----- Original Message ----- 
From: "Lloyd Fosdick" 
To: 
Sent: Wednesday, April 18, 2001 7:41 AM
Subject: Illiac manual


> I have an Illiac manual that I am preparing to make available on the web. 
> Since this is the first project of this kind I have undertaken I'd 
> appreciate any advice that you can offer. If this is successful then I have 
> some other early manuals that I will try to install along with some commentary.

I detest the usual .PDF files.
True some semiconductor houses put their small format 
catalogs into .PDF (probably with professional help)
but I generally refuse to waste my life and eyesight
fighting with most .PDF on the web.

Well, I'm searching for a better way.
I have searched for several years for a better way!

The following is the best I know - and it is
   - slow
   - painful
but I like the results
   - searchable
   - compact (fast down load for folks)
   - easy to read 
         - .html based fits screen better than any .pdf I have found
         - user can select the font size (to ease aging eyes?)
         - little if any horizontal sliding necessary
               
so - I currently put up with the drudgery.
AND am wide open to suggestions.

------ current method --------
1) Scan text in black and white and OCR using 
        OMNI PRO, Version 10
     which may be as good as any in the consumer class programs

    I tend to scan about 8 page sides, then process them
    (I get too confused if I scan too many before processing.)

    Problems with OMNI PRO, Version 10 too numerous to list
       - sub and superscripts not processed correctly
       - non ASCII characters garbaged
       - poor horizontal spacing 
       - if it does not recognize a character it is supposed to emit a tilde -
            well, it much prefers to emit garbage instead -
       - on and on past infinity   ;-((

     I manually zone the text areas before OCRing  -
     If there are two columns, I zone each column separately.
     This reduces confusion, and permits the user to  view
     the contents of a single column on the screen in the 
     font size the user chooses, usually permitting
     the .HTML and browser to do text wrap.

    Unfortunately,  OMNI PRO, Version 10 usually does not retain
    horizontal spacing well - sometimes it does - drives me nuts.

    Do the automatic spell checking on each page -

2) Store scanned text as a .TXT file 
     If stored as .HTML,  you fight font size problems, and lots of 
       other problems - like when it gets confused about font
       or font size.
     If stored as .DOC - may God have mercy on your soul.
        I cannot make WORD do anything useful to scanned
        stuff.   It sets modes such that if you try to increase
         the font size to easy reading, the tops of the letters
         get cut off - weird.   And several other problems
         to painful to describe.

3) Append to ypur previously processed text in the .HTML file

4) Insert .HTML tags as needed
     (I also insert image < IMG ... tags at this point
       even though I have not scanned the images
       properly yet.)
5) Proof read and correct as needed
6) Save .HTML file and view it with a browser
     (I used to view it with both Netscape and Microsoft
       browsers - trying to make the results look good
       with both.
       HOWEVER
     I am currently annoyed with Netscape 6.0,
         and I screwed up Netscape 5.4,  so I now 
         totally ignore the way Netscape might process
         my .HTML - let Netscape users struggle -

7) Go back to 4 above, make things "right"
8) Scan images at 300 dpi in either gray scale or color
     as appropriate -
       I currently use ADOBE PhotoDeluxe 2.0
       much prefer ADOBE PhotoDeluxe 1.0
        (it doesn't try to save stuff  into scrapbooks, ...)
       but I screwed it up and can't find the CDROM

[You may note that I have a lot of troubles with
 computers, most of the troubles I cause myself,
  which does not improve my humor :-((((     ]

9) I trim the resulting scan to contain only image
    (I could trim in the scan phase but I don't do that any more - )
10) Reduce the size of the image to by a factor of about 4
    (the PC screen thinks you scan at 74 dpi,  
        but I get more pleasing results (for me) my way.

    The image processing takes time, so I upgraded my
    processor to 700+ megahertz and an INTEL mother board 
    (to take advantage of a builtin network connection).
    What a disaster!
    The thing could not find my CDROM to read
    in INTEL motherboard code to find my other peripherals.
    I took it to the computer store, and $180 later
    about half of my system works properly.
    (Much of my e-mail files permanently disappeared.)

11) Run the sharpen tool once (the reduction seems to make
     fuzzy edges)
12) Save the image -
       name - I tend to name things by page - saves imagination
                   and is easier to trace back
       type - .JPG is SUPPOSED to be lossy, and give
                 smaller images than .GIF (no loss).
                 I tend to save both ways and throw away the larger image.
                 (Sometimes .GIF at compression 3 is much larger than 
                   uncompressed .GIF   -   strange stuff!)
               Line drawings almost always are smaller in .GIF
13) I used to make thumb nails and save user load time
      then the used could click on in interesting thumb nail
      and get the full sized image -
      (getting too tired and impatient - sorry)

14) Go back to 6)  above until you are satisfied

16) I will not let you you update the Table of Contents
     to point to your completed work

15) Take an aspirin, (or illegal drugs?  strychnine is recommended ;-)
       then go back to 1) and scan some more



> 
> Lloyd D Fosdick
> 
> 3455 Table Mesa Dr, B114
> Boulder, CO 80305-5833
> 
> tel: 720 304-9044
> 

If you proceed, in spite of the above,
please let me know so that I can link to your work.   :-))

I am also looking for a scaner that scans very close to
the center of a book.   The best I have found in 2 centimeters -
I really need one centimeter.

Good Luck,  please keep in contact  (There has GOTTA be a better way!)
   Ed Thelen