marcello said:

I don't want to unload hundreds of image files on the user.

except that's precisely what happens when you bundle them all together. the user who only wants one of the pagescans has to download them all. there are a raft of good reasons why a person might want a single scan. lets think in the pluralistic mentality for which project gutenberg is known. there's no reason you can't have one place that has the individual scans, one that has them zipped up, another that has them as a djvu (both forms), one that has them as a .pdf, and even one where they are a quicktime movie. per lockss, you want multiple copies scattered throughout cyberspace anyway. speaking of pluralism, any all-image format that can't play on a playstation, or on machines like that new nokia770 that's coming out, is a non-starter.

Can you imagine reading Ulysses from 1.000 tiff files?

why not? the person is only looking at one or two pages at a time, so the viewer-program only has to open one or two files at a time, who cares how many there are in the whole book?

Is there even any picture gallery software that can handle that gracefully?

i'm sure there is. but why use that? instead, use e-book viewer software that's geared for scans. my viewer-program can use scans _or_ text, or even _both_, in combination, for people who want to view them side-by-side (e.g., to do proofing). *** this is a good demonstration of one problem with project gutenberg: you don't have any application programmers helping you lead the way. so you are forced to settle for the least worst program at your disposal, instead of grabbing the bull by the horns and creating the best program. i would suggest that the next time a programmer like me comes along, you be nice to him or her, instead of spitting in their face. -bowerbird