View Full Version : Epub Revision - enhanced metadata support


Nate the great
04-08-2010, 08:13 PM
Need for enhanced metadata support. EPUB 2.0.1 provides basic support for expressing embedded publication metadata. However, native support for standards such as ONIX and RDFa is not yet provided, which has a negative impact on the fidelity of metadata provision. Likewise, PRISM has article-level and inline metadata support that could integrate into EPUB.

Who wants to get this one started?

charleski
04-08-2010, 11:28 PM
PRISM appears to be more relevant to magazines than books. It may be useful in encouraging a move to epub instead of relying on custom-built apps, though.

I don't really see the point of ONIX, but then I'm not an accountant. ONIX has certainly added a few new epub-specific tags and looks like it's aiming to get integrated. But a lot of the information it covers it potentially quite sensitive - publishers and distributors aren't going to be entering discount codes and the phone no.s or email addresses of employees into an epub to be read by anyone. ONIX looks like a schema for a back-end stock-management database, I don't see why it should be part of the epub spec.

kovidgoyal
04-09-2010, 12:17 AM
series, rating and section level metadata in the OPF (all of which calibre already does via extensions).

Section level metadata: support for specifying the author, a summary and a date for every section.

Oh and support for specifying raster cover images in OPF

dmikov
04-09-2010, 12:47 AM
FB2 2.1 has pretty exhaustive metadata schema, might be worth looking into.

Terisa de morgan
04-09-2010, 03:22 AM
series, rating and section level metadata in the ncx (all of which calibre already does via extensions).

Section level metadata: support for specifying the author, a summary and a date for every section.


I totally agree, first for global book and second for anthologies (if we can ask, series and rating for every section too, with the same purpose).

ghostyjack
04-09-2010, 07:38 AM
What also might be handy is the ability to specify for a series how many books there are in it and which one this is. Therefore it should be possible for a reading device or software to extract the info and display as:

Book # of x.

Where # is the book number and x is the number of books in the series.

pholling
04-09-2010, 11:53 PM
I hope that the metadata includes attribution for a block of material, as Kovid suggested. I see textbooks becoming a blend of material from different sources crafted by the instructor to meet the needs of the current course. Whatever business model makes that work will want to know who wrote which section.
I also see metadata tallying compliance with the constantly changing landscape of standards involved for representing different kinds of information.