- From: Hugh Winkler <hughw@wellstorm.com>
- Date: Wed, 15 Jun 2005 08:47:29 -0500
- To: "'Mark Baker'" <distobj@acm.org>, "'Mark Nottingham'" <mnot@mnot.net>
- Cc: <public-web-http-desc@w3.org>
Yes to Mark B's comments... and to Mark N's. URI for semantics-- yes++. If the pie in the sky stuff is the machine to machine interaction I mentioned in one little sentence at the end, I deny I ever intended to address that in this forum. I agree with mnot that addressing that problem is a very long project indeed. Hugh > > Though I agree with Hugh's intent, I think his choice of words was > unfortunate for the pie-in-the-sky image I'm sure it planted in most > folks' heads. > > What we're (well, me at least) really talking about here is the > difference between this; > > <parameter name="language" /> > > and this, where clients bind to the "application type" (my term) of the > parameter, not its name. > > <parameter name="language" > type="http://example.org/types/language/" /> > > (any similarity to WADL is purely coincidental - "type" is certainly > very different) > > It's just a layer of indirection which enables a client to depend on > data which is expected to change less often. Using a URI also provides > the hook into the more advanced "semantic" features, but it needn't be > dereferenced by a non-SemWeb app (or a SemWeb app for that matter). > RDF Forms uses a URI in this way. >
Received on Thursday, 16 June 2005 01:32:01 UTC