- From: Felix Sasaki <fsasaki@w3.org>
- Date: Thu, 28 Apr 2005 20:18:15 +0900
- Cc: public-i18n-its@w3.org, "Lieske, Christian" <christian.lieske@sap.com>
Hi Christian, hi all, Sorry for being absent for a long time. It takes some time to get used to the "new" world ... Since this is a bit lengthly, I write a separate mail as a comment about its0504ReqPurposeSpecMap. I think we have to specify our notion of an "XML vocabulary", which we want to map. I would propose the following list as the sources of a mapping (to keep it simple, there should be just one kind of target, e.g. one of 1-5): 1 (Qualified) element names, e.g. <img> 2 (Qualified) attribute names, e.g. @img 3 attributes with fixed values, e.g. @type="img" 4 elements with a specified required attributes (i.e. combining 1 and 2), e.g. <figure img=".."> 5 elements with a specified attribute with fixed values (combining 1 and 3) <figure type="img"> As you will see from the list, all variants express the same meaning, .e.g. markup for an image object. But this meaning might be conveyed by such variants. Of course you could imagine more complex variants, e.g. elements with attributes of a certain type (like ENTITY) or elements in a certain context which is selected by XPath, but this would leave some schema languages out (i.e. RELAX NG and Schematron). The list is not by myself, it is a subset of the architectures which are part of the HyTime Standard [1]. It has been developed for mapping SGML (or XML) vocabularies on the DTD Level. It is important not to leave the DTD level, i.e. arbitrary XPath should be forbidden, because such mappings would be applicapable only for XML instance documents, not for a vocabulary. Another question is whether we should use XML or s.t. else (e.g. RDF ...) to specify the mappings. I think mappings should not be part of the ITS vocabulary itself, because they have a different purpose. I have submitted a draft paper to extreme markup languages which describes the mapping between various schema languages with RDF, but the methodology can also be used to describe 1-5. I hope this is a useful input. Best regards, Felix. [1] http://xml.coverpages.org/archForms.html or http://xml.coverpages.org/kimberArchIntro980113.html vor a tutorial on
Received on Thursday, 28 April 2005 11:18:25 UTC