Bowerbird@aol.com writes:

david starner said:

the catalog just ain't gonna help someone know what kind of book they'd like to read. (and a marc record won't help them either.)

that's a job that collaborative filtering will eventually do much better than anything you can do in the form of a catalog of any type.

But this won't be of any help to brick and mortar libraries who want to integrate PG etexts into their existing catalogs. MARC is best way to accomplish this. This would also let PG to offer a Z39.50 gateway to the catalog which would be very cool. I like the distributed cataloging idea, but it's not the same as DP or Wikipedia which are brilliant and making it as simple and easy to contribute as possible. Cataloging is not simple and it's not easy, and if it's not correct and consistent it will result in a mess which will do more harm than good. That said, there are a number of steps in the process that can't be easily automated which can be done in a distributed environment by people fairly easily, but these steps have to be identified and then a mechanism for people to contribute. The catalog as it stands represents a lot more effort than a lot of people realize. I hope people keep that in mind when they slam the existing catalog. b/ -- Brad Collins <brad@chenla.org>, Bangkok, Thailand