- From: Tim Bray <tbray@textuality.com>
- Date: Sat, 10 Jun 2000 15:09:19 -0700
- To: xml-uri@w3.org
I apologize for rehashing this, but I feel I must point out that I (at least, probably others) strongly disagree with Dan's contention that the NS rec really meant to say NS names were URIs and the fact that it doesn't is a "typo" or equivalent. I believe that the consensus was that the rec means exactly what it says, that NS names are names that use URI syntax. I recall that an earlier rev of the spec provided for *two* URIs to be attached to a NS prefix; one being the *name*, and one giving the reference for the *schema*. The analogy to sgml's SYSTEM and PUBLIC identifiers was strong, for those who cares. The schema pointer was abandoned, one of the reasons was people like me arguing that the notion that dereferencing a single URI was a hopelessly inadequate bridge from an instance to the complex space of resources that might identify its semantics in whole or part. -Tim
Received on Saturday, 10 June 2000 18:10:40 UTC