Steve Thomas writes:

That's not the point. People don't go to PG thinking, "hmmm, I wonder if they have any XML files". They go looking for a book. If you want the text of a particular book, you'll use it whatever format it comes in, so long as you have the software to handle that format. Nobody "needs" XML or PDF. They "need" the words of the book. Formats are secondary.

What if they have a page reference to the standard (or only) edition of the book? Then they "need" the page numbers. What if they have a speech synthesizer smart enough to do multiple languages, but they need the languages marked? Then they "need" language tagging. What if they "need" to process a table? Then they "need" a system that doesn't ASCII-format tables. (And, BTW, a speech synthesizer that just skips accented letters is just lame. Removing the accents could be done in one line of Perl or a dozen lines of Fortran.)

Could it be better to put the PG effort into getting plain text editions out, and leave it to others to do the extra conversion to XML etc.? This is a model that has worked really very well for quite a few years, without complaint from any but a few tech-enthusiasts.

No, it doesn't work real well. The value of XML is in what it includes that plain text doesn't, and a lot of that is lost in the plain text version. You need the original book to fix that. Even with the original book, it can be a pain, whereas it's trivial to keep page numbers (for example) in the original processing. -- ___________________________________________________________ Sign-up for Ads Free at Mail.com http://promo.mail.com/adsfreejump.htm