Lists Home |
Date Index |
On 5/26/05, Dave Pawson <email@example.com> wrote:
> I do hope its a + and - debate, rather than listening to the
> the ra-ra brigade. Recall the debate over the 'survey' recently
> on this list?
It will be a plus and minus debate if both sides show up or submit
papers! Actually, I can't imagine it being a rah-rah session. The
question is how to move forward given the very real problems. Anyone
denying that there are problems will be laughed out of the room.
> I'm more curious about this.
> Would you go to your <name your opposition team/> home base,
> and sound off about how good your <name your home team/> is?
Who's the opposition team? The whole point of XML is interop, so
we're all on the same team. It's 3rd down for XSD and there are
differences of opinion in the huddle about whether to punt or pass or
grind it out on the ground. [Apologies to non-North Americans for the
> Michael, what expectations, if any, do you have in this area?
If you're asking me, MS's expectations are that we learn what other
people think about whether or not XSD has critical mass , whether the
best way forward is to just nail down the interop issues, which issues
can be resolved by just gritting teeth and implementing the spec vs
which need to be fixed/clarified in an erratum, and whether there are
any features that are so rarely used and cause so much trouble that
they should be scrapped..
From what I've learned writing the MS position paper: I can't really
forsee anyone making a convincing argument that XSD is hopelessly
broken and should be scrapped, but I can very easily imagine that
people are finding ways to use XSD in a complementary way with
Schematron, RELAX NG, XSLT, etc. to address its limitations. There are
a bunch of problematic features that need work to be implemented the
same way by everyone, but contributing to the test suite and doing
test-driven development seems to the most appropriate way forward.
[Yup, "grind it out on the ground".] For better or worse, there
don't seem to be any subsets that remove significant problems without
losing functionality that a significant community of people out there
find important. I for one would love to get pushback or confirmation
on any of those points..