22 Apr
2005
22 Apr
'05
9:12 a.m.
Tom Jennings wrote:
On Thu, 21 Apr 2005, vrs wrote:
Split the task up, maybe by stealing some software from DP
http://www.pgdp.net/ to start with.
1. One person scans a set of fiche, and put them up on a site somewhere.
2. Someone takes a fiche, and breaks it into page images.
<for text versions>
3. Someone OCR's the images.
4. Run through a couple of rounds of proofing
5. Someone massages the individual pages into a complete manual.
<for raw-pdf version>
3. Someone merges all the pages for one document into a pdf.
My low-tech, low cost solution to this problem
was to buy a fiche
printer,
print the fiche on paper, then scan that. The quality is sort of barely
adequate.
Unfortunately, it is also very much a "low volume" solution, just barely
adequate for a manual or two. (My printer requires you to manually
align
each page.) 25000 sheets calls for something more sophisticated.
Maybe get a little more systematic? Is there any rhyme or reason
to the fiche, eg. the first images are title/doc
ID/table-of-contents pages? If so, tediously capture those, then
you'll know what's worth bothering with and what's not.