- From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Date: Tue, 28 May 2002 14:28:56 +0100
- To: "Martin Duerst" <duerst@w3.org>, <www-i18n-comments@w3.org>
- Cc: <w3c-i18n-ig@w3.org>
> Can you give > actual examples of W3C work that are not 'web languages'? Charmod, IRI :) > Why do you think that something that > is not a 'web language' would not need to address > internationalization? Hmm, that bit might have been badly phrased. I think that some W3C stuff may even define a language which is not intended to be used on the web, but is, for example, simply for use in examples. Such a language need not, IMO, be internationalized. For example consider the Java language mappings of the DOM. Whether or not Java conforms with charmod is wholly irrelevant; and the DOM WG need to have the wiggle room to decide that. > > Do you think that a 'web language' implies that there has to be > a process model? Certainly not normatively defined. The RDF specs talk about an XML document and an equivalent RDF graph, there is some implicit processing, but there is no definition of requirements for RDF conformance. e.g. most of the WG would see Adobe's XMP which only uses a subset of the RDF/XML syntax as conformant. It quite possibly does not create an RDF graph; it does not accept every RDF/XML document; what it does do is to assign a semantics to the XML documents that it does accept that conforms with the semantics assigned by the RDF Model theory to the graph assigned to that XML document by the RDF/XML syntax. This makes it hard for the RDF recs to conform with: "Every W3C specification MUST specify that implementations MUST conform to the requirements applicable to software" The RDF recs barely discuss implementation. For example, an RDF implementation may allow searching and sorting over stings in the RDF graph; but the character of the RDF specs is not one which would mandate: "Software that sorts or searches text for users MUST do so on the basis of appropriate collation units and ordering rules for the relevant language and/or application." It is not clear from the scoping of charmod that RDF Core will not be asked to echo such text in either the current or a future version of RDF specs, even when no other discussion of sorting or searching is present. I think what the RDF Core WG is looking for is some recognition that W3C specs come in different flavours and conformance with charmod will take many flavours as well. My sense is that the I18N group is asking W3C for a hammer to hit other WGs with; it should be possible for the other WGs to make a case that their recommendation, at least in part, is not a nail. It is unclear from the words of sections 1 and 2 that intelligence will be used in wielding the hammer. In all cases WGs must meet internationalization requirements and thus charmod is at some level merely an articulation of current best practice. Jeremy
Received on Tuesday, 28 May 2002 09:29:04 UTC