- From: Dave Beckett <dave.beckett@bristol.ac.uk>
- Date: Wed, 29 Sep 2004 14:25:29 +0100
- To: kendall@monkeyfist.com
- Cc: public-rdf-dawg@w3.org
On Tue, 28 Sep 2004 13:48:42 -0400, Kendall Clark <kendall@monkeyfist.com> wrote: > On Tue, Sep 28, 2004 at 06:06:35PM +0100, Dave Beckett wrote: > > > <result> > > <hpage href="http://work.example.org/alice/"/> > > I think the attribute here should be "uri" instead of "href"; in this > case it's probably the case that Alice's hpage could be derefenced, > but in many cases the URI in subject position isn't meant to > be. "href", while well known from HTML, doesn't match the case where > URIs are used as names or logical constants or just unique identifiers > that aren't supposed to be deref'd. > > I'd prefer, then, that this example look, in pertinent part, like > this: > > <result> > <hpage uri="http://work.example.org/alice"/> > ... I can live with that change. No need to imply retrieval or html-like semantics. And 1 char shorter :) > > <blurb><p xmlns="http://www.w3.org/1999/xhtml">My name is <b>alice</b></p></blurb> > > I prefer the escaped version, though I'm not sure why. Initially I > thought it was because I suspect some triple stores don't *really* > support XML literals per the specification (which requires Canonical > XML, iirc), but now I guess I'm just vaguely unease about the whole > thing. I think still prefering the unescaped since that reads more naturally. If the triple stores do XML bad, that's their problem. Switch to the competition or fix it :) Generally I think &-escaping gets a -1 from XML, or -lots from some. Dave
Received on Wednesday, 29 September 2004 13:27:37 UTC