Scanning microfiche?

From: CLASSICCMP_at_trailing-edge.com <(CLASSICCMP_at_trailing-edge.com)>
Date: Sat Jul 29 17:01:51 2000

*>As for the $0.07 per page, maybe Tim should indeed buy his own scan/copying
*>machine for many thousands of bucks.

Well, I already do have a fiche viewer, I was just looking at a way of
making it more widely available, and making it digital, with the intention
of putting it on modern/denser digital media as time goes on.

I can duplicate/enlarge the fiche in my darkroom, the material
cost for a duplicate 4"x5" sheet is about $0.50. The stuff prints onto
conventional photographic paper really nicely, although at about $0.50
per 8"*10" print it isn't cheap to do even a single sheet of fiche
(somewhere between $50 and $100) photographically. Those Xerox-machine
based page-printers are the real winner costwise, though I'm not real
happy with the printed pages I've had made for me in the past quality-wise.

As to the quantities, there seem to be about 250 fiche sheets in a one-inch
stack, and my stack is about 18" high. That's 250*18=4500 pieces of fiche.
And each piece of fiche is somewhere between 50 and 200 printed pages of text;
that's about 450 thousand printed pages, by my calculations, using an average
of 100 printed pages per fiche.

In my experience, a 300 DPI scan of a page of text or line drawings reduces
down to about a 30 Kbyte GIF. As Eric pointed out, there are more efficient
ways to compress images, maybe 10Kbytes is reasonable. 10 Kbytes * 450000 pages
= 4.5 Gigabytes. So it'd fit on a 8 or so CD-ROM's.

As a smaller sub-example:
The stack of fiche containing all the PDP-11 related pocket service manuals
(CPU's, boxes, terminals and peripherals) is only about 45 pieces of fiche.
They're packed pretty full, so that's about 10000 printed pages. They'd all
fit nicely onto a fraction of single CD-ROM, by my calculations, even
if it was as crude as a 30 kByte .GIF per page.

OCR would be nice for keyword searching, too...

* I would certainly kick in $1K toward such a project. My experience is that
* you _have_ to have an auto scanner, this move/frame/press stuff is just too
* hard to keep your brain focussed.

I was thinking (I could be wrong) that a scanner could take in the whole
4" x 5" microfilm sheet at once, and then I could write some software that would
parse the whole-sheet image into frames, do some automatic cleanup on each
individual frame, and compress. Am I walking down the wrong path?

Tim.
Received on Sat Jul 29 2000 - 17:01:51 BST