So far... the best results I have obtained (in experiments) has been to *photograph* the pages, then scan the photos... I am using 35mm film, but I would imagine that the newer hi-res digital cameras would work as well. I will post some of these scanned docs on my website when I get a Round Tuit. I use a page-sized piece of 1/2" glass to smooth the page... works for sheets as well as bound material, then a cheesy copy stand with some photo-grade lamps. A tripod would work too. I gave up on the OCR for the reasons set forth in the previous posts. Cheers John

For pages that consist solely of text and line art, scan them as 300 DPI TIFF Class F Group 4. That takes only 40-120K per page. I put the resulting images into a PDF file, since most people don't have any other G4-capable reader, and G4 is supported as a native PDF image format. Some people always flame me about disliking PDF because they can't run Acrobat Reader on their Commdore 64, but realisticly I've found that more people have access to Acrobat Reader than any other viewer. My attitude is that if I spend the time to scan the docs and make them available free on my web site, people that don't like it can take a hike. I've written a program using PDFlib to automate creating the PDF from a directory full of G4 files. For greyscale and color images, I'm working on a process to separate out the images, use G4 coding on the monochrome portion of the page, and overlay the images in JPEG format. This will also work nicely with Acrobat reader, since it can support overlaid images, whereas most other viewer software can't. Some results of my scanning can be seen at www.36bit.org. Note that most of those scans were done *before* I got a sheet feeder. In my experience, although there is some skew with the feeder, there is less skew than when I do the pages manually, and the skew is more consistent software. Eric