- From: Dave Reynolds <der@hplb.hpl.hp.com>
- Date: Tue, 28 Aug 2007 14:32:06 +0100
- To: Jos de Bruijn <debruijn@inf.unibz.it>
- CC: Michael Kifer <kifer@cs.sunysb.edu>, RIF WG <public-rif-wg@w3.org>
Jos de Bruijn wrote:
>>>>>>>>> Axel: Strings with @ signs in RDF - will this be OK in the proposed
>>>>>>>>> format for such literals?
>>>>>>> The original proposal was that the text type is a pair of lexical form
>>>>>>> and language code. The XML syntax would use attributes for the language
>>>>>>> code as normal. For the presentation syntax (but see below) I'd suggest
>>>>>>> following N3/Turtle:
>>>>>>>
>>>>>>> "string"@lang
>>>>>>>
>>>>>>> which would correspond to the literal value ("string", lang)^^rif:text.
>>>>>> For uniformity of the syntax, it is better to use "string@lang"^^...
>>>>>> where @ is special (i.e., needs to be escaped, if one wants to include it
>>>>>> in the string.
>>>>>> By the way, why can't we use xsd:string data type for these?
>>>>> Because, in this case, it would be impossible to distinguish between
>>>>> strings with a '@', and strings with a language tag.
>>>> No. Read again about having @ a special symbol that needs to be escaped.
>>> So, you propose to always escape it? It seems that you would then change
>>> the xsd:string datatype, which is something we should avoid.
>> Perhaps using xsd:string is not the right way, since this type does not
>> have the language specifier (right?).
>
> Correct.
>
>> So, we should use a different data
>> type. Perhaps rif:text or rif:istring.
>
> So we are in agreement.
> I sent an email to the XSD folks earlier, to see whether/how such a
> datatype could be aligned with the XML Schema datatypes.
>
>> But the syntax should be uniform:
>> "...@lang"^^rif:text. Inside the "..." the @ sign should be escaped if it
>> is not the language sign.
>
> I have no problem with that; sounds reasonable.
In terms of the abstract syntax the rif:text datatype should have two
components - the lexical form and the language tag.
This should be translated in the XML syntax to character content plus an
xml:lang tag.
If there is some reason the BNF is clearer with this strange escaped-@
then I guess I won't oppose it but the BNF is not the important syntax.
However, I would have thought it would be better to have the bit between
the "..." correspond to the character content in the XML.
Dave
--
Hewlett-Packard Limited
Registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England
Received on Tuesday, 28 August 2007 13:32:32 UTC