- From: Tim Berners-Lee <timbl@w3.org>
- Date: Mon, 29 Nov 2004 20:53:27 -0500
- To: Stuart Williams <skw@hp.com>
- Cc: www-tag@w3.org, Joshua Allen <joshuaa@microsoft.com>, Graham Klyne <GK@ninebynine.org>
On Oct 29, 2004, at 5:23, Stuart Williams wrote: > data:text/plain,some%20percent%20escaped%20literal%20value > You'd need to call out a specific syntax -- ie authorize the parser to construe the text as a literal. data:application/n3-literal;2.0 or data:application/python-literal;2.0 for example. You aren't identifying a piece of text, you are identifying (in this case) an floating point number. So application/rdf+n3 wouldn't work either, as that would imply an n3 document grammar. cwm has some code to do that on output, just to allow a literal object to return something for the uriref() function. It doesn't treat data: URIs specially when it sees them, though. But I guess it would make sense. Give a certain symmetry to the situation. Yes, there were constraints on what you can do with literals from reasoning in OWL, just as there are constraints that you can't for example have a datatype property be an inverse functional property (like social security number is supposed to be). But I don't believe it is the place of the RDF layer to try to enforce these things. datatype IFPs are really useful in other systems (e.g. database keys) as are literals as subjects. > Seems a bit ugly... and has probably been suggested before. Sometimes the effort of digging up the reams of discussion on this point becomes hard. It may be possible in this case.M > Stuart > --
Received on Tuesday, 30 November 2004 01:53:36 UTC