Tunnel Site".

Somewhat further west was the Spenazuma Mine, shortened
from getting "them" to "spend their mazuma". And a
classic example of blatant salting. Today, this is on a private
ranch on which visitation is strongly discouraged.

What might or might not have been a scam was the Bear Flat
Irrigation District. In which artesian water was run over a
long series of canals and lakes in what today is totally barren
and dry as a bone.

More modern is the saga of the Banana Farm scam. Which older
Thatcher residents do not want to talk about. Also, for some
strange reason, nobody but nobody in the entire Gila Valley
wants to talk about the "Golden Letter" scam whose pyramind
scheme flushed out the entire area during the 1980's.

Next eBook project will be Micro Cookbook Volume II.

Which eventually should be up here. I'm waiting for
Acrobat 10 to do the preliminary build on it.

I still feel that this text never got the attention it
deserved. Much of its content remains relevant.

Updated and expanded our Auction Help library page.

While the total number of auctions seems sharply up,
there sseems to be a dearth of "Stuff I can Sell on
eBay ones.  For many months now.

An interesting new humidity sensor can be found here.

Welcome to hell.
Here is your accordian.

I have a hollow feeling that I've missed many of the
obvious candidates in our Gila Valley Day Hikes. Or that my
own personal interests or biases are tinting the collection.

I just added a few obvious omissions that should have gone
in long ago. Please email me with any new suggestions of
your own. No, not that one.

A reminder that there will be a talk on many of these this
saturday at 6:30 PM in the Discovery Park Jupiter Room .

Yesterday's canal chasing was more or less a debacle,
missing the intended locations by a few hundred feet.

We did more or less prove that the Mud Springs Bajada
route seems viable and that a rancher ran a pvc pipe
over pretty much the same terrain in what sure looks
like "steal the plans" to me.

The premise so far: If "they" in fact were to exploit every
drop of Mt Graham water, their crown engineering jewel would
be routing Ash Creek water "up" and "over" the Mud
Springs bajada.

So far, there is a half mile of mid canal "proven" to well beyond
acceptable on-ground evidence. And the Jernigan Canal
"needs" a water source that is topographically compatible.
But any linking proof remains elusive.

Meanwhile, the only feasible source for the mid canal would
in fact be Ash Creek. But, again, linking proof remains elusive.

Expanded, updated, and added bunches of new links
to our Gila Valley Dayhikes library page.

A reminder that I'll be talking on this Saturday at 6:30 PM
in the Discovery Park Jupiter Room. See you there.

Just got an email from someone wanting to make a hot
chassis guitar amplifier "safer" by adding an isolation
transformer.

My response, of course, is DON'T EVEN THINK OF IT!

If the amplifier was hot chassis, it was done so to keep things
as cheap as possible. No way could it have even then sounded
decent. And it certainly would not have improved with age.

Low end leaky capacitors, single ended pitiful output, excess
heat, lots of hum, and all. The basic premise of the circuit was to only
moderately electrocute the guitar player in order to save a
few pennies. The circuit was so bad that special protection
was needed for the preamplifier tube!

An interesting collection of tube type el-cheapo amp schematics
can be found here. Most of these are not even appalling.  

If you like the hum, noise, and distortion that a tube type amp adds
to otherwise clean audio, that's fine. But the way to do it these days
is to start with a decent Class D solid state amp and add a DSP chip that
purposely emulates your choice of make and model of tube type
hum, noise, and distortion.

Which, of course, would only be an approximation, because no two
tube amps of the same make and model ever sounded exactly alike.
You can beat this with a decent pseudorandom sequence generator.
Turn the crank till it sounds right.

Best avoid electrocuting the drummer. Better they die in a bizarre
gardening accident.

Duh. Sometimes you can increase the resolution of
a scanner simply by asking it for more. If you are
scanning into Paint, you can use your Custom Settings
to switch from 200 DPI default to, say, 600 DPI.

For nearly ten times the small object resolution!

Which is perfect for small eBay image generation.
Per this example.

Naturally, you'll want to limit yourself to small subjects
or your total file sizes will quickly get outrageous. Be
sure to quickly crop to subject!

Adobe just announced what they are variously calling
Acrobat 10, Acrobat X, or Acrobat 2010. Supposedly
it has further improved its already outstanding ClearScan
OCR techniques. The new code is supposed to be a
lot faster. more compact, and even better looking.

I'm still trying to find out if an Acrobat 9 glitch is a
mistake I am making, a serious bug, or an intentional
crippling: An obvious post-processing improvement in
acrobat is to replace certain bitmaps with raw PostScript
procs that are better looking and far more compact. This
also allows substitution of different types of bitmaps,
particularly to improve and upgrade a halftone.

At any rate, if you export from Acrobat 9 Pro to PostScript
and then import back to Acrobat 9, everything works just
fine, except the search ability is completely lost! It is as if
the mapping betweeen the ClearScan glyphs and their
underlying English letters never gets properly exported.

Meanwhile, the cheapest quasi-legal way to buy Adobe
is through the Academic SuperStore. These folks will demand
something that looks vaguely like a pay stub from a school.

Another route is to pick up an older version of Acrobat at
an auction or wherever and then buy and install the much
cheaper upgrade.

At one time, the only popular ways to emphasize text were
underlining, italics, or bolding. Of these, underlining was the
only reasonable option available for hand writing or an
ordinary typewriter.

These days, of course, underlining should be tolerated in the
sccond grade, strongly discouraged in the third, and severely
punished in the fourth. Fortunately, most compulsive underliners
eventually grow so much hair on the palms of their hands that
they no longer can type.

Since color is now essentially free, it has largely replaced
traditional emphasis forms. We will often use blue for links,
brown or aqua for plain old emphasis, purple for sourcecode,
and a rare red for urgency.

It turns out you can simply fake an italic font in PostScript,
largely eliminating any need for them. There are two ways
to specify a PostScript font. One is to simply state the size
as an integer. The other is to use a six element array of
[ wide climb lean high xshift yshiftt ] to create your font.

Ferinstance, /Fontname findfont [10 0 2 12 0 0 ] makefont
setfont would create a slightly compressed italic font. As
would my Gonzo Utilities equivalent of /font1 /Fontname
[10 0 2 12 0 0 ] gonzofont.

Many more font tricks here.

I'm wondering what our prehistoric hanging canal explorations
can tell us about climatic reconstruction.

The evidence seems to be accumulating that ( even factoring
in our present severe drought ) things seemed wetter then.
And possibly very much so.

Ferinstance, today there is not nearly enough water in Spring
Canyon to justify any canal project. Let alone a major seven
mile long one. Hand built with stone age technology.

And one credible explanation for an early lake in the failed
Allen Resevoir area could be a large cienega. To this day,
there are cienega hints there as well as both up and down
canyon.

Specialized interpretive help is obviously needed.

MAJOR BREAKTHROUGH!

It turns out that it is ridiculously simple to edit Clearscan
glyphs when you are scanning a book for web distribution.

Completely legal and without any encrypt/decrypt hassles
as well. You can easily end up with absolutely perfect
typography and also exactly preserve fill justfification.

While a bitmap is temporalily involved, it is no worse than
the bitmap you used during your original scanning. All results
eventually end up fully stroked vector and fully searchible.

Have Acrobat 9 pro export a .TIF file for you. Open a new
version of good old Paint, converting its attributes to B/W and
then load the file. Create an optional second similar paint instance
to use as a convenient alphabet stash.

Start with the first character or the first really bad character,
depending on how much time the project warrants. Improve
this character till it is "perfect"

Typically, you will eliminate internal or external collisions, make
the feet smaller and the stems straighter. You might also make the
character a pixel narrower or shift it a pixel or two sideways.
Mirroring or copying can dramatically improve certain characters.

IMPORTANT: If you change one instance of a character on the
page, be sure to change them all. Otherwise, your "good" character will
just get averaged in with the bad ones.

When finished, drag the corrected TIF file back into Acrobat 9
Pro and create a new PDF file. If needed, the process can be repeated
as often as you like.

Some of the process can likely be automated, perhaps with B/W techniques
othwerwise similar to our Bitmap Typewriter.

Should a bitmap page end up the wrong size, simply correct its
horizontal and vertical resolution in its header. One easy way
to do this is with IrfanView.

The more engineering I become involved in and the more of it
I see, the more it gets down to its fundamental definition of
a sense of the fitness of things.

The local prehistoric canals I've been exploring are so utterly
beyond incredibly brilliant that they clearly blow away any high
tech in the Gila Valley to this day. Given the available tools
and technology.

The reason you hang a canal on the edge of a steep sided mesa
is that this makes your slope independent of terrain! Leading to
astonishing savings in energy and the amount of material that
needs moved.

One of their latest features to be pondered over are cutouts,
apparently used for flood control or mud removal. Of the eleven
canals of over 30 miles to date, at least three seem to include
these cutout features.

Even more amazing, if the premise was made that their goal
was to totally exploity every drop of mountain stream water,
an extremely conspicuous and difficult site seemed totally
absent. That site has now been found and verified. And appears
to route Ash Creek water "up " and "over" the Mud Springs
Bajada.

We are now so ridiculously far beyond world class that terms like
"second only to Phoenix" or "second only to Tucson" can
clearly be dropped. Engineering a flatland river canal is utterly
trivial compared to mesa hanging.

Field mice are definitely needed. As are toys. Topping the toy
list is a precision MEMS altimiter and, of course, a Draganfly.

Please email me if you want to be part of the team.

I'll be doing an upcoming Discovery Park lecture on
365 Gila Valley Dayhikes. 6:30 PM in the Jupiter
Room on November 13th. Free admission.

One of the continuing surprises of our hanging canal
studies is how much "they" relied both on micro and
mega topographic features. In many cases, an unusual
and highly localized topo feature that is totally counterintuitive
plays a key role in making an entire canal system possible.

One obvious example is where the Deadman Canal
purposely crosses the narrowest and highest point on its
mesa, exactly where a three way switch can be best
used to route water to three wildly different drainages.

Or the several instances where a canal makes a "U" turn
and purspurposely routes itself back UP canyon, contrary
to the predominate mega terrain direction.

Or many instances where "up" is really "down" along a mesa
edge. The wonderments continue.

I'd forgotten how stunningly simple a technique called
arcto crowding can be to make fancy borders. It is
detailed in our Gonzo utilities and examples can be
found here ( Particularly ATG 59 and 72 ) and here.
( Particularly LC 45B ).

Gonzo has a routine called roundpath that rounds the
edges of any arbitrary path...

/roundpath {/rpdata exch def /rprad exch def
rpdata length 1 sub cvi /rppoints exch def rpdata
0 get rpdata 1 get moveto 2 2 rppoints 2 sub
{/rpvalue exch def 0 1 3 {rpdata exch rpvalue
add get } for rprad arcto pop pop pop pop} for rpdata
rppoints 1 sub get rpdata rppoints get lineto} def

Note that only the internal edges are rounded. The initial
and final data points close on themselves and complete the
figure.

Now for the sneaky part. If you put a box or whatever at
each corner of your figure, the arcto operator may not have
enough "room" to operatate properly. Giving you a mind
numbing variety of special border effects.

These can be especially useful combined with the Gonzo
superstroke operator or intentionally using negative radia.

The optimal grade for a prehistoric canal is often
something around two percent. Or Four feet of drop
for each 200 feet of reach.

The canal will not work at all with positive slope, and might
self destruct with stronger negative slopes.

The resolution needed is much better than GPS,
especially with low angles on the birds. And older
sport altimeters also have limited resolution.

It turns out there is a new generation of MEMS
pressure sensors that might do the job. Such as
a THREE CENTIMETER (!) one or this
NINE CENTIMETER (!) one. Prices for the
bare sensors themselves are in the $39 range.

Differential operation should be trivial, since you
would simply walk back to your starting point.

Finding suitable terrain would go a long way to deciding
exactly where the "missing" portion of a canal can
or cannot go.

It would be best to replace any line art bitmaps left in
ClearScan with raw PostScript vector procs, but several
yet unsolved problems may prevent this.

Here's an alternate "bitmap improver" scheme I thunk
up that should be fairly easy to implement:

Change the bitmap from B/W to some color mode. I'd
likely stay with full RGB because of our previous tools
and experience in this area.

Get in good old Paint and trace everything you want to
to keep in red. Make your image much larger than you
need and stash an ongoing library of useful shapes and
lettering where they can be cut and paseted into position.

Then write some simple raw PostScript code similar
to our postproc tools that scans the bitmap one pixel at a timt.
Replacing all red pixels with black ones and all the others
with white. Recrop to original size and resave as B/W.

Several of the prehistoric hanging canals seem to have
branches that do nothing but drop uncontrollably off a
mesa edge.

A possible credible eplanation is that these are either
storm cutouts to prevent high flood damage. Or that they
are a maintainence tool to periodically remove mud
from the main canal.

The Allen canal cutout seems associated with considerable
caliche scouring. Suggesting uncontrolled flow. Another
potential use problem is that the bottom of the caliche is
only 1400 feet away from the below-dam canal continuation
on rather unfavorable terrain.  Other uses seem unlikely.

A newly discovered cutout on the Mud Springs canal
is remarkable similar.

The ebook version of the ISMM Incredible Secret
Money Machine is now available for free download
here.

There's still a few minor details I'd like to improve.
I purposely did NOT do any revisions or updates
beyond correcting area codes, adding a URL, and
updating the ISBN.

So much has changed and so much has stayed the
same. And so much more is coming around again.

I used the tan Edition II cover. If you have one of
the original covers with the False Aurelia on it,
please hires scan it and email me a copy.

Type I font glyphs have long been well and freely
documented in the Type 1 Font Format. I've
written some crude PostScript code to ecrypt,
decrypt, and plaintext convert these glyphs
here and here. Using my Gonzo Utilities.

A current project around here is finding some
way to post edit ClearScan fonts out of
Acrobat 9. After some checking and as near
as I can tell, there seems to be additional encryption
added to these font glyphs. Perhaps because it
would be trivially easy to steal fonts you do not
own using ClearScan.

The usual keys do not work. Nor do the new
"Passwords". I'd expect many glyphs to end with
the closepath 9 and endchar 14. Exhaustive
searches using standard Charstring encrypt
gave no useful results. I suspect they are using
a different pseudorandom sequence generator.

Here's an alternate method towards ClearScan
font editing that does not mess with any "secret"
encryption: Export as a bitmap with the same
resolution as the scanned original. Retouch
the export and use it as a new "original".

Obvious things to retouch are missed OCR
recognitions, improved kerning, and straightening
out problem letters. Often "black space collisions"
can be eliminated by making each and every letter
somewhat less wide by a pixel or two.

Do note that each and every occurance of any
sized glyphs will need replaced, so that ClearScan
will exactly "average" to your new character.

The process can be made a lot faster and a lot
less tedious by also exporting to PostScript and
extracting an absolute character position list.

Techniques similar to my Bitmap Typewriter can
then be used to automate glyph rerplacement.

Chances are that file sizes can also be significantly
reduced with these sort of stunts.

Meanwhile, if you have found a simple and legal
method of reading and modifying a ClearScan glyph,
please email me.

Just found another new mountain stream fed prehistoric
hanging canal!

It is much too early for even speculation, but it may
be five miles long, may source far up Ash Creek,
and may be the long sought delivery venue for
the Jernigan Canal fields.

Tentative Name is the Mud Springs Bajada Canal.

We are now up to at least eleven hanging canals
with a total length approaching thirty miles! The
engineering here is orders of magnitude beyond
stunning.

Please email me if you want to help with the exploration.
You do have to be the type of hiker that brings
along your own catclaw, just in case there is not
enough along the route.

Next trip leaves Sunday morning.

It turns out there were several variations on the ISMM
Covers. The earliest was the "Dirty Hippy" one that did not
set too well with either Sams or certain booksellers.

In which the guy had a beard and his shoelaces untied.
The false aralia was also considerably bushier and larger.

This cover is very rare and hard to find. I do not seem to have
a copy.

I guess the real problem was that I was trying to make the ISMM
both the best small business book of all time AND a parody of all
the scam ripoffs of the era.

When we took over publication as a Synergetics Press
title, we did not own the cover art and could not afford anything
but "one ink color on duplex cover stock".

A slightly brightened version of this tan cover appears here.

Found two curious "features" in Acrobat ClearScan.
I'm not sure if I am doing something wrong, if they are a
bug, or if they are an intentional crippling.

If you have the time and dollars available, spending
extra time post-processing an eBook you are uploading
for web distribution should certainly end up worthwhile.

An obvious way to improve a mangy bitmap of line art
is to replace part of all of it with raw PostScript procs.
Besides looking better, chances are the file sizes will
be ridiculously smaller as well.

Using my Gonzo procs.

But if I export PDF-->PostScript, edit, and then
import PostScript ---> PDF, everything weems to behave
just fine, except the text search no longer works! It
is as if the links between the ClearScan glyphs and
their "real" alphabet characters have vanished without
a trace.

Retouching a bitmap might seem best done before any
ClearScan or OCR. And a strong case can be made to
scan into Paint rather than Acrobat 9 because of this.
But if there is any deskew needed at all in the ClearScan
process, permanant "wrinkles" will end up built up into
your bitmaps. Ot they will simply "lean".

I seem to be having problems with the obvious route of
PDF-->Bitmap, edit, then Bitmap-->PDF in that the sizes
often may end wildly different in the process. The
choices seem to end up too big to process or too ugly
to use.

Some more ancient history that came up on our
helpline: A connector being male or female is
determined by the INNERMOST conductor or
conductors.

This is no problem on phone or phono jacks.
Or on power line cords. The one that usually
Causes the most confusion is a BNC connector.
Female BNC jacks normally panel mount. Male plugs
normally terminate a cable.

Even more subtle is the type F cable connector.
The inner coaxial cable wire serves as the male
connector. Again, almost always, the male type
F connector plugs go on the cable and the female
jacks panel mount.

The weekend saw many new canal discoveries.
Which, as usual, created more questions than
they resolved.

Many deep cuts were found in the lower Marijilda
area. Which added strong credence and credibility
towards the Culibra Cut in the Allen Canal
being in fact prehistoric.

The mid portion of the Allen Canal was extended
nearly to the north mesa edge. But it raised a distinct
possibility that the upper and lower reaches are
in fact two different canals.

One question is whether the Allen Resevoir received
most of its obviously high inflow from the upper
Allen canal or from once large but now defunct springs.
The watershed is simply too small and runoff free to
offer other obvious alternatives.

Meanwhile, the upper Allen canal could in fact
end up as the missing source for Jernigan Canal
water. But link evidence to date is sorely lacking,
owing partially to sheet flood damage.

If these two canals are distinct, we might
rename the upper Allen Canal as Spring Canyon
Canal.

More here and here.

As we've seen, the "find OCS Suspects" selection
does not work in Actobat 9 Clearscan, and it is not
obvious whether it is going to work in newly amnnounced
Acrobat X Pro.

Here are four different routes to find out which text is
nice looking and compact and searchible, and which
remain sloppy unsearchible bitmaps...

RETOUCH METHOD -- Select Tools-->
Advanced Editing-->Touchup Text. Then
mouseover all available text. Solid blue
words will be both smooth and searchible.
White letters are still bitmaps. Words with
white stripes are smooth but not yet serchible
as a single word.

RESIZE FONTS METHOD -- Select Tools-->
advanced editing-->Touchup Text. Then
mouseover all available text. As above.
Then use Properties to switch to a one point
type size. OCR candidates will remain in their
original positions, small converted fonts
will move.

MAGNIFY METHOD -- Select extreme
magnification and then mouse scan the full
page. Rounded but slightly erratic characters
are compact and searchible. Crude bitmaps
remain just that and likely need fixed.

POSTSCRIPT METHOD -- Export PDF as
PostScript and view in an editor or word
processor. Most smaller bitmaps using the
sepimg operator will be text that still needs
repaired.

The latest buzzword scam de Jour seems to be "Energy
Harvesting". And piezoelectricity seems to be their
favorite dead horse being rewhipped.

We looked at piezo fundamentals way back here.

There are two fundamental flaws involved: First,
piezo is an e-field machine, meaning that it has an
infinite source impedance. And thus is inherently
and intrinsically horribly inefficient as an energy
source.

Secondly, all of piezo is fundamentally a capacitor.
Meaning that it only responds to changes and only
produces an ac output.

Which explains why virtually all piezo uses today
are below the one watt level and have negligible
efficiency. They also self-destruct if their max
Curie Point temperature limit ( comparable to very
hot water )  is exceeded.

Yet another problem with most energy harvesting scams
is their duty cycle. There is no way in hell thaat cars
running over a sensor can generate useful energy.

Power yes, energy no.

Too bad nobody knows the difference.

There is a shareware program called IrfanView
that is an interessting competor to ImageView32.

Its feature that I need the most is the ability to
rotate in tenth degree increments. It also has
bunches of "Gee Whiz" special features that
I am not sure who would use for what.

High resolution rotates are essential for architect's
perspective and scanned text deskewing.

Gamma correction, and sharpness seem far
more convenient in ImageView32. I'll likely
keep using it for most of our image post processing.

I've been making only limited progress in a goal of
easily editing Acrobat 9 CearScan OCR pages
to further improve them while remastering a book
for web delivery.

There are hints that Adobe has once again gotten
snotty about attempting to protect font glyph paths.

At any rate, if you export a ClearScan .PDF file to
raw PostScript, you can easily read and adjust parts of
it using an ordinary text editor. It starts out with bunches
of working routines. Then it outputs one bitmap at a time
of amy images it was unable to recognize. Each bitmap
is preceeded by its own clipping interval. And may be
overwritten by OCR recognized text.

You can simply write over or replace these images with
these techniques. Just enter any "paste over" code
before your showpage. Then redistill. Files can optionally
be shortened by eliminating any images you are going
to overwrite anyway. Many images can be dramatically
shortened by replacing them with "raw" PostScript code.

Following the images are the new fonts, created by
averageing out the recognized shapes of the originals.
This is amazingly powerful and downright brilliant,
especially at preserving fill justification. But some
of the fonts remain hokey owing to limited samples
of limited resolution.

Easy and simple font editing sure would be nice.

For any given font, a list of glyphs is provided. These
glyphs are then placed and spaced as needed to build
up composite text messages.

However, so far I have been unable to decrypt the
glyphs. The detailed documentation for eexec or
charstring does not seem to work. Neither does directly
using their defined password.

Please email me if you know what I am missing here.

Some of what I thought was ancient history:

When Adobe first came out with their font
technology, they introduced a bizarre eexec
encryption scheme. Which they thought was
absolutely unbreakable and bulletproof.

I think I was among the first to discover that
eexec had a fatal flaw that any reasonably
patient seventh grader could sight read:

All you had to do is insert an extra character
and the PostScript error reporter would tell
you what was on the stack and what your
"spelling error" was. With patience, you
could un-eexec just about anything.

And, thanks to the PostScript stack, you
could jump ahead dozens of characters at
any time.

Especially fonts. It was sort of like a safe
combination lock telling you "Uh, try two clicks
to the left".

Adobe long ago decided to publish the
eexec secret encryption in their free Type 1
font Format text.

While normally done with C language routines,
with two sneaky tricks, you can easily encrypt
and decrypt "regular" eexec in raw PostScript.

PostScript has no explicity defined 16-bit integers
But you can force these with a 65536 mod. And the
high byte can be extracted by a 256 idiv .

One step in either the encrypt or decrypt requires
multiplying a 16-bit word by 52845. Doing so
may exceed the PostScript integer limit of 2^31-1

Fortunately, 52845 can be factored into 10569*5.
Simply multiply by 10569, do a 65536 mod, then
multiply again by 5. Perhaps like this...

/rstart 4330 store
/c1 52845 store
/c1f1 10569 store
/c1f2 5 store
/c2 22719 store

/decryptcharstring {/r rstart store
{dup /curinchar exch store r 256 idiv xor
buildoutputhexstring  % details omitted

/r curinchar r add cvi 65536 mod c1f1 mul
cvi 65536 mod c1f2 mul cvi 65536 mod
c2 add cvi 65536 mod store
} forall }store

While this does the format book examples beautifully,
I can't yet seem to get it to work on ClearScan glyphs. I
am not sure if I have a mistake here or if Adobe is
once again being snotty about encryption.

A dilemma: Acrobat 9 Clearscan does its OCR by far
the best with a black and white scan. But this makes
halftones dark and murky at best. Halftones prefer
a grayscale scan instead.

A workaround is to rescan the halftones as grays
and then paste them into the final PDF using these
techniques. And overlaying any bad halftones.

It is always difficult to web remaster decent halftones.
But there are some tricks you can try to bring them
up from "mesmerizingly awful" to "barely acceptable...

Seek the best alternate source of halftones.
If possible, scan as grayscale and repaste.

Color tints may help. Knock out backgrounds
to white. Untilt to architect's perspective if
applicable.

Outline edges gently with thin grays or blacks.
"Chase" details along their length for
consistency. Make surfaces more uniform.

"Average out" areas. Eliminate any shadows
that are unneeded or murky.

Make sure scans are horizontal! Deskewing
does bad things to halftones.

Try lightening image. Somewhere between
murky dark and washed out gray.

Make open holes white. Use repitition to
make pins or legs more consistent.

Symmetry via mirroring or flipping can
mreplicate the best part overall.

Always try to do your own scans, rather
than relying on others. 600 DPI whenever
reasonable or possible.

Use raw PostScript and my Gonzo Utilities
for enhancement.

Preserve legality and provenance for
anything of historical context.

Replace true circles where needed.
Add extra white or black dots to
darken or lighten.

Make the worst the best. If all else
fails, divert attention elsewhere.

Take half and leave half on the worst
of the corrections.

More on image postproc here.

Well, I wouldn't exactly call it progress, but I did
make a bunch of lateral arabesques in my quest to find
a simple and easy way to edit Acrobat 9 fonts.

The ClearScan in Acrobat 9 is utterly amazing
in its ability to scan pre-computer books and
greatly improve their fonts while fully preserving
fill justification (!). Espeically if destructively page
scanned at 600 DPI or higher and with very careful
and utterly flat page alignment.

Per this "still needs improved" example.

But the final fonts still can look rather ratty,
since they are based on an average of ( sometimes )
only a few slooppy samples. It sure would be nice
to reach in and "retouch" or "adjust" all these
fonts to greatly improve them.

So far, I've found five rather gruesome and painful
approaches to the problem...

Export to a TIFF file, repaste all of
the characters in Paint, and re-OCR
with ClearScan.

Export to a TIFF file, repair and
collect the characters using Paint
into your own new alternate font and
selectively replace characters using
the well hidden Properties section of the
TouchUpText tool.

Export to PostScript, read in a text
editor, and modify the font glyphs
as needed. Then redistill. The only
tiny problem here is that the glyphs,
while easily read, are Charstring
encoded. Per these details. Some
simple raw PostScript encrypt/decript
code should be possible.

Snoop around in the Acrobat 9 SDK
to see what can be built up as a custom
plugin. This takes C-language skills.

Please emai me if you find any simpler or easier
to use Acrobat 9 font editing alternatives. Meanwhile,
I'll try to work up a raw PostScript Charstring
encoder/decoder.

Started the web conversion of the Incredible Secret
Money Machine.

You can follow the progress here.

Latest GuruGram #113 is on the Allen Dam Failure Docs.

Its sourcecode is found here, and additional GuruGrams here.

The RTL Cookbook upload should be fully complete
now. An Acrobat reader is required for download.

Please report any problems you find with this by
emailing me.

Other recent classic reprints include our Thesis
and our Intro to PostScript Video. Which can
also be found on YouTube here and here.

There are many "secret" font characters in most of
the Adobe fonts. These are coded but not immediately
keystroke accessible. The possible characters can be
found here, and an older tutorial of ours here.

Two characters often of interest are the copyright symbol
and the trademark. Here is how to recode a font that
includes these secret symbols so they can be reached...

/bd {bind def} def

/IsChar {exch/CharStrings get exch known}bd
/MapCh{3 -1 roll/Encoding get 3 1 roll put}bd
/MapDegree {dup 16#b0 exch/degree IsChar
   {/degree}{/ring}ifelse MapCh} bd
/MapBB {dup 16#a6 exch/brokenbar IsChar
   {/brokenbar}{/bar}ifelse MapCh}bd

/reencode{findfont begin currentdict dup length
dict begin{1 index/FID ne{def} {pop pop}ifelse}
forall/FontName  exch def dup length 0
ne{/Encoding Encoding 256 array copy def 0 exch
{dup type/nametype eq{Encoding 2 index 2 index
put pop 1 add}{exch pop}ifelse}forall}if pop
currentdict dup end end/FontName get exch
definefont dup MapDegree MapBB}bd

/LATENC [ 0/grave/acute/circumflex/tilde/macron
/breve/dotaccent/dieresis/ring/cedilla/hungarumlaut
/ogonek/caron/dotlessi/fi/fl/Lslash/lslash/Zcaron
/zcaron/minus/.notdef/.notdef/.notdef/.notdef/.notdef
/.notdef/.notdef/.notdef/.notdef/.notdef/.notdef/space
/exclam/quotedbl/numbersign/dollar/percent
/ampersand/quotesingle/parenleft/parenright
/asterisk/plus/comma/hyphen/period/slash/zero/one
/two/three/four/five/six/seven/eight/nine/colon
/semicolon/less/equal/greater/question/at/A/B/C/D
/E/F/G/H/I/J/K/L/M/N/O/P/Q/R/S/T/U/V/W/X/Y/Z
/bracketleft/backslash/bracketright/asciicircum
/underscore/grave/a/b/c/d/e/f/g/h/i/j/k/l/m/n/o/p/q/r
/s/t/u/v/w/x/y/z/braceleft/bar/braceright/asciitilde
/.notdef/.notdef/.notdef/quotesinglbase/florin
/quotedblbase/ellipsis/dagger/daggerdbl/circumflex
/perthousand/Scaron /guilsinglleft/OE/.notdef
/.notdef/.notdef/.notdef/quoteleft/quoteright
/quotedblleft/quotedblright/bullet/endash
/emdash/tilde/trademark/scaron/guilsinglright
/oe/.notdef/.notdef/Ydieresis/.notdef/exclamdown
/cent/sterling/currency/yen/brokenbar/section
/dieresis/copyright/ordfeminine/guillemotleft
/logicalnot/hyphen/registered/macron/degree
/plusminus/twosuperior/threesuperior/acute/mu
/paragraph/periodcentered/cedilla/onesuperior
/ordmasculine/guillemotright/onequarter/onehalf
/threequarters/questiondown/Agrave/Aacute
/Acircumflex/Atilde/Adieresis/Aring/AE/Ccedilla
/Egrave/Eacute/Ecircumflex/Edieresis/Igrave/Iacute
/Icircumflex/Idieresis/Eth/Ntilde/Ograve/Oacute
/Ocircumflex/Otilde/Odieresis/multiply/Oslash/Ugrave
/Uacute/Ucircumflex/Udieresis/Yacute/Thorn
/germandbls/agrave/aacute/acircumflex/atilde
/adieresis/aring/ae/ccedilla/egrave/eacute/ecircumflex
/edieresis/igrave/iacute/icircumflex/idieresis/eth/ntilde
/ograve/oacute/ocircumflex/otilde/odieresis/divide
/oslash/ugrave/uacute/ucircumflex/udieresis/yacute
/thorn/ydieresis]def

LATENC /_Times-Roman /Times-Roman reencode

The copyright will now code as \251 if it is available. The
trademark will code as \256 after you change "/registered"
to "/trademark above.
|
Be sure to embed all fonts during distilling.

More on similar topics in our PostScript library.

We always try to be prompt.

No matter how long it takes.

It sure would be nice if you could press a magic button
and get a list of all available fonts for your machine,
and press a second one to generate a font sample
catalog of what you have immediately usable.

There seems to be a way to do this in Acrobat 9, but
as near as I can tell so far, the process is extremely
user hostile.

Create a page from OCR using ClearScan. Then
select Tools --> Advanced Editing --> Touchup
Text. Then select any block of text. Then right
click on Properties. Then Text. Then scroll down.

Presto. A complete list of every font in every
directory that you preplaced in Distiller under
Font Locations.

The only tiny problem is they seem to want you to
"look but don't touch". Click anywhere else on
screen and Poof gone. Even a screen grabber gives
the wrong answer. No obvious copy and paste.

Hint: Fonts with hyphens in them are likely to be
PostScript. Those with commas are apparently
Truetype. The rule, of course, is to NEVER use a
Truetype font!

The only ways I have found so far to extract this list
in machine readable form are extremely gruesome.

Such as memorizing them one or two at a time and
hand entering them to Wordpad. Or using Cunieform
and clay tablet to make a list by hand. Or typing them
into a laptop or a second computer.

I would think the Acrobat SDK should let you build a
plug in that you can C-Compile to provide this handy
feature. Or possibly I am missing something here.

Obviously, this is an Acrobat feature that certainly
should be provided.

After further testing, there is an alternative that
should work but doesn't: The Acrobat Typewriter tool
also gives you a list of fonts that are slightly easier
to extract. But they only give you Windows XP fonts
and clearly omit any fonts in Adobe directories!

Ferinstance, the Stone font has been installed in my
machine, but the Typewriter feature misses it totally.

This surely is a mistake or a leftover from earlier
versions. For no way would Adobe intentionally
leave their own fonts out in favor of Truetype.  

The RTL Cookbook free upload to our new eBook
library directory can be previewed here. Much more
of it should go up in a few hours, and a few days for
completion.

The results turned out surprisingly good for a pre
word processing book. We scanned to Paint at 600 DPI
and then used Adobe's ClearScan for conversion. The
intermediate Paint step gave us flexibility for retouch
and whatever. At 600 DPI, ClearScan performance is
superb, leaving us with 99 percent or better searchibility.

This low budget approach gives us file sizes of 30K per
page, and figures that, while not as good as can be done,
are fully legible and reasonably pleastant to view. The
halftones turned out reasonably viewable and not
overly murky.

There is a deskew feature in Acrobat 9 that is quite
good at realigning text. But it has some possible
side effects that can cause problems.

Ferinstance, deskewing can make murky halftones
even darker and even worse.

If you scan to Paint instead of Acrobat 9
( usually a VERY good idea ), any repaired
image artwork may climb or lean on you after
deskewing or have "Hershey Bar" chinks
in what are supposed to be horizontal or
vertical lines.

Finally, if you happen to have a black or color border
such as the last cover of our RTL Cookbook, the
deskewing may add ugly slanty white lines.

The rule is to make your scan as accurately aligned
as possible so that Acrobat 9 does not do anything
at all in its deskew pass. Sadly, the finest rotation in
ImageView32 is limited to one degree increments,
and thus not quite good enough to be useful here.

Also, any image retouching is best done as a second
pass AFTER deskewing. This can be done by exporting
as TIFF, reworking in Paint, and re-OCRing using
ClearScan.

Mathematicians usually accept that 2+2=4. But this
may not be true for extremely large values of two.

Managed to get some very meager but apparently
accurate info on the Allen Resevoir. Which suggests
to me an utterly fascinating posible relationship between
it and the ( presumed ) prehistoric Allen Canal.

Many thanks to Nicole Spence Gibson and
Michael Johnson of the Arizona Department of
Water Resources for their valuable input.

First the apparent facts:

The dam was built by the Soil Conservation Service in
the 1930's using labor from the PWA and possibly
the CCC.  This was part of the "Gila Project".

In 1948, the overflow pipe became clogged and
nobody bothered to fix it. The dam failed spectacularly
twenty years later in the late 1960's after a heavy storm
overflowed the emergency spillway .

The local rancher/farmer involved was Geral Claridge.

It is not clear why the resevoir was named for a
different local historical clan. Use seemed to be
mostly as a large cattle tank and for recreation,
especially duck hunting. Plus flood control.

No modern irrigation control structures seem
associated with the dam.

The dam held quite a bit of water most of the time,
and even unlikely "water sking" was mentioned
as a common use. Curiously, the watershed for
the Allen Resevoir is quite dry and quite small,
being only a few square miles at most. There
are no obvious springs or access to Mount Graham
snowmelt or runoff. Heavy monsoon storms would
be limited to a week or two in July.

Followed by my rank speculations:

To me, the obvious cause of the dam's demise was
neglecting to clean the overflow pipe for twenty years.
The investigation of the failure appears perfunctory
and cursory at best. Specific immediate causes
of the failure appears to be poorly consolidated
materials and improper moisture control during
the dam construction.

Plus, just possibly, a very proud gopher.

To me, the ONLY credible water source for such a
large resevoir would have been cleaning out and
reusing the prehistoric Allen Canal! As sourced
from a Spring Canyon perennial reach.

The canal itself seems to have been completly covered
by the western abutment of the dam. Without any
attempt at piping, control, or preserving flow in any
manner. Obliterated at a level significantly above
the lake bottom.

A careful study of the canal immediately above
the dam might resolve some of these issues.
This area is fairly difficult to get to.

A plausible explanation for the name disparity is
that a pioneer member of the Allen clan diverted
an "old ditch" to create a duck pond or stock tank.
In the same manner that the nearby Robinson Ditch
did so to the east. Long before the dam construction.
And presumably, before the Hawk Hollow tank.

My present belief remains that the Allen Canal is
in fact prehistoric. And that its significant cut
is in fact world class. But that stronger proof is
still clearly needed.

Made the Thesis upload smaller and more compatible
with earlier Acrobat Readers. I'm also going to call
this #111 in the GuruGram library. Mostly for some
housekeepig reasons.

Which we shortly should have updated, along with
its sampler link.

Next major project in the ebook woks is the RTL cookbook.
Hopefully followed by the ISMM.

Uploaded a copy of my thesis to our Classic Reprints
Library. On a deep penetrating metal locator.

On one level, this was an outright scam to con Goodyear
Aerospace to pay Arizona State University to bribe Ziff
Davis to buy a story off me. Which they all did.

The thesis defense consisted of holding the metal locator
near a blackboard rail. Where it said "beep". No further
comment was needed.

On another level, this was one of the first published practical
uses for linear integrated circuits. At the time, "real"
linear IC's were outrageously expensive and hard to use.
So we faked them by creating differential amplifiers by
biasing cheap RTL logic gates into their active region..

While a plain old TRF receiver design that only hinted at
some signal processing opportunities, it completely blew
away existing vacuum tube designs. And did work
reasonably well at finding deep pipes.

Yeah, the upload looks kinda hokey. But I wanted to
preserve the original historical context and flavor. At
the time, most thesis figures had to be done using
traditional pen-and-ink drafting. So I faked mine mostly
using printed circuit tape and instant transfer lettering
instead.

The actual construction project appears here.

An excellent video on what caving is all about can be
found here.

Many years ago, I was involved in a major cleanup
project in earlier portions of this same cave. We had just
discovered that a weak acid solution seemed to work well
in removing the carbon from carbide lamps from formations.

Speaking of which, the Central Arizona Grotto is having
its 50th anniversity (!) dinner here this Saturday October
2nd. Anyone with an interest in Arizona caving is welcome
to attend.

Particularly if you have a copy of the UAAC Songbook.

Yesterday's trip extended the Allen Canal a few hundred
feet more into the sheet flood damaged area. And, as
usual, generated more questions than it resolved.

About 12,000 feet remains to be explored in eight areas.
These include...

Tracing an obvious route from takein to
Hawk Hollow tank.

Resolving the transition from CCC
spillway to prehistoric continuance.

Determining which direction the canal
leaves Allen mesa.

Finding the continuance south of Allen
Resevoir.

Resolving the route through the flood
damage area.

Finding a water source for the Jernigan
Canal.

Finding iron-clad absolute proof that the
Culebra-like "big dig"is in fact prehistoric.

Determining the north terminus of the
Allen Canal.

You can email me if you want to participate or add your name and email
added to our continuing interest list.

Should you decide to contribute a much needeed Draganfly, you
can ship it to 3860 West First Street, Thatcher, AZ, 85552. .

More on all this here and here.

Many thanks to all of you that attended our Prehistoric
Hanging Canals of the Safford Basin lecture yesterday.
The turnout was surprisingly good, considering some of the
many competing events were giving away free food.

You can email me if you want your name and email
added to our continuing interest list. "Field mice" are
definitely needed to continue mapping and exploration.

Normally, graduate students are used for field mice.
They are not at all endangered and nobody ever gets
emotionally attached to them. They are quite durable.
Surprisingly, many of them are even housebroken.

The presentation can be viewed here with its sourcecode
here. The original paper is found here with its sourcecode
here. A related Dr. Neely paper is found here, with his book
access here.

Some limited info on the Allen Dam is hopefully on the
way to me.

Which is expected to reveal this is a combined SCS
and PWA project from the 1930's that failed spectacularly
in the late 1960's. It is also expected to not be in any manner
irrigation  related.

Supposedly, the original drawings are available in
Phoenix. But they are on parchment and not readily
made web friendly. I'll try to repost what is available
on this. Probably to our Tinaja Questing or Gila
Dayhikes libraries.

Presumably the west abutment of the dam ran right
over the Allen Canal without any provision for access
or flow maint. Which reinforces my belief that the
Allen Canal is in fact prehistoric.

It IS expected that the prehistoric Allen Canal was usurped
as a water source. This is the only feasible water supply
consistent with long term storage. Otherwise, the watershed
in only a few square miles at most. With no obvious access
to any springs, streams, or mountain runoff.

For obvious reasons, I am going to call the huge canal
cut just northwest of the resovoir Culibra.

Although the amount of dirt moved is comparable to
the Marijilda aquaduct, the sheer size of this project
demands that exceptional effort be made finding absolute
and unquestionable proof of its prehistoric origins.

A sledgehammer method of post-editing Acrobat PDF
files is this: Output to PostScript. using Acrobat 9 Pro.
Opaquely write over the problem areas. Then redistill.

This is particularly handy to retouch or replace images.
After exporting of a single page to PostScript, get into any
old editor or  word processor and find the showpage line
followed by %%Trailer.

Add lots of white space just before showpage. Then insert
your overwriting code, using either raw PostScript or
my Gonzo utilities.

Here's some image replacement code I used for our
recent Thesis Project...

/jpegimageprocwithlink {
      % hoffset voffset hres vres
save /snap2 exch def
/infilename exch store % passed pix file
/inurllink exch store % link filename
/photoscale exch store
/vpixels exch store
/hpixels exch store
translate % optionally adjust position
inurllink
setareaurl % autolink sizing
/DeviceRGB setcolorspace % pick model
0 0 translate % set page position
hpixels vpixels scale % magnify unit square
photoscale dup scale
/infile infilename (r) file def % establish read file
/Data {infile /DCTDecode filter} def % define source
<< % start image dicationary
/ImageType 1 % always one
/Width hpixels % JPEG width in pixels
/Height vpixels % JPEG height in pixels
/ImageMatrix [hpixels
0 0  vpixels neg 0 vpixels ]
/DataSource Data % proc for filtered JPEG
/BitsPerComponent 8 % color resolution
/Decode [0 1 0 1 0 1] % per red book 4.10
>>
image % call the image operator
snap2 restore} def

/setareaurl { % for auto include routine
/cururlname exch store
mark % start pdfmark
/Rect [ 0 0
hpixels photoscale mul
         10000 div % remove to activate link
vpixels photoscale mul
        10000 div % remove to activate link
]
/Border [ 0 0 0] % [0 0 0 ] = none;
/Color [ .7 0 0 ]
/Action <</Subtype /URI /URI cururlname>>
/Subtype /Link
/ANN % annotation type
pdfmark % call pdf operators
} def

/new37_image {save /af1 exch store
200 200 5100 6580 0.118
     % xpos ypos hpixels vpixels scale
(https://www.tinaja.com)
     % url first smallified box for here
(C:\\Documents and Settings\\don 2\\Desktop\\
Thesis\\newmicro.jpg)
jpegimageprocwithlink
af1 restore} store

0 0 new37_image % this does it

If you are retouching or improving an image, first export
it as .JPG. Make sure this is both legal and ethical.

Our usual reminder that current versions of Acrobat do not
let you manipulate disk files unless you first run the program
by doing an Acrodist-F from the command line.

This method will result in a longer file length, and some issues
involving transparency, links, and OCR may remain.

Custom consulting available.

Just got an email from a friend of some Fort Thomas residents
who have some strange rock alignments on their property. Could
these be prehisstoric canals?

Sorry to disappoint, but these are almost certainly totally worthless
CCC busywork boondoggles from the 1930's that are almost ( but
not quite ) as bad as today's failed economic stimulus ripoffs.

These are usually piled six deep in Fort Thomas. Examples are easily
found on the Black Rock road near the old dump, or on Runway 32
of Eden International Airport. Or especially on bajadas to the North.

Just in the unlikely case the find is real, have them go to Acme Mapper,
and center the cursor on their location, do a page link, and then email
me the result .

The usual rule is: "If it is obvious, it is CCC."

More here.

There is a fundamental mismatch every time you try to use
a CFL compact fluorescent lamp in a track lighting fixture.

Most of the light goes out the side and gets absorbed by a
usually black or dark inside portion of the shade. Leaving
you with horrible efficiency and a largely dark room. Plus
hot fixtures that can raise safety issues.

These were intended for PAR parabolic reflecting incandescent
lamps that output nearly all of their light out the front.

One workaround is to try and paint the inside of the shade a very
bright white. But PAR shaped LED's will be an obviously better
choice as their price drops.

BTW, it is a myth that CFL's are always cheaper than incandescents.
Ferinstance, a closet light will never breakeven. And a case can be
made that CFL's are only clearly cost effective when run a
minimum of four hours a day.

The deli was unable to collect their aviary bill, so
they took a tern for the wurst.

Yesterday's trip pretty much proved that the Allen Canal
is in fact sourced by an apparently still perennial reach of
Spring Canyon. While there might be some possible modern
rework or maint, it remains primitive and totally free of
any concrete, rebar, or other modern constructs.

But many questions remain. Especially a strong and thorough
proof that this structure is in fact world class stone age prehistoric.
Most evidence to date clearly supports this conjecture.

Also remaining unanswered is the relation to Hawk Hollow
Tank, Allen Resevoir, Jernigan Canal, and exactly where
this seven mile long (!) canal is in fact ultimately headed. Plus why
there are at least four different architectural construction
styles.

I'll be talking about our hanging canals in this Saturday's
Discovery Park lecture at 6:30 on September 25th.

A new directory of free online college and university courses
can be found here.

Finally figured out how to find text problems in Adobe
ClearScan. The obvious "Find First OCR Suspect" does
NOT work. Instead, an indirect but highly useful process
can be used.

Start with Tools > Advanced Editing > TouchUp Text Tool.
Click on the text you want to test. Then left click and drag
to highlight the converted text words. Or use Select All.

Any uncolored word has not been converted. A word with one
or more narrow white breaks in it has been converted, but will
not be searchible. A word with a white letter background has
only been partially converted.

Alternately, you can right click and chose properties > Select Font
and then pick any available system font. Preferably one that is
obviously different in size and shape than your original.

Your result should be all of the "good" ClearScan in the new
font and the "bad" ClearScan in the original font.

With a decent initial document and enough resolution, there
should only be one or two error words per page. Get your
bitmap back into Paint and retouch whatever it was that
made ClearScan barf. Usually a subtle change in spacing,
some blackfill, a white edge retouch, improved kerning, or
substituting a known  good character will work.

Then redo your ClearScan OCR text recognition. If you have
too many errors per page, you will have to rethink your
original document or its scan resolution.

Even with attempted repairs, you may not be able to get
100% of your words searchible. Espeially if the text sample
is too short. Usually it is best to let a few slip by. But you
definitely should get all words readible.

This technique also should let you replace with a "perfect"
font. Some experimenting with font sizes will be needed,
and fill justification will likely not work very well.

Updated and expanded our Gila Valley Day Hikes page.

We may be doing a lecture on this in the November
Discovery Park series.

Latest GuruGram #112 is on a Hanging Canal Slideshow.
Its sourcecode is found here, and additional GuruGrams here.

Some pr from my upcoming presentation,,,

================================================

PREHISTORIC "HANGING CANALS" SUBJECT OF NEXT
SATURDAY'S DISCOVERY PARK FREE LECTURE

Some recent archaeological discoveries have revealed an astonishing
series of prehistoric canals just South of the Safford area. These
mountain stream fed canals have the remarkable property that
portions are literally "hung" along the steep edges of remnant
Quatranary bajada mesas. Some as much as ninety feet above
their valley floor.

At least seven hanging canal systems are known. Their total
explored length now exceeds eighteen miles. Other amazing
features of these canals are that some include above-grade
aquaduct portions. Others provide for elaborate switching to
route water between different drainages. And most include
a "breakaway" feature that makes flood repair fast and simple.

The stone age technology involved seems well beyond world-class.
Reaches of at least some of these canals still flow to this day.

Some engineering details on these hanging canals will be
presented as a talk in the continuing fall Discovery Park lecture
series by local author Don Lancaster. This free presentation will
begin at 6:30 PM on Saturday, September 25th in the Discovery
Park Jupiter Room.

The lecture may be previewed at https://www.tinaja.com/glib/hangcan1.pdf
and at https://www.tinaja.com/glib/hangshow.pdf

Everyone is invited to these family-oriented talks. Students of all
ages may qualify for extra credit. 20 inch telescope viewing and
free planetary flight simulator rides may also be available. Discovery
Park is located near the corner of Discovery Park Blvd. ( aka 32nd
street ) and 20th Avenue in Safford, AZ.

For more details, contact Discovery Park Dean Paul Anger at
(928) 428-6260 or email paul.anger@eac.edu .

We try to have a daily entry into this WHAT'S NEW ?
blog. But if I turn my back for an instant, we get a week
behind.

Also, it is usually much easier to do three or four entries
at a time. So, we tend to "backfill" occasionally.

Please always check back a week or two for any new entries.

A very interesting low cost book scanner can be found here.
And bunches more on lower cost DIY scanners here.

And info on the Scan Tailor here.

Scanning remains best if you can destroy the book or if
your document is already on separate pages. The Clearscan
OCR in Acrobat 9 is exceptionally useful during the
conversion process.

Othewise, a  high resolution digital camera can sometimes
( with care ) be a better choice.

How high is high? A true one megapixel camera imaging a
flat and correct 10x10 inch B/W area would give 100 DPI resolution.
For 200 DPI, you would need four megapixels and for 300 DPI
you'll need nine megapixels. Or 16 Megapixels for 400 DPI and
you can clearly forget about 500 DPI.

But color processing involves a resolution robbing Bayer Filter.

And other factors besides raw resolution may determine the
camera's image quality. Bottom line: Use 12 megapixels miniumum
and try several cameras to see what works for you.

I long ago decided that Powerpoint totally sucked. And wrote
my own PostScript PowerPoint emulator. That offered totally
open fast and compact single file loads, ridiculously better
graphics quality, full searching, device independence, auto
linking, photo click/expand, and total freedon from .GIF hassles.

Among other benefits.

You can find the basic tutorial here. With supporting info on our
Gonzo Utilities here. And flashing extensions here. And full
animation here. And .JPEG conversion here.

And our latest use example here. With its sourcecode here.

I thought we migh review the "click-to-expand" image code,
based on this tutorial ...

/jpegimageprocwithlink { % hoffset voffset hres vres
save /snap2 exch def/infilename exch store
% grab passed pix file
/inurllink exch store % grab link filename
/photoscale exch store
/vpixels exch store
/hpixels exch store
translate % adjust position for final figure ???
inurllink
setareaurl % autolink sizing
/DeviceRGB setcolorspace % pick color model
0 0 translate % set page position
hpixels vpixels scale % magnify unit square
photoscale dup scale
/infile infilename (r) file def % establish input read file
/Data {infile /DCTDecode filter} def % define a data source
<< % start image dicationary
/ImageType 1 % always one
/Width hpixels % JPEG width in pixels
/Height vpixels % JPEG height in pixels
/ImageMatrix [hpixels
0 0
vpixels neg
0
vpixels ]
/DataSource Data % proc to get filtered JPEG
/BitsPerComponent 8 % color resolution
/Decode [0 1 0 1 0 1] % per red book 4.10
>>image % call the image operator
ypos snap2 restore /ypos exch def
} def

/setareaurl { % for auto include routine
/cururlname exch store
mark % start pdfmark
/Rect [ 0 0

hpixels photoscale mul
vpixels photoscale mul
]
/Border [ 0 0 0] % [0 0 0 ] = none; [0 0 2] = debug
/Color [ .7 0 0 ]
/Action <</Subtype /URI /URI cururlname>>
/Subtype /Link
/ANN % annotation type
pdfmark % call pdf operators
} def

And here is how you use it..

/frey1_image {save /af1 exch store
0 0 1000 800 0.5 % xpos ypos xres yres pixscale
(https://www.tinaja.com/images/frey1.jpg) % url first
(C:\\Documents and Settings\\don 2\\Desktop\\
gurugrams\\112_hangslide\\frey1.jpg) % then link
jpegimageprocwithlink
af1 restore} store

One gotcha: The horizontal and vertical resolution must exactly
match that of your original .JPG images. A totally plowed pix
means you got these wrong, while a slanty one means a
minor error in your horizontal resolution setting.

Consulting services available.

Acme Mapper can sometimes be a useful substitute for
a GPS receiver. If you do not have differential GPS
available, it can even give you more accuracy!

Even at its "rural" resolution, you are looking at 200 feet
in 70 pixels, or about three feet per pixel. At the "city"
resolution, you have 20 feet in 70 pixels, or well less
than a foot. But the actual data is usually somewhat worse.

The cross in the middle of your image will give you its
GPS location. These are easily extracted and emailed.

You can also flag various GPS locations. But note that
there is apparently a bug in which the letter sequence
of the flags may change unexpectedly.

Properly converting an older printed document from a
thick bound book or a weak Xerox copy is a lot more
complicated than it seems. I'm in the process of
web formatting my thesis and most of my earlier
books and found some rude surprises along the way.

Simply scanning will not hack it. After scanning, you
first should go through some background improvement
and other retouching. After that, you will need some
elaborate OCR. This can make your file sizes much
smaller and allows at least partial full text searching.

Getting into Acrobat PDF format also, of course,
lets you magnify figures, have text read aloud,
combines pages into full documents, and provides
reformatting and other benefits.

Adobe has a feature in Acrobat 9 called Clearscan.
Given a good enough and a high resolution enough
copy, it is utterly amazing what it can do. On the
other hand, it still seems buggy and a still work in
progress.

Clearscan attempts to average out all the character
defects and creates a new font that looks significantly
better. But I cannot seem to get the Find First OCR
Subject "missed stuff" to work at all.

It also would be super nice to be able to edit your final
font and improve it. Particularly if an old typewriter
was weak in dotting "i" characters. Being able to
globally mark backgrounds would also be super
useful to get rid of Xerox speckle and artifacts.

A minimum of 600 DPI resolution is recommended for
scanning. Using plain old Paint to get from document
to initial bitmap seems to work well. Particularly
since you can easily improve backgrounds before
continuing. And even attempt to repair the
worst of the lettering.

Needless to say, a squeky clean scanner is a must!

One hint, on Paint at 600 DPI, you cannot view the
entire page at once and may miss some background
problems. Put a separate copy into Imageviewer32 and
adjust its size to whatever is convenient.

It is a good idea to give Acrobat 9 several similar
pages at once to work with. This way, it has a
reasonable number of character samples to use.

You can follow the progress of the thesis project here.

I see something enormously distressing taking place on
the web. Completely ignored and of profound significance:

Only history that has a champion survives the web!

Ferinstance, you can easily find the rules for Mad
Magazine's 43 Man Squamish from their June 1965
issue. .

But Sallen and Key's crucial 1955 paper defining the entire
world of Active Filters is not freely available anywhere!
Heathkit schematics are readily available from dozens
of web sources, while details on automotive diagnostic
interfaces are few and far between.

In other words, trivialities survive while essential defining
documents go begging.

One obvious solution: MANDATE that all scientific papers
older than three years have free web access from multiple
sources! It is way past time to stake the gatekeepers to an
anthill.

Note that most published scientific and quite a few technical
documents are paid for with your tax dollars. YOU own
these and should have every right to freely access them.

A second obvious solution: Restrict copyright to 36 months
renewable once upon careful review to an individual, and
to 18 months nonrenewable to any corporation or heir. With
manditory free web conversion afterwards.

Yeah, I am trying to not be part of the problem. Many
classic reprints appear here, with others in the works.

Latest GuruGram #110 is on Remastering Video for Web
Distribution.

Its sourcecode is found here, and additional GuruGrams here.

The National Inventory of Dams did approve my password
after a few days. There is no information on the Allen Reservoir
in the directory.

What is really amazing is that there is apparently ZERO
overlap in county, state, and national dam directories.

I am mildly curious: Even with "oops, it broke", how do you lose
a 500 foot long, 30 foot high dam with a lake behind it?

And hide every written document on it anyplace ever?

And instill in the locals not the slightest clue nor the least
interest that it exists? Even when you can almost throw a rock
at it from downtown? Its only a short 4WD trek from
Thatcher International Airport.

Conspiracy enthuasiasts might note that this is where they
buried the UFO's before faking the moon shot.

We get lots of requests for efficiently drawing a smooth
curve through multiple dots.

This is amazingly more complicated than it seems. Especially
when the dot positions are beyond the capability of the math
used, when continuity with other dot groups is needed, when
the dots represent fuzzy data, or when fast and efficient
computation is required.

I thought we might list a few resources here, and possibly
eventually make a GuruGram out of it.

A key doncument on solving the problem with cubic splines
can be found here. This is not easy reading.

A cubic spline intro tutorial can be found here.

We have long had some curvetracing routines in our Gonzo
PostScript utilities. These were based on using weak and
inefficient splines, one per dot pair. They were rather
painful and tedious to use in that you had to enter a X value,
a Y value and a slope at each point.

It occurred to me that you could sometimes fake the
slope by using the previous point, the present point,
and the next point to temporarily draw a circle. The
slope at the present point is often a good fit. Some
more details of this improvement appear here.

It is almost always possible to draw a cubic spline
through four points, as shown here. Since there
are usually an infinite number of splines that can
do this, an optimization process can be used. This
is often the shortest spline as well.

Things get considrably uglier when trying to fit
fuzzy data with lots of points to a single cubic
spline. Some details are here on a least squares
optimization routine. Some of the points may be
missed per the details here.

Details on finding the minimum distance from a
point to a cubic spline appear here.

And on the length of a cubic spline here.

And on subdividing a cubic spline here.

You can always draw a power series through n
data points, but the behavior between  points can
end up surprising, and certain limits ( such as
overshooting 1 or 0 ) may arise. Per the details
you'll find here.

Many of the approaches may need solution of
nxn linear equations. Per these details and
these complications.

I seem to be having an outrageously difficult time finding
anything at all about Allen Reservoir, a spectacularly failed
and quite large dirt dam two miles southwest of Thatcher.

Yeah, there's lots of web sites that lavishly praise the fishing
at a spot that has been bone dry for decades. And dozens
more will give you the exact GPS locations.

The geneologcal historian for the Allen clan never heard of
who built it or why. The historical society president is in the
hospital. The SCS ignores completly my local and national
emails. And the national dam directory has locked me out
because they apparently think I am a terrorist intending to
further destroy an already completely destroyed dam.

All I am really looking for is proof that a canal the dam
completely obliterated and ran over at its western abutment
is in fact prehistoric. The CCC does not seem to be involved
as there is waay too much dirt and not the slightest trace of
fancy rockwork.

There is a thirty foot hole in the dam, likely caused by a
very proud gopher. I suspect the owners do not want to
own up to it, because they apparently destroyed a perfectly
good seasonal duck pond in one of the usual government
boondoggles.

I suspect the houseboat franchise may be available.

Please email me if you can help on this .

Found two new good books that realistically cover alternate
energy. Exposing, among other things, the hydrogen ludicrosity
and the outright corn ethanol scam.

Check out Vaclav Smil's Energy Myths and Realities.
And Robert Bryce's Power Hungry; The Myths of
"Green Energy".

I don't care for the nuclear option at all. Because (A) We
tried it and it did not work; (B) Whether it is a net energy
source at all remains unknown until the decomissioning
realities and exactly how the results of the next five or six
Chernobyls are going to pan out. and (C) Solutions that use
current solar flux are unlikely to cause long term warming
problems compared to solutions that add old sources of
energy to the present heat flux budget.

Much more here.

Leave it to the Military to point out to me via email
that a Epiphany is usually overwhelmingly positive,
while a "Holy Shit Momement" can be either very bad
or very good.

Some personal examples...

Way back when I was first developing our
curvetracing routines, I decided to use
Mickey Mouse as a subject. Halfway
in the development process, about 2:15
AM, I entered ONE wrong coordinate.
And Mickey instantly got an appropriate
sized and positioned erection! Sorry, but
for obvious reasons, code ( with or without )
is not available.

I witnessed a tanker rollover in Phoenix
that created the usual movie-style fifty
foot fireball. In the length of time it took
to say "Holy Shit", the fireball blew itself
out, leaving the rest of the zillion gallons
of diesel fuel below its flashpoint.

For decades I had been finding tantalizing
hints that Magic Sinewaves were in fact
real. Sure enough, at 3:15 AM a waveform
popped up with all low harmonics precisely
zero. The Holy Grail was found!

Finding a continuance of the prehistoric Allen
Canal in an unexpected place, superb
preservation, and world class size was a
recent example.

But the most recent "Holy Shit Moment" was
realizing exactly WHY the prehistoric obsession with
hanging canals halfway up the sides of mesas:
This made the slope INDEPENDENT of the terrain!
And, factoring in the available technology, makes
the LBT look like a tinkertoy set.

Latest GuruGram #109 is on Resolving Stability Issues in
Gauss-Jordan Equation Solutions. Its sourcecode is found
here, and additional GuruGrams here.

More on Gauss-Jordan equation solutions here.  

The Central Arizona Grotto of the NSS is having a fiftieth (!)
anniversity bash. Due to a groundswill of popular demand,
I was asked for some comments. Which I'll repeat here...

==================================

I guess Bee and I have been long term CAG members since
1967 or so. We were attracted to the Grotto by an Arizona
Highways story. We had done some minor caving in Bear Cave
and elsewhere in Pennsylvania and wanted to know more about
Arizona's Pivot Rock cave.

I was never a gonzo supercaver. That award surely should go first
and foremost to Jerry Hassemer. Or at least to his dog Suzy.
With such illuminaries as Ron Bridgemon, Pete Delaney,
Bob Buecher, Dwight Hoxie, Tom Strong, and Lang Brod
( and dozens of others ) certainly in the running.

Over the years, I was often the grotto vice president and edited
the Cave Crawler's Gazette newsletter for three years. In those
days, properly doing a newsletter involved an awesome amount
of time, effort, and cost. I still hope to get these newsletters up
on the archive of my website at https://www.tinaja.com . A classic
cover involved the DISMAL I, a "Diet Smith" magnetic car
levitator that really was Pete Delaney and a pair of crutches
standing in a garbage can. Overlayed on an impressive cliff. Such
photo edits were a lot harder before PhotoShop.

We often hosted grotto meetings, sometimes involving a "who
turned out the lights?" drunk Saint Bernard. One of the grotto
meetings was literally sixty five feet in the air on the Gentry
lookout tower near Heber.

While we regularly attended CAG events even after moving first
to Parker and then to Thatcher, my primary "loyalty" was always
to the ARA. As I felt that outreach from the cave community was
(and very much still remains) a crucial issue. We probably hosted
more ARA paper regionals than anyone else, mostly at EAC or TFD.

My biggest Arizona cave "find" was the extension Jerry Hassemer
and I made to Dum Ditty cave. I remain convinced major wet passage
remains beyond its present terminal crawlway.

Finding virgin passage remains one of the ultimate highs of the
caving experience. Out of state, I was involved in finding thousands
of feet of virgin passage in California's Lilburn Cave. But this was
largely a trivial task at the time. Bee and I also did some fuzzy
elephant research in Natural Trap cave in Wyoming as part of a
scientific expedition. Where I managed to blow up a cave pack
full of spent carbide. We also played a negligibly minor role in
Kartchner development.

My focus was often on the Redlake area. We did quite a bit of
pumping both in Redmond and Bear Sprinhgs. I am convinced
there is a lot of potential remaining in Bear Springs and in Hot
Dog pit. There still is a largely unchecked lead half a mile
northeast and above Hot Dog in a slanty entrance that just might
lead to big time new cave. The crucial issue of the area remains
Pishiboro heading South to the Frog Pond with Columbine's
water source remaining an enigma. Bear Springs may remain
the key to the puzzle.

One major effort for little return was the search for Mescal Pit.
Dozens of trips and one regional were made to the Sombrero Butte
area, eventually finding a minor 95 foot pit in precambrian limestore.
A second unfound "rocks roll forever" pit, tentatively named
"Strawberry Awful", is rumored a mile to the northwest. You can't
get there from here. This did eventually lead to the exploration and
mapping of El Diablo. Which remains in the most spectacular
scenery in Arizona and involves rather challenging hiking.

I was also overly enameled of El Capitan canyon. While only
"almost" a caving experience, it remains a "must visit" for any
caver. The "next" El Capitan Canyon is likely to be San Carlos Falls.

There are many long term legends of Arizona caving. But four of the
more persistent are the "hall tree", "Bee and the flamethrower",
the "watermelon caper" and the "jawbone of the ass". Seems Greg
Lazear had a hall tree he no longer wanted, so he hid it in a remote
and secret Arizona cave. Other cavers found it and rehid it elsewhere.
Over and over and over again. Simply referring it as "it" and carefully
never mentioning what it was or its history. It probably has been in more
caves than anybody and probably recirculates to this day.

Bee was in Peppersauce Cave once and kept asking "Is my lamp lit?"
At the time her carbide light was putting out a foot long flame. Turns
out that caving with prescription sunglasses on can be suboptimal.

Students of Dave Thayer knew he was planning a Grand Canyon river
trip the next week, so they backpacked and preburied a watermelon,
pineapple, crutons, a silver fingerbowl, and, of course, beer in the sand.
Only to lose track of time and having to sleep on the Tonto Rim in garbage
bags on the way back up. One participant who was brand new to caving did
not notice anything out of the ordinary when they handed him a watermelon
and asked him to backpack it. The prank ended up an overwhelming success.

A jawbone of an ass was found at the bottom of the Grand Canyon and
placed in the pack of a known volatile caver. Who was furious about finding it.
After SIX TIMES of refinding it in his pack, he violently threw it over the
redwall. Sure enough, it ended up back in his pack and it made it all the way
to the rim. Requiring heroic physical effort on the prankee.

But the ultimate Arizona caving legend had to be the UAAC songbook.

As people age, their endurance and flexibility wanes. So we are more into
papers, meetings, support, and parties these days rather than gonzo exploration.
Some of our current projects include studying the Mount Gram Tramway
https://www.tinaja.com/glib/tramshow.pdf, compiling a directory of Gila Valley
day hikes https://www.tinaja.com/gilahike.shtml, and exploring an utterly mind
blowing group of Safford Basin prehistoric canals. Details on the latter at
https://www.tinaja.com/glib/hangcan1.pdf .

It is important to remember that most rocks can be classified as sedentary,
ingeneous, or metaphoric. The most crucial issue facing Arizona cavers today
is continuing test and development of the caver's wrist sundial.

And, of course, never store carbide in a nonlocking carabiner.

To paraphrase Buckaroo Banzai, no matter where you go, there you are.

This has been an unusual year for desert wildflowers locally.
Some say the best in forty years.

Fall flowers are often relatively rare and minor. But we now
have a spectacular bloom of something. Possibly Isocoma
Tenuisceta. Who their friends call burroweed, shrine jimmyweed,
or burrow goldenweed.

What is mind boggling is this: In two smaller areas, a very
conservative calculation leads to ONE HUNDRED BILLION
(with a "B" !) blooms.

They have been around for several weeks now.

This borders on "several" or "an adequate supply".

Another Sonoran wildflower directory can be found here.

How can you tell the difference between an "epiphany" and
a "Holy Shit (!) momemt"?

Found the ADOT Proposals to reroute US70 here.

Again, there is not the slightest mention of any
prehistoric considerations. The proposed routings
may impact many of the hanging canals.

Oops. Back in February we did a fairly major upgrade
to our Fundamental Factors Underlying Technical
Innovation.

This made it to Wesrch ok, but we somehow missed
updating the GuruGram library. The new tutorial
should now be downloadable.

One sure sign that you are into serious four wheeling:
When the entire windshield is blue.

The latest report on pv solar panel prices is found here. Where we
see pricing holding steady at $3.50 per peak panel watt for the summer.

This pricing level, of course, GUARANTEES that pv panels today
remain a humongous gasoline destroying net energy sink. This price
is SEVEN times that required for pv panels to become an utterly
pointless "paint it green" transfer scam. AKA the "holy grail"
of one dollar per fully burdened installed system watt.

And are NOT in any manner renewable nor sustainable.

For net energy breakeven at ten cents per kilowatt hour amortization,
and genuinely becoming a useful energy source, the price has to drop
by a factor of FOURTEEN to twenty five cents per peak panel watt.

An interesting pv solar trade journal can be found here.

More detailed analysis here.

Found another small chunk of the Allen Canal. While
there still are some "minor" gaps of 15,000 feet or
so, evidence is finally preponderating that this in fact
is one entity. And quite likely is prehistoric.

Running over TEN KILOMETERS (!) from Spring Creek
to the Jernigan site. The middle half is astonishingly large.
But mostly in "badlands" style terrain that is fairly easy to
dig or scrape. As before, the engineering is astounding.

Age hints are accumulating. There are full size barrel
cacti growing in the wind blown fill in the canal center.
A spectacularly failed flood control dam ran right over
the canal without any regard to access. As did an
apparent CCC Hawk Hollow Tank project. Apparently
very old fencing slices through twice without concern. As
does a major and historic dirt road.

There flat out were not enough historic pioneers to take
on a task like this. And they certainly would have mentioned
it in their carefully kept famiily records. While the prehistoric
population is estimated to be near the present of 30,000.

And it would be much easier for an early historic rancher or
farmer to"dig out an old ditch" than it would to engineer a
world class project from scratch in their spare time.

Proof of age and continuity remains elusive, but should be
forthcoming.

Just had a nickel-and-dimeing noncustomer read us the
riot act for using prioity mail for shipping.

Other shipping methods ( such as parcel post ) are too
slow and too unreliable to even consider if you are serious
about running an eBay business. Not to mention their
staggering hidden costs of needing to wait in line at the post
office, compared to Priority's instant online processing .

Or that our prices including shipping are much less than
anywhere else without.

Our real customers lavishly praise our delivery speed.
There is no way we will favor a troublemaking nickel and
dimer over proven customers. This epsilon minus, of course,
has now been permanently blocked.

We do combine for the lowest possible eBay shipping costs
where and when possible.

Much more here.

Another factor that may help "prove" the hanging
canals are in fact prehistoric: What was the Home Deopt
in Thatcher during the 1880's likely to have in stock?

Well, mule rentals fer sure. Plus scrapers, picks, shovels,
prybars, block & tackles, wheellbarrows, cement, headgates,
and all sorts of blacksmith odds and ends.

Nothing at any of the hanging canal sites suggests any use
whatsoever of these obvious historic tools and materials.

The mesmerizingly awful Google Analytics Bug we looked at
a few days ago can apparently be eliminated by use of a
Firefox add on named TACO.

This seems to work quite well.

The September 2010 issue of Wired Magazine has the usual
photo of the Apple I on page 134. Few people realize that it
is my keyboard you are looking at.

A reprint of the original construction project can be found here.

Woz apparently used our updated version that had LSI added.

Yeah, the one-chip solution had fewer parts, but something
elegantly simple was lost with the original "encoderless
encoder". Not to mention the extra power supply.

Other classic reprints here.

An interesting new scheme for homebrew laser cutters
can be found here.

By moving the focal point up and down through the material,
a much lower power laser can be used. This apparently
works particularly well with acrylic plastics.

More discussion here.

Similar stuff in our Santa Claus Machine library.

Found another large chunk of old canal. As with most
any scientific discovery, it raises more questions than it
resolves.

IF this is a continuance of the prehistoric Allen canal, it
is huge, has a different architecture, and appears well east
of where it was expected. But it IS a hanging canal.

If historic, CCC, or modern, it makes no sense at all
( even for CCC who specialized in the utterly pointless )
because there is no credible water source except for the
prehistoric Allen Canal .

The low end is somewhat closer to the Jernigan site and
intermediate topography remains barely slope favorable.
But the befuddlement factor is clearly raised.

Much more here.

As we have seen here, a key major advance in
publishing in general has been to typeset first and
edit second . Rather than the other way around.

I very strongly feel that the solution to the current
scholarly publishing debacle is that we should always
publish first and peer review second. And NEVER the
other way around.

Such as the Wesrch Model.

If something is wrong with what you said, you fix it.

Instead of the present system that involves years of
delays, gatekeeper greed, and journals so expensive
and so elusive that even most libraries cannot afford
a subscription.

ALL scholarly papers over three years old should be
freely and instantly web available from multiple sources.
Without any access restrictions whatsoever.

Failure to do this will result in traditional scholarly
endeavors pricing and otherwise restricting themselves
out of the market. And out of business.

Fortunately, there are moves underfoot to guarantee
government funded research to become freely available
after a reasonable quarantine delay.

The old peer review model no longer works. Live with it.

Updated and expanded our PostStcipt Sampler page and
our Classic Reprints Sampler pages.

Other sampler links are found here.

Just uploaded two RE ASCII Keyboard Encoder projects
here and here to our classic reprints library.

Revised and updated our Prehistoric hanging Canals of the
Safford Basin.

The source and sink of the Allen Canal is taking form but
remains uncertain. It appears a CCC "steal the plans"
rebuild/adaption of Hawk Hollow Tank straddles the original
prehistoric canal. About 12,500 feet remain to be visited.

The hanging part is sort of vague in that its route is less
steep than the original canyon bottom. It is, however, very
well defined. And accomplishes the same tasks as a true
hanging canal route would.

Turns out that nearly all of the hanging canals are newly threatened
by a proposed realignment of US 70 well south of its present
urban route. A major alignment study appears on the web, but
I cannot refind it. Apparently it was a scanned copy that does not
allow search terms. Prehistoric considerations are not even
mentioned.

These are world class archaeologcal features that very few
seem to have the faintest clue over. A pair of poorly thought out
water tanks already ran roughshod over the Twin Boobs canal and
trashed many grids and mulch rings. Plus a CCC project.

There is a terrible, terrible Google Analytics bug that is
highly likely to drive your website viewers away in droves.

The symptom is that any visitor sees an unwanted and certainly
unrequested white screen that is trying to link Google.
Sometimes the screen waits forever; other times there is
a delay of minutes to hours.

Sometimes, if you are quick enough, you can click on the
"stop loading" X and view the underlying page. Naturally,
your back arrow will not work either. This is, to say the
least, is maddengly infuriating.

There are apparently three factors that cause the bug:
(1) A website host failing to place the analytics link at
the VERY END of their website code. (2) Google response
being slow and creating a 403; and (3) Firefox gagging on
a Google 403.

Naturally, everybody will blame everybody else. Meanwhile,
it is your website that gets trashed.

More on the bug here.

One workaround is to use the Chrome browser for any
problem site.

Several published workarounds do not seem to work at all.

Meanwhile, DO NOT USE GOOGLE ANALYTICS ON
YOUR WEB PAGE! There is no conceivable benefit
that can remotely approach the damage it is dong to you!

Just read a fortune cookie:

"Help! I am trapped and being held prisoner in a
Fortune Cookie factory".

The EAC Discovery Park lecture series is apparently going
to continue.

I've got a Prehistoric Hanging Canals paper scheduled for
September 25th and a 365 Gila DayHikes one for November
13th.

I still have a mesmerizingly awful problem involving
Acrobat Distiller. Apparently the history of XP "run"
use can lock out any access to "Acrodist-F".

Sometimes it eventually recovers after a week or
two and other times it seems permanent. It usually
applies only to one user and not another.

The latter is essential for most of our Gonzo stuff
and anywhere else you want unrestricted diskfile
access.

Yes, the machine has had its registry cleaned and
has been antiviruesed.

The onlly present workaround is to use another user.
Which means a painful double switching everytime a
simple change is made in a .PSl file.

Note that the obscure "windows key - L" command
lets you switch between users.

Please email me with your suggestions.

We now have a pair of lower res Introduction to PostScript
remastered classic videos up on YouTube.

< http://www.youtube.com/watch?v=W6kD_SwKGlY > is part1.
< http://www.youtube.com/watch?v=_ui1TIgapcI >  is part2.

These are in the 320 x 240 iPod format. Apparently YouTube
now lets you have 2 gigs, 15 minutes, and up to HDTV,
so we will be experimenting with at least classic NTSC
resolution of  640 by 480. Only progressive.

We'd still be stuck with the original live VHS copy of a
copy, but we should be able to significantly improve all the
stills. Details whenever.

I may have found a credible explanation for why the prehistoric
canals were all purposely hung on the edges of steep sided
mesas:

Doing so gives precise control of canal slope INDEPENDENT of
terrain!!!!! And minimizes cuts and fills.

Which would ( and probably does ) represent utterly brilliant
engineering.

The gently sloping top of a mesa makes a suitable canal
surface, while the steep sides let you precisely control the
initial slope. Taken together, these might form an unbeatable
combination for transporting water long distances.

The latest revision of our Introduction to PostScript video
should be close to a final cut.

As mentioned before, about three minutes of downloading
is involved because of ISP delivery limitations. Our first
attempt at You Tube failed because we have to split the
half hour vid into two fifteen minute pieces.

Again, we chose .MP4 format for convenience, its wide
audience, and tiny file sizes against image quality.

Much, of course, has happened since the first release of
this video two decades ago. Manic cutting edge enthuasism
for Cobol, Pascal, or the Vax seems to be somewhat waning.

And Kroy Kolor long ago went belly up caused by pricing
themselves out of the market, reliability issues, and
color inkjets running away with all the marbles.

But the big positive development in PostScript has been
that Acrobat Distiller makes a superb host based
PostScript interpreter. Greatly improving the convenience
and ease of use of the language.

Much more here.

Found another hanging canal! Now up to at least eight.

This one has been well published as the Jernigan Site.

Its a lot shorter than the others but somewhat wider.
Apparently they bought a "large" from the canal
factory instead of the usual "medium".

It also does not hang nearly as much, ending up "only"
twenty feet or so up in the air. Total length is about
a quarter mile and it is used as a "local" to link two sets
of fields. Rather than an "express" to deliver distant water.

At one point, the cut is over a meter deep! This is the
deepest I have seen and involves a tremendous amount
of hand labor.

Two interesting stunts could improve video editing.
We might call these scripting the script and editing the
editor.

Naturally, you should save your edits often. You should
also do as much with the edits as possible, avoiding any
and all reuse of MP4. For edits of an edit can lead to
blotchy titles, poor sound, and other problems.

Yes, generation loss is much, much less with digital
video. But it is not gone completely.

At any rate, a plain old Wordpad reminder script can
keep track of which saved version does what. This is
particularly useful for preserving start, end, and length
points for latter clip inserts and upgrades.

Thus scripting the script.

The .vpj files created by the VideoPad editor are simple
plaintext lists of clips used, time sequence, and sound
mods. These are surprisingly short, typically only a
few K long. They are easily read and modified with
Wordpad or any other text editor.

Hand modifying a .vpj file could have lots of interesting
uses. Ferinstance in changing a video clip name for
an update. Or upgrading a .JPG to a true .BMP. Or
changing a directory for a new but similar project.

Thus editing the editor.

I'll eventually try to work up a GuruGram on all this.
Other GuruGrams are listed here.

I'm having troubles proving such is so with the prehistoric hanging
canals. At present, the Allen Canal delivered water somewhere
to an unknown location and purpose. The Jernigan Site
obviously needs a major unknown water source.

The two are 5000 feet apart with a reasonable slope between
them. But with three intermediate stream crossings. The area
is hard to access partly because of a locked gate on a dam and
is patrolled 24/7 by roving bands of Gila Monsters.

There are possible other Jernigan sources and Allen sinks.
None of these seem compelling. Sheet flooding could have
long ago complicated the picture.

I sure could use some field mice on this project. email me
if you are available.

A newsgroup poster questioned why I had to switch my
membership from the Gurus and Swamis Union local
#415 over to #204.

This had to do with the Godzilla Versus the Night Nurses
cross genre classic. Because of the restraining order
from the Tapioca Pudding Institute, the movie release had
to be pulled and reissued directly to 8-track.

Back when our PostScript class was doing resumes, one
student wondered "what if" one single individual was responsible
for each and every major disaster or debacle over the
previous few decades.

And how a positive spin would be put on it.

Their "proven ability to command a strong media presence" would
be  based on their Edsel ad campaign, tanker safety coordinator for
Exxon,  fire prevention manager for Yellowstone, Bhopal tank
maintainence director, new  Coke creativity group leader, Cherynoble
training supervisor, BP safety containment officer, etc... etc...

This might make an interesting cloud project. Please email me
with one or more of their employment experiences.

Org. The initial title page of our new Intro To PostScript
video was looking awful. And, amazingly, did not get any
better with a full bitmap rather than a .JPG still.

It turns out that MP4 files have to have at least four
frames before they can get up to speed on any change.
And your first frame may look awful if it is just sitting
there waiting for your viewer to hit the play button.

The workaround is to start with a nondescript frame of
pretty much a single color, fade into your title background,
and then fade to the actual title.

Note that all this activity has to preserve the exact sound timing
of the original. The "uncouple sound" feature of Videopad
can be handy here.

Flying an aerial platform is ( literally ) quite a trip!

Added a shift register tutorial to our Early Classics Reprints
library.

I've been fighting a really obscure bug involving running
the Acrobat Distiller on Windows XP. Not sure if it is only
my machine, or something others can use. But here goes...

As we have seen here, Acrobat Distiller makes an
outstanding host based general purpose PostScript
interpreter.

Starting with Acrobat 8.0, Adobe has default locked out most
disk access on Distiller. Bypassing some obvious abuse
potential.

Their published workaround is to run "acrodist-f" from the
command line. This restores full disk access to Distiller.
Unfortunately, their early docs wrongly told you that acrodist-f
disables, rather than enables access.

At any rate, acrodist-f normally works just fine for enabling
and is essential for most of my PostScript-as-language apps.

But it would stop working for me for weeks at a time but only
for certain XP users. The bug and the cure may go something
like this: The windows command line apparently will refuse
to process acrodist-f if the previous command resulted in an
error!

An apparent solution is to type HELP from the command line.
Or some other known good thing for it to do. This seems to
restore proper operation.

Our newly digital remastered Introduction to PostScript
video still has a remaining glitch or two.

We are working with our ISP on a better .MP4 delivery scheme.
So far, you have to download the entire half hour video before
you can start to view it. This takes four minutes or so, depending
on your comm rates.

The original artwork for most of the inserted clips can be
found here as our PostScript Show and Tell series.

Again, this stuff may seem dated, but it was outstandingly
cutting edge revolutionary when it came out.

Here's a brief summary of how to get from an old
VHS video to online distribution...

0. Be sure you have lots of disk space available.
    Make sure any free software has been upgraded
    to allow MP4 saves and eliminate other hassles.

1. Convert the VHS tape to a computer readable
     digital format, possibly by using a dubbing
     dual VCR/DVD machine.

2. Convert to .MP4 format using Leavo or similar software.

3. Make sure all artwork clips end up at the
    needed final resolution. One good route
    with hires originals is to crop and reduce
    to 4X, blur strongly with Imageview32,
    reduce to 1X, and gently resharpen.

4. Edit and produce your .MP4 video using
    Videopad or a fancier program.

5. Upload to a suitable website or YouTube.
     Make sure your participant.

Our neighbors stopped selling Cowboy Jerky because
it was too tough and stringy.

So they switched to beef.

Here's some of the "figure improvement" techniques we are
using in our ongoing update of the Introduction to PostScript video.

Much of the work is being done with plain old Paint and Imageview32.
Helped along by our Bitmap Typewriter.

We purposely picked the iPod resolution of 320 by 240, first because
of the challenge of getting good lowres results, and second to keep
the file sizes down. It is very important that all stills and titles have
the native resolution you are seeking.

It is also super important to minimize any and all high frequency
info in your images. Fail to control this, and the results will be awful.

We typically would start with a high resolution original and crop it as
tightly as possible. Also considering a pan or a double shot to show
what is needed. Some tints or color were added wherever they could
indeed add to the result.

The hires original was then reduced to FOUR TIMES the needed
final resolution. Typically this would be 1280 pixels wide or 960
pixels high. If the original was not in a 4:3 format, the critical height
or width gets used.

Now for the sneaky part: You have to get rid of higher frequencies
or the results will look awful. Add some very significant BLUR to the
4X image! Perhaps four or more clicks. Then reduce 4:1 to 320x240
and gently resharpen.

If the final is too narrow or two short, get into Paint and click on
Attributes. Reset to 320x240 and recenter. Always stay in
the .BMP format till one final .JPG save.

In general, the blur-resize-sharpen dramatically makes light lines
darker and heavier. And even eliminates many dropouts.

The 4,752 perspective bricks in the bulding proved to be a problem
and created severe Moire issues. This was gotten around by
replacing the bricks with an indinstinct lowres texture. Taken in context,
this does not look too bad.

Should have some viewable results shortly.

Thatcher Fire Department is about to take delivery on
a 95 foot aerial platform.

This means we can now hunt with the big dawgs and
no longer have to stay on the porch. It also means we
are no longer hopelessly outclassed by our newer three
story hotels, dormitories, and auditoriums.

Less well known and less appreciated is the fact that an
aerial ladder or platform can go down as well as up.
Which should greatly ease incidents on the wrong side of
a canal or behind a hard-to-climb fence.

We have had numerous incidents involving crossing the
Norton Lane bridge. The usual tiny problem is that we
have no Norton Lane bridge. Leading to post flight
awkwardness.

Updated and expanded our Gila Day Hikes page.

A sick person checks in with the witch doctor in darkest
Africa. Who takes a rawhide strip and boils it in all sorts
of vile concoctions. And tells them to eat two inches of it
a day till it is gone.

Person comes back in a week and reports they ate it all
and they are feeling much, much better. But still are
not quite cured.

"I guess the thong has ended, but the malady lingers on"

Found some supporting evidence that suggests the Allen
hanging canal is in fact prehistoric. Or at least "fairly old".

In three separate places, there are large barrel cacti growing
directly in the middle of the filled canal path. Considerable time
would likely be needed for the canal to fill, germination to take
place, and the cacti to grow to these large sizes.

It might be interesting to study the fill particles to try and see
whether they are wind or water borne. If wind based, this would
likely add considerably to the age of the last use.

It still is not clear where the canal is headed. Part of the area
to the north is largely alkali and more or less badlands. And
the one known habitation site is pretty much to the north and
west, with several drainages in between.

Total length of the canal seems to be approaching five or more
miles. Construction style and energy levels definitely seem
consistent with prehistoric. And the number of manhours
involved boggles the mind.

As does the world class engineering.

Here are some of the steps we are going through in our
making our older Intro to PostScript video available online:

The analog VHS tape first needs converted to a digital
format that is computer readable. While there are quite
a few adaptors and systems available, the easiest is to
go with is a Magnavox combination VHS/DVD dubbing
recorder ZV427MG9 . These are available from
Walmart for just over $150.

It is super important your choice allows DVD recording
from VHS. The magic word is "dubbing" The output
will be a computer readable DVD/R with files in the
digital MPEG2 format.

The "best" modern format for online distribution is
probably MP4. A $30 MP2 to MP4 converter
is available from Leawo and seems to work just
fine for us.

Next, a MP4 resolution has to be chosen. We decided
to go with the highly limited but popular 320x240 format
since it is iPod compatible, has small file sizes, and
gave us quite a challenge in reformatting stills and
titles.

The low resolution is not quite as bad as it sounds.
Compared to NTSC (Never the same Color) tv,
it allows wall-to-wall useful info. Because there is
no wraparound and no retrace. Plus much better
chrominance/luminance separation since there
is no longer any luminance. And, of course, no
copy-to-copy degradation.

It is super important the all stills and titles have
exactly the needed resolution! Otherwise they will
look awful. More details in a day or two.

A video editor that is newly MP4 compatible will be
needed. We found the free Videopad editor somewhat
limited but more than useful for this current project.

The ISP provider on your website may have to make
a change or two to allow video downloading. Such
as giving .MP4 file extension permission. And
YouTube may have length limitations. We are not
quite there yet, but it should be only a day or two.

There seems to be a new attempt at reviving Popular Electronics
Magazine. While I remain singularly unimpressed by this venture
so far, you might find it of interest here or here.

Bits and pieces of the old PE lie scattered around the landscape.
I have many of my early reprints here. Jeff Duntemann has
revived the Carl and Jerry Series here. Michael Holley has
much of the SWTP stuff here. And one of their editors remains
alive and well in a nearby town.

The sixth hanging canal is real! Haven't found the
hanging part yet, but it seems to be over five kilometers
long (!) and its characteristics seem totally consistent with
other nearby bajada prehistoric finds.

Size is the usual meter wide by ten or twelve centimeters
deep. We can call it the Allen Canal for now. Have only
explored half a mile or so in the middle of the bajada, so
it is still  unclear where it comes from or where it goes.

Or what its ultimate purpose was.

Best present guess is that it sources somewhere near
Hawk Hollow Tank or Upper Central Wash and delivers
somewhere near Allen Reservoir. Some parts may be
barely visible on Acme Mapper. But remain unchecked.

Sure could use some field mice to help with the
exploration. This could easily become world class.
email me if you are interested.

Digi-Key appears to be by far the worst offender in not
having the faintest clue what the difference is between a
radial and an axial component is. And then spreading "not
even wrong" info about them.

As is intuitively obvious, and as any 1929 radio book
will tell you, an axial device has its leads going out its
axis: A radial device has leads going out its radius.

Detailed image examples appear in this tutorial.

Two popular forms of electrolytic capacitor are the
single ended axial in which both ends go out the
bottom and the double ended axial in which one
lead goes out the left and one out the right.

Radial electrolytic capacitors are exceptionally
rare. It is highly unlikely you have ever seen one or
that Digi-Key has ever sold one.

It  still continues to amaze me how many people
try to "correct" our eBay listings while adamantly
making making utter fools of themselves.

Guy enters Mercy Hospital in Brisbane, keeps complaining about
the afternoon tea. The staff finally suggests a native tea made from
the hide of a koala bear.

He than praises the excellent taste of the tea but complains about the
floating hairs and fat in it.

"I'm sorry sir, but the koala tea of Mercy is not strained."

We are in the process of converting and uploading our
original Intro to PostScript video both to the Guru's Lair
and YouTube.

Boy, things sure have changed. This was originally done
three years before the initial Video Toaster release.
Video at that time was outrageously crude and expensive.

Especially -gulp- any frame by frame editing.

We started off by finding a hungry rural video outfit that
specialized in vids of people falling off horses. The
best available cheap reasonable quality at the time was
a system called Industrial Betamax. Analog, of course.

The taping session ended up poorly promoted, so it wound
up being single camera shot with an audience of only three
people. After the shoot, we hijacked an entire community
college English class and had them pass around all the
handouts and pretend they were interested in them.

We managed to include such advanced state-of-the-art
special features as zooms, video titling(!), awful music loops, and
special effects such as inserted still images(!). Wowie Gee!

The video house ended up losing our master tape. So, all we
had  left was a fairly mangy VHS copy to work with for the
current processing.

So, yeah it is kinda hokey. But it remains a historical record
of the early days of PostScript. The tools and techniques
shown were utterly mind blowing revolutionary at the time.

We are trying to improve the video by shortening it to
only the PostScript Show and Tell part and redoing the
titles, stills, and possibly the horrible music. The portions
on getting started with PostScript, toner cartridge refilling,
Kroy Kolor, and binding systems may become additional
"historic record" videos if there is enough interest.

We are only halfway through the conversion process, with
everything now digital and in computer. Full details will
show up here soon when we get everything worked out.

After several years of outstanding service, Harry Swanson
has retired as director of Discovery Park.

The newly selected director is Paul Anger. His phone number is
(928) 428-6260 and his email is paul.anger@eac.edu.

It is not yet clear whether the Saturday Night free lectures
will continue.

Extending our Magic Sinewave calculators to higher
values beyond n=23 are likely to create problems with
Gauss-Jordan stability issues.

Stability is caused by coefficient blowups that eventually
can easily get far beyond the range of even 64-bit math.

One sledgehammer cure might be to go to 128-bit math.
Since only subtraction, division, and translation are needed,
the routines might not end up too ugly.

High nxn blowups are very much dependent on the column
and row positioning in the original equations. Messing with
the colums gets ugly fast, but rows can be interchanged in
any order without changing the solutions.

A Monte Carlo approach might work. In which you shuffle
your rows several dozen to several hundred times and then
pick one with the least coefficient buildup.

I've been exploring a different route in which you progressively
find the best next row position. Eventually creating a list of
optimal or near optimal row sequences. You then do the
Gauss-Jordan reduction on the best choice on the list.

At any time you will have a candidate row and a column of
interest. The coefficient in the candidate row gets subtracted
from each remaining row to create a distress value. The
distress value gets normalized in the next step, sometimes
causing coefficient blowups.

A distress value of +1 or -1 will cause no coefficient blowup.
An absolute distress value less than one will cause blowups.
The smaller the distress value is, the worse the eventual blowups.
Because of the unity normalization in the next step.

It is not immediately clear what an absolute distress value greater
than one will do, but it usually should reduce blowups. Further
the odds are good that any extremely small normalized value will
get "swallowed" in its upcoming subtraction.

The algorithm might proceed as follows: The distress value is
modified by abs dup 1 lt {1 exch div}{1.0} ifelse creating a
nastyness factor of unity or higher. The nastiness factor
is calculated for all remaining rows and least square summed
and square rooted. Creating a potential blowup index.

The process gets repeated for each candidate row and the
row with the lowest blowup index goes on the sequential
row list. I should have some demo code shortly .

Got several requests for more info on color organs.
Reprints on several classic construction projects can
be found here. And good old number one here. I am
working on eventually uploading quite a few more.

But, except for rock shows and night clubs, these are pretty
much obsolete. Since you can do so much more so much better
with a computer based display driving an HDTV.

There's also a bunch of technical issues that were never
properly solved with the previous batch of color organs...

DYNAMIC RANGE -- The dynamic range of
lamp brightness is ridiculously less than that
of most music. An automatic level control or
automatic gain control would help bunches.
As would very heavy audio compression.

LINEARITY -- Dynamic range is further
compromized by extreme nonlinearity of
most light bulbs. One cure might be table
lookup linearity correction. In which incremental
(or log?) changes of input level produce some equally
incremental changes in output light.

RFI -- Astonishing levels of radio frequency
interference are produced by classic color organ
systems. This can be reduced by extensive inductive
filtering on the power lines. Another and possibly
better solution is to go to reverse phase control
techniques where the light goes on at the beginning
of each power half cycle and then is turned off
in inverse proportion to the audio strength. This
demands power MOSFETS rather than SCR's
or triacs.

DISPLAY ISSUES -- Getting fully saturated colors
is tricky unless you go to lasers or LED's. Heat
management can be a problem. And the static
light positioning can easily and quickly become
boring. This is likely a fatal flaw.

AUDIO FILTERING -- Early active filters could
certainly be improved, and digital filters would be
even better. But best of all would be elaborate
computer signal processing in which each instrument
would be assigned a color, rather that colors being
frequency specific.

I'd be happy to assist others in developing a new generation of
color organ related devices, but it is extremely unlikely that the
interest and the funding would be there.

Updated and expanded our GuruGram Library.

I think we have most of the site rebuild complete.

Please email me with reports of anything strange.
Especially problems in downloading longer PDF files.

I think we have a 100 percent recovery. But, just in
case nobody mentioned this to you before, ALWAYS
BACK UP EVERYTHING SIX WAYS FROM SUNDAY!

And do so often and consistently.

Could there be a sixth hanging canal? Some tantalizing
hints showed up on a Acme Mapper satellite image west
of the Frey Mesa Road. Both the location and the slopes
are credible.

It is a tad warm to run out and check this instant, and the
area is hard to access. But, if real, the hanging canal
systems would approach an astonishingly staggering
30 kilometers! This area would also seem pristine in
the sense of zero pioneer or CCC overwork as well.

Meanwhile, some minor grid and mulch ring finds North
of the Mud Springs area strongly imply a Ledford to
Lefthand occupational continuium.

All this stuff has apparently just been sitting there for
eight centuries. And nobody seemed to notice. Piles
of rocks can hardly be subtle.

Some of our current goals and projects include working off
a horrendous backlog of eBay listings, getting our Intro to
PostScript video both on our website and U-Tube, starting
a release of "all we have" in thumb drive format, and
beginning the ebookization of the ISMM.

The rest of the books and most of the remaining early
classic reprints hopefully will follow.

Please email me with your needs and requests.