- From: <Patrick.Stickler@nokia.com>
- Date: Thu, 4 Oct 2001 11:41:04 +0300
- To: phayes@ai.uwf.edu
- Cc: www-rdf-logic@w3.org
> >My point was that whether you have some typed anonymous node > >pointing to a literal string value or a URI scheme that contains > >a typed value, the same issue exists. It's not *created* just > >because you start using URIs for typed data literals. > > That is true. The general issue of how to handle equality in an > assertional language is going to be there, and treating literals as > identifiers only reduces literal equivalence to that general problem. Right. > In effect, though, this seems like a proposal to abandon the notion > of literal altogether, as far as the language semantics is concerned. That was suggested as a possibility offered by the ability to encode typed data literals as URIs, but not (necessarily) a serious proposal that elimiation of literals as now defined in RDF should happen in actuality. > It seems more like a step backwards than a step forwards. Unlike > general names, literals have the character that one *almost* can > determine identity of reference from syntactic identity. That is the > whole point of using literals in the first place, after all: it is > why we write '5' rather than 'the successor of the successor of the > successor of the successor of the successor of zero'. But is '5' the integer '5'? I mean, exactly just how do you *know* that '5' represents an "integer". How can your application be sure? How can you test that it is in fact a valid integer, if it is presumed to be one? At present, unqualified literals are only of one type: a string. Period. Any further interpretation is *entirely* within the scope of a custom application. Providing for an explicit data type for a literal in any way -- whether by typed anonymous node or data type URI -- is essential for safe interchange of strictly typed knowledge between disparate systems (SW agents). Non-typed literals are nearly useless except for presentation to humans, so if they went away, in place of typed data literal URI resources, how would that be a step backwards? I consider that a huge leap forward. > So tossing this > aspect of literals aside, as it were, because we are having some > problems making it work *perfectly*, seems rather draconian, for all > practical purposes (I say 'practical' since it would certainly make > my job easier when writing the model theory.) Well, I'm wondering if we're able to make untyped literals work at all for the definition and interchange of business or mission critical knowledge ;-) Cheers, Patrick -- Patrick Stickler Phone: +358 3 356 0209 Senior Research Scientist Mobile: +358 50 483 9453 Nokia Research Center Fax: +358 7180 35409 Visiokatu 1, 33720 Tampere, Finland Email: patrick.stickler@nokia.com
Received on Thursday, 4 October 2001 04:41:12 UTC