- From: Booth, David (HP Software - Boston) <dbooth@hp.com>
- Date: Fri, 6 Apr 2007 12:12:23 -0400
- To: "Misha Wolf" <Misha.Wolf@reuters.com>, "Henry S. Thompson" <ht@inf.ed.ac.uk>, <www-tag@w3.org>, <newsml-g2@yahoogroups.com>
> From: www-tag-request@w3.org [mailto:www-tag-request@w3.org] > . . . > Henry pointed out at the Edinburgh AC meeting that if we used simple > concatenation to give people access to information about terms in > our taxonomies, we could end up with illegal fragment IDs. So: > > If: > <subject qcode="iptc:123456"/> > and if: > iptc -> http://www.iptc.org/NewsCodes# > and if we used simple concatenation, we'd get: > iptc -> http://www.iptc.org/NewsCodes#123456 > > There is, of course, the other option: > iptc -> http://www.iptc.org/NewsCodes/ > then if we used simple concatenation, we'd get: > iptc -> http://www.iptc.org/NewsCodes/123456 > > We've decided to side-step this by specifying that the concatenation > rules are taxonomy-specific and are up to the provider of each > taxonomy. So any URI-based program using multiple taxonomies must have special concatenation rules built in for *each* taxonomy? That sounds awful. Was there some reason why the group could at least recommend that the namespace part end with either "/" or "#" (along with corresponding constraints on the local part)? David Booth, Ph.D. HP Software +1 617 629 8881 office | dbooth@hp.com http://www.hp.com/go/software
Received on Friday, 6 April 2007 16:13:25 UTC