XML-ER thoughts for input to TAG f2f discussion tomorrow

Hello folks,

I've mostly failed to get back up to speed in time for this week's TAG
f2f. I claim extenuating circumstances, but that's not really very
helpful.

I've read through my backlog of mail and pulled out a few high-level
issues that I don't think are resolved (I could have missed some, I
don't assert that this is a definitive list). There are also detailed
comments on Anne's draft[1] to be worked out. And there's the question
of use cases and sample documents[2].

1. Processor or declarative text transformation

There are clearly a range of opinions on what kind of thing an XML-ER
processor is and how it's described. Some folks think it's a processor
that you use instead of an XML processor (so it parses and produces
a tree). Some folks think it's a declarative thing that pre-processes.
There are probably as many opinions as there have been commenters in
the discussion.

2. What about non-XML characters

I don't recall what Anne's draft says about them, but they came up a
couple of times. What do you do with tag name that starts with a
digit? Or contains a non-Name character of any sort.

3. Schema aware algorithm?

If you know the schema (lowecase-s) for the document you're trying to
parse, you can sometimes do better. I don't think anyone argued for
trying to do this for XML-ER, but to the extent practical, it might be
nice if one could build on XML-ER to add that functionality.

4. Prioritizing use cases

I don't think we've identified all the use cases and I'm sure we haven't
reached agreement on which ones are "must support" and which ones are
"nice to have" etc.

5. Media type

There's plenty of room for debate about how an XML-ER processor fits
into the landscape of media types.

Hopefully these items, as well as any review of Anne's draft that might
be ready, will prove useful fodder for discussion tomorrow.

                                        Be seeing you,
                                          norm

[1] http://dvcs.w3.org/hg/xml-er/
[2] http://www.w3.org/community/xml-er/wiki/Samples

-- 
Norman Walsh
Lead Engineer
MarkLogic Corporation
Phone: +1 413 624 6676
www.marklogic.com