A use-case, if it's of interest: We have an online digital library
application, which contains images (jpg,png,gif,tiff), also PDFs, audio
files (mp3,wav), video (mpeg4, etc). It's widely used here at UT Austin,
mostly for instructional purposes, but also for content-managing media
rich web sites [0]. It uses Atom extensively for moving data around and
for archiving collections.
I typically get EXIF data from jpegs, id3v2 medatadata from mp3s, and IPTC
data from images when possible. I can use the content element in Atom to
capture metadata about the primary media file, but it's that metadata that
varies per media file that I need to deal with (height, width, bitrate,
stereo/mono, aspect ratio, md5 hash, etc.) in atom:link.
Yahoo's Media RSS Module [1] seems to do most of what I need, and perhaps
just using that is the best option. There was a need expressed some time
back for the Atom group to address these media file issues [2] and again
in the current thread. It strikes me that any of: 1. saying that using
Yahoo Media RSS is a "best practice", 2. "porting" Yahoo Media RSS to a
more Atom-like syntax and implementing that, 3. doing something else
entirely -- would be fine.
I really like the idea of the atom "link" element being the appropriate
starting point and working on the best set of "rel" values and a set of
agreed-upon child elements (based on Yahoo Media RSS). Of course, Yahoo
Media RSS has the idea of a "media group" (i.e. a set of files that are
different versions of the same thing) which I really like and which
should, I think, be handled. I'm not 100% sure what all needs to be done
to begin defining extensions, but I would be glad to take a stab at
"porting" Yahoo Media RSS to an atom-like syntax for others to comment on
if folks think that'd be a useful exercise.
One more point: ours is a "closed" system in many ways, so it doesn't
really matter what we do, but time and again I have folks wanting to
repurpose their materials (usually "podcasting", but other ways as well).
Being able to share collections externally and/or to ingest collections
born elsewhere would be a tremendous benefit.
--Peter Keane
[0] http://elucy.org
[1] http://search.yahoo.com/mrss
[2] http://www.imc.org/atom-syntax/mail-archive/msg11813.html
On Thu, 17 Jan 2008, James Abley wrote:
On 16/01/2008, James M Snell <[EMAIL PROTECTED]> wrote:
All it would really take are implementors willing to support the
extensions once they were defined and a group of individuals willing to
do the work to define them. I'm willing to help.
- James
I'm wary of making commitments that I can't keep, having recently had
to drop out of one community due to most of my spare hacking time
being taken up with working at the startup that I'm currently with,
but I'd like to take this forward if possible. Any suggestions for
ways to do that?
James
Sylvain Hellegouarch wrote:
Peter Keane a écrit :
On Mon, 14 Jan 2008, James M Snell wrote:
I had been involved in some discussions a while back geared towards
working on a standard set of media extensions derived from the itunes
and yahoo media extensions. Unfortunately, that work never
progressed. Such work would certainly not be unwelcome.
- James
Absolutely. It would be a huge boon for the sort of work we are
trying to do. MIT's OpenCourseWare, ITunes U, etc. are all the rage.
A Media-rich Atom standard would be a no-brainer for that area (and
with FeedSync or a FeedSync-type standard along with it, a very
powerful platform indeed).
Any thoughts on how such a process might be re-booted?
I concur and I'd join such work if it existed.
- Sylvain
