Vote resolution and metadata (Was RE: Early ECMAScript transform)

Dimitris Dimitriadis wrote:
> As far as our vote is concerned, I can't see that we have a 
> conclusive vote; however, it seems from Mary's and Fred's 
> mails that we could go for the .domtest option (given that 
> Fred will get used to it in 5-6 years and that Mary doesn't 
> have a strong preference). Please advise if my reading is wrong.

It was a tossup and since Mary has the tests, I'm willing to go with her preference which was to maintain the .xml extension and to isolate resource files in a files subdirectory.

> 
> Metadata is one of the few things still in need of being 
> resolved. I haven't dug through the archives, but I remember 
> that the discussion was RDF vs. our own metadata. Back then, 
> myself and Mary went for the latter, whereas Curt leaned 
> toward the former. Any changes there?

The metadata elements are in the test definition namespace (for ease of test authoring) but patterned after Dublin Core expressed in RDF/XML (for interoperability with RDF).  An RDF representation of
the inline metadata could readily be generated by an XSLT transform.  

The in-line metadata only tries to capture information that is relatively constant about the test itself (its description, what it tests, who wrote it) and not opinions about the tests as expressed in
emails or in DOM WG deliberations,  judgements on its correctness, or its inclusion in test suites.  That type of information would probably initial be in the SourceForge tracker, but could be in
external metadata files and my preference would be that these files be RDF.

Received on Wednesday, 15 August 2001 12:59:27 UTC