- From: Julian Reschke <julian.reschke@gmx.de>
- Date: Sun, 27 Sep 2009 12:37:02 +0200
- To: Maciej Stachowiak <mjs@apple.com>
- CC: "public-html@w3.org" <public-html@w3.org>
Maciej Stachowiak wrote: > ... >> Not sure what counts as "dependency" specification. I'm ready to >> believe that HTML5 currently references other specs that get this >> wrong, too. But what's relevant here are the specs that define the term. > > To give a specific example, Unicode, ECMAScript and CSS all define the > term "property" in completely different ways. And HTML5 adds yet another > distinct definition in a different context. As far as I'm concerned, > this is not a problem, because it's always clear what is meant. > ... Understood, and agreed. I think everybody in the computer business is aware of the fact that the precise meaning of "property" depends on the context. > Another example: RFC3986 has a different (much more general) definition > of "resource" than the HTTP RFC. Yes, and that's fine because it defines a subset for the context of HTTP. Any RFC2616-resource should be an RFC3986-resource, otherwise that would be a bug. >> Furthermore, as explained earlier, HTML5 is inconsistent in itself; >> and that's something that should be fixed. If "Foobar" is the thing >> identified by a URL (HTML5) then it simply can't be a bag-of-bits at >> the same time. > > Natural language is context-sensitive. I don't think any actual > confusion is caused. If no confusion is caused, fine. That's something we can check. On the other hand, can we please agree that "A resource is a bag of bits" is plainly incorrect? If it would be correct than HTML5 would need a different term for the thing a URL identifies (because of, as mentioned earlier, content-negotiation and also uses other than GET, such as POST). BR, Julian
Received on Sunday, 27 September 2009 10:37:44 UTC