greg said:

I am not sure we can automate things much more than they are already,

this speaks volumes. (look, a pun!) in my ever-humble opinion, you _need_ to create files that _can_ be parsed unambiguously. and the number-one test of that is whether _you_ can parse 'em. a digital library where every book has to be handled _individually_ sacrifices far too much potential.

I expect that these types of problems will go away when we eventually start having richer (or at least better formatted) metadata in eBook files "born as XML," but with our current procedure there's a little too much variety in the eBook layout, author names/roles, and so forth to be able to create completely automatic catalog entries from the eBooks themselves.

no comment. -bowerbird