The general consensus seems to be that bi-level scanning with a resolution of at least 300dpi but preferably 400dpi (although I tend to use 600dpi). G4 encoded TIFF is pretty good space wise (obviously lously compared to text). OCR is (almost) certain to introduce errors. You'll need a significant investment in proof-reading to fix this! If you can OCR, then any format that can represent that text in whatever fonts and layout the original document used (and uses an efficient openly-documented format) should do. Most of my text documents are PDF. You can turn PDF into text (or html I guess) where appropriate. But you cannot OCR (or at least, I bet you cannot OCR without introducing errors). Antonio -- --------------- Antonio Carlini arcarlini(a)iee.org