- From: Karen Coyle <kcoyle@kcoyle.net>
- Date: Thu, 12 Aug 2010 08:37:29 -0700
- To: Ross Singer <ross.singer@talis.com>
- Cc: "Young,Jeff (OR)" <jyoung@oclc.org>, William Waites <ww-keyword-okfn.193365@styx.org>, public-xg-lld@w3.org
Quoting Ross Singer <ross.singer@talis.com>: > Karen, I generally agree with this summary but would amend it slightly. > Instead of: > > "You code an mp3 as a song, not as a file." > > I'd say: > > "You code an mp3 /first/ as a song and then as a file. Then link the file > resource to the song resource (but probably not the other way around)." I agree. Even better, I should have said that you *catalog* an mp3 as a song (symphony, whatever), and you include information about the physical format as the "carrier" part of the cataloging. We're back in the "content v. carrier" discussion that is part of the library cataloging discussion. (This came up a lot in early discussions of digitization, where some folks wanted to catalog the digital file and list the person who did the digitizing as the author. In the end, no one found this a useful approach.) Not that this is universal... in some contexts, the carrier is primary (this can be true some times in archives, I believe). And note that I am only referring to recorded information here -- of the kind that libraries manage. I think that if we insist that our metadata models can work for *any* kind of metadata, we'll set ourselves an impossible goal. Scientific data (genome, etc.), data sets (census, etc.) and other things that may be covered by metadata have other requirements and a different set of user behaviors. We shouldn't compromise service to our users in an attempt to stretch our metadata beyond our own boundaries. What this gets down to is the goals you have for your metadata. I don't want us to lose site of the goal of serving information seekers. I should emphasize that these are *human* information seekers. While we need our metadata to be processed by machines, the machine is a conduit, not the end-user (at least in my context). kc -- Karen Coyle kcoyle@kcoyle.net http://kcoyle.net ph: 1-510-540-7596 m: 1-510-435-8234 skype: kcoylenet
Received on Thursday, 12 August 2010 15:38:07 UTC