- From: Harry Halpin <hhalpin@ibiblio.org>
- Date: Tue, 24 Jul 2007 12:38:18 -0400
- To: "C. M. Sperberg-McQueen" <cmsmcq@acm.org>
- Cc: Dan Connolly <connolly@w3.org>, Andrew Eisenberg <andrew.eisenberg@us.ibm.com>, public-grddl-comments@w3.org, w3c-xsl-query@w3.org
On 18 Jul 2007, at 16:08 , Harry Halpin wrote: >> An aside, although the official response will come from Dan himself. >> >> The reason GRDDL does not use third-party schemas is because it >> already has a mechanism (for both XHTML and XML) for pointing to >> third-party transformations. Since the only reason GRDDL uses a >> schema at the namespace URI is to find a transformation (as it uses >> no other part of the schema), it seems reasonable to assume that if >> a schema is somewhere else rather than at a namespace URI, then >> the user can simply directly point to the transformation given by >> the "third-party" schema. This keeps indirection to minimum. > So if I have a few thousand, or a few hundred thousand, documents > in the same vocabulary, and use a schema for that vocabulary which > is not at the namespace name but which is accessible to my processes, > the advice of the GRDDL group is not to document the transformation > in the schema document, but to edit each document and add an > explicit link to the GRDDL transformation for that vocabulary? > Given that the semantics are associated with the vocabulary, > and the annotations for the vocabulary are documented in the > schema document I use -- and given that there's no particular > technical difficulty involved here, that I can see -- why would > I as a document owner want to incur that kind of cost and > introduce that kind of redundancy? > --C. M. Sperberg-McQueen I understand that's a perfectly sensible use-case, but not for GRDDL as it stands. There is a technical difficulty: Since the schema is not at the namespace document and not marked up in the document, how would the GRDDL processor ever discover your schema annotations by "following its nose" from the source document? We did also think about a similar use-case, as the GRDDL Use-Cases draft mentions [1], but the proposed solution of linking to the transformation relies on a HTTP Header Linking Draft that, well, seems to not be making progress in the IETF. For your use-case, since the location of the schema and associated with is known by you, but not described the document or namespace document, it would make more sense to explicitly write your transformation to RDF for the type of vocabulary using something like XProc [2] I think rather than GRDDL, or simply use the transformation using a processing language like XSLT or XQuery directly. The only solution would be to add an arbitrary parameter to GRDDL. However, we have endeavored in the WG to make GRDDL "parameter-free" and instead rely on "following-your-nose" and "following other specs" to find the transformation, and if one wants to add an parameter to GRDDL to locate a transformation, one should just use the processing language like XSLT or XQuery locally and directly rather than relying on GRDDL, since there is no advantage using GRDDL would provide in this case over existing software. Does this answer satisfy you? [1]http://www.w3.org/TR/grddl-scenarios/#header_use_case [2]http://www.w3.org/TR/xproc/ -- -harry Harry Halpin, University of Edinburgh http://www.ibiblio.org/hhalpin 6B522426
Received on Tuesday, 24 July 2007 16:38:56 UTC