Re: typed literals and language tags - two proposals

I see the advantage thus:

The current treatment of literals is over-complex and 
confusing.  Eliminating language tags for typed literals would be a 
considerable simplification, and as such would help to improve the take-up 
and correct implementation of RDF.

#g
--

At 19:45 08/05/2003 +0100, Jeremy Carroll wrote:



>Brian McBride wrote:
>
>>At 13:39 08/05/2003 +0100, Jeremy Carroll wrote:
>>
>>
>>>These are for the Option 1 and Option 3, I will keep those names.
>>>
>>>Both options:
>>>
>>>PROPOSE reopen
>>>   pfps-08 reagle-01 reagle-02
>>
>>This looks like a larger change than I had realised.
>
>
>The reopen these issues is essentially a formal device for indicating that 
>in my opinion we should notify pfps, reagle, and the others of any changes 
>to rdf:XMLLiteral that we make.
>Option 2, which had no support on the list, makes no changes to XMLLiteral 
>and hence would not require this.
>
>
>>Can someone clearly state what advantage is gained from this.
>>Brian
>
>
>If we were to go with option 3 in particular, (but to a lesser extent 
>option 1), we have made a change that allows us to be more positive about 
>pfps-08 - that seems like an advantage.
>
>Having the language tags on the xsd typed literals is decidedly odd - so 
>option 2 seems like a no-brainer (that really is the small change of just 
>syntactically omitting semantically irrelevant language tags for the types 
>other than rdf:XMLLiteral) - we then have to decide whether we are happy 
>to have rdf:XMLLiteral as both a syntactic and semantic anomolous datatype 
>or fix it - either option 1 by not having it as a datatype or option 3  by 
>not having anomolous.
>
>I think either of those are advantageous. Maybe not sufficiently so.
>
>Jeremy
>
>

-------------------
Graham Klyne
<GK@NineByNine.org>
PGP: 0FAA 69FF C083 000B A2E9  A131 01B9 1C7A DBCA CB5E

Received on Thursday, 8 May 2003 18:19:45 UTC