On Thu, Feb 09, 2012 at 02:24:13PM -0800, James Adcock wrote:
Greg> PG uses Dublin Core, with some additional namespaces.
We are confusing catalog storage with book storage. When PG sends a say HTML book out into the world the catalog information does not travel with the book. In the case, as an example only, with an EPUB book (properly coded) the book information *does* travel out into the world with the book. I suggest again if PG is in the business of preserving books then it should send the whole book out into the world, not keep some parts of the book only in the PG catalog.
We have partial metadata in the text/html. Are you talking about adding DC metadata in what is visible? Or, encoded but not necessarily visible? Either seems pretty achievable. One slight snag is that some of the metadata is only added after the book's files are pushed to the server (notably, LCSH). Note, however, that the RST *does* include all the metadata (as I pointed our in a section of the message you cut). Getting that same data into derived formats would be quite easy. -- Greg