[
Lists Home |
Date Index |
Thread Index
]
On Sat, May 10, 2003 at 01:03:01PM -0400, Mike Champion wrote:
> - Lots of people simply don't have time to track these huge specs, so they
> wait until last call because it's confusing to keep looking at a moving
> target. OK, no more excuses, it's time to plow through this stuff! ... but
> it's pointless to complain about people who haven't wanted the headaches
> previously. Without stable specs and interoperable implementations, this
> stuff simply doesn't get the attention of people who are out there actually
> building applications with XML.
That's a good chunk of the problem. The bigger issue is that we've moved
past an era when a few hackers could come together, identify a problem to
be solved, and come up with a small spec to describe a resonable,
interoperable solution. Not the *best* solution, mind you, but a small,
reasonable one.
Back in the Good Old Days, you could choose the base features that your
problem needed for interoperability (or the code you were willing to write
in the name of interoperability). DTD-Validity? Optional. Namespaces?
Good idea, but optional. RDF? Take it, leave it, incorporate it later
when people actually understand it. But at the end of the day, your data
is either well-formed or it's not XML, and that's a surprisingly solid
foundation for doing a lot more than we ever could with icky HTML, binary
files or a plethora of plaintext formats.
What do we have today? A standards body that's dismissive of its hacker
roots, acting more like an arbiter between vendors who are trying to build
solutions to problems we're not trying to solve, and standardize products
that we do not want to buy.
What I fail to understand is why anyone is confused about this. Of course
the big vendors are offended by the old, curmudgeonly denizens of xml-dev
who won't just roll over and die already. After all, we saw how a few
bright people could get together *here* and solve a nasty interoperability
problem in a simple and easy-to-implement manner. We're still looking
for the next big lever that will simplify our programs and increase
productivity -- without forcing everyone to do their homework and read
five reams' worth of specs and abstractions to get a modicum of benefit.
Z.
|
