Re: XDM blog entry

Rui Lopes wrote:
> Alex Milowski wrote:
>> This is very tempting.  I don't think our discussion of a design
>> principle based on infosets and their extensions precludes this.
>> I also think we have a requirement to support XDM.
>> In the end, if we end up with a specification for pipelines
>> that says "we use XDM between components", I don't think we violate our
>> design principle.
> After another reading pass through the first draft of requirements and 
> use cases, I've check that from the 28 use cases listed, I've counted, 
> at least, 8 use cases (correct me if I'm wrong) that directly need 
> pipeline support from an augmented data model.

Can you list these for me?

> These use cases mainly address HTML and text files (de)serialization 
> issues, and the XQuery/RelaxNG factors. Do we want to support these use 
> cases? Or should they be discarded/postponed?

Maybe we should consolidate serialization issues into one use case? 
That's an important feature of any XML processing technology and we
can, hopefully, point to the work done by XSLT 2.0/XQuery to facilitate
our futher spec work.

> As XDM can provide support for these use cases, I say "support XDM".

That's another vote for a requirement.  I'll make sure I add it as
a requirement that we support XDM.

--Alex Milowski

Received on Friday, 20 January 2006 14:30:12 UTC