- From: Felix Sasaki <fsasaki@w3.org>
- Date: Tue, 12 Aug 2008 15:19:11 +0900
- To: Axel Polleres <axel.polleres@deri.org>
- CC: Jie Bao <baojie@cs.rpi.edu>, public-rdf-text@w3.org
Axel Polleres さんは書きました: > > It was only Addison and me in the call, so I just summarize by putting > some comments inline to my summary sent before, see below. > > Jie, if you are ok with that, and nobody else objexcts, I will try - > over the next week - to incorporate this statues into > > http://www.w3.org/2007/OWL/wiki/InternationalizedStringSpec > > best, > > Axel > > Jie Bao wrote: >> There is an unplanned emergence meeting I have to attend right now. >> Sorry for missing the meeting today. I will check the irc log for >> summary. >> >> Jie >> >> On Mon, Aug 11, 2008 at 12:59 PM, Axel Polleres >> <axel.polleres@deri.org> wrote: >>> Let me attempt to summarize what "is left" from last time and where we >>> seemed (from my point of view) to agree... this should give us kind >>> of an >>> agenda for today. >>> >>> best, >>> Axel >>> ------------- >>> >>> Summarizing starting points for today: >>> >>> 1) We seemed to agree last time on the following: >>> >>> *) One datatype rdf:text >>> >>> *) value space for rdf:text: >>> >>> pairs such that the first argument of the pair is a Unicode string >>> and the second one is a valid language tag following >>> http://www.rfc-editor.org/rfc/rfc4646.txt >>> >>> *) lexical space (notation in presentation syntax): >>> >>> "string@tag"^^rdf:text >>> >>> rather than > >> >>> (string, tag)^^rdf:text >>> >>> with shortcut notation: >>> >>> "string"@tag > > Addison and I agreed that the for the lang tag we would allow the same > lexical space as xml:lang. Actually, xml:lang is defineid implicitly > anyway, referencing BCP 47, BTW, see also the last issue, below. > >>> 2) We seemed to agree last time that subtag matching according to >>> RFC4647 >>> can be done by built-ins in RIF and datatype facets in OWL: >>> >>> - for RIF this means a builtin: >>> >>> pred:matches-langtag( arg1 , arg2 ) >>> >>> intended domains: >>> - arg1 rdf:text >>> - arg2 valid language range according to >>> http://www.rfc-editor.org/rfc/rfc4647.txt >>> >>> - for OWL this means: >>> >>> ??? Jie to elaborate on datatype-facets in OWL. > > > Jie, please elaborate. Actually, this doesn't seem to need to go into > http://www.w3.org/2007/OWL/wiki/InternationalizedStringSpec > but is rather part of the OWL spec. Can you send a pointer again, > where the datatype facets of OWL2 are explained in more detail? > >>> 3) Issue 1: What about "subtypes" which cannot be determined by >>> sub-pattern >>> matching? >>> >>> "art-lojban" and "jbo" >>> "zh-cmn" or "cmn" or "zh" >>> >>> see http://lists.w3.org/Archives/Public/public-owl-wg/2008Jul/0288.html > > The replacement for RFC4646 will resolve that by defining formal > canonicalization for lang tags which can be seen as a preprocessing > step for the langpattern matching of RFC4647 ... > > This relates to issue 3 below: If we reference BCP 47 instead of > RFC4646 we would get this "update" for free. > > Addison says that lang tags are always defined in a forward-compatible > manner, so we shouldn't run into trouble with that. > >>> 4) Issue 2: xs:string >>> >>> did we agree on here xs:string is located here, i.e. whether it >>> should be >>> the subtype od rdf:text with an empty lang-tag? >>> >>> Note that would mean that >>> >>> "blabla"^^xs:string is syntactic sugar for "blabla@"^^rdf:text, > > Addison and I both agreed that defining rdf:text as a supertyp of > xs:string would be desirable... the question is more: > Are we "allowed" to do this or would it have implications to existing > XML specs? As long as you do not expect XML-processors to "understand" that supertyp relation you should be fine. But it cannot hurt to ask the XML folks, when you have a concrete piece of text in a draft about the relation. Felix > >>> What about "blabla" (aka plain literal in RDF)? is this also a >>> shortcut for >>> "blabla@"^^rdf:text? > > We didn't touch that issue at the moment. In RIF, currently there is > no distinction made between "plain" literals and xs:string. > >>> 5) Issue 3: Whether to supersede RFC 3066 (the one used by RDF and >>> currently >>> by RIF) with RFC 4646 (Tags for Identifying Languages) ... >>> I kind of imagine tat I sensed last time agreement towards the newer >>> spec >>> RFC4646, would that cause trouble wrt 3066 upwards-compatibility? >>> Kind of >>> similar to the "plain literals" issue 2 above. > > BCP 47, see above. > > > > best, > Axel >
Received on Tuesday, 12 August 2008 06:20:16 UTC