- From: Anthony B. Coates <abcoates@TheOffice.net>
- Date: Thu, 12 Sep 2002 10:20:30 +0100
- To: www-tag <www-tag@w3.org>
** Reply to message from Tim Bray <tbray@textuality.com> on Wed, 11 Sep 2002 15:04:06 -0700 > Wrong. There is not a one-to-one relationship between languages and > schemas. I could rev a language by changing the semantic descriptions > of how its elements ought to be rendered, without changing the syntax > constraints expressed in the schema. I could rev the schema because > it's hard to read or needs to be re-modularized or validation runs slow, > without effectively changing the language. No, not wrong at all. The point you make about semantics being able to change without the schema changing is theoretically possible, but in practice it doesn't happen like that. Semantic changes are the worst kind of change, and so are most likely to occur at a major version change. Even if they occur at a minor version change, they will occur at a version change, because otherwise the user groups would start rounding up lynching parties. I don't deny that versioning can be tricky, and maybe there isn't a one-size-fits-all solution. However, I think there are potential 80/20 solutions that would be of enormous benefit to users. Just leaving a void around this because of the difficulty will not be a help, and will just lead to versioned namespace URIs. Some kind of architectural and/or best-practice guidance is necessary if namespaces are to be used (mostly) unversioned, as many W3C luminaries seem to want. Cheers, Tony. ==== Anthony B. Coates, Information & Software Architect mailto:abcoates@TheOffice.net MDDL Editor (Market Data Definition Language) http://www.mddl.org/
Received on Thursday, 12 September 2002 05:21:06 UTC