[
Lists Home |
Date Index |
Thread Index
]
At 09:18 PM 6/18/2003 -0400, Rich Salz wrote:
>But that's the arrogance. Sure, for every use-case someone posts,
>someone can post a counter-example. That doesn't make the uses less
>useful.
To me the arrogance arises from standardizing such things to the maximum
extent possible. Taking a use case which seems plainly marginal and gluing
it into a specification which affects a large number of people who don't
actually need it imposes costs that the arrogant prefer to ignore.
>The more the infrastructure can do, the better.
The more _my_ infrastructure can do, the better. The less _your_
infrastructure imposes on me, the better. Finding a balance between those
has always been a better idea than building the largest possible contraption.
> Why should I need a special tool to see if my manifest is valid when the
> laws of physics and the FAA make it feasible to use XSD?
Because the tool can operate in _your_ environment without inflicting side
effects across the landscape.
> > Perhaps that's hubris to you, but acknowledging that the world is beyond
> > the precise control of schemas seems like a useful form of humility.
>
>Nobody's refuted the calendar set of examples.
Sure, it's useful. Does it belong in a schema describing markup? No, I
don't think so. I don't think it adds any more value than the current
calendar disaster the WXS datatypes have already created with
gHorribleKludge. Calendaring is a tough set of problems. I don't think
maxOccurs and minOccurs in WXS does much to solve them.
>And yes, it's hubris to say "nobody ever works in a controlled
>environment" and "schema shouldn't be allowed such fine-grain control."
It's hubris to say that "because I want this goddamn feature, you're all
going to have it too." Finding the right balance of features keeps us off
each other's property while still letting us talk across the fences. Not
that WXS is anywhere near balanced, of course.
|
