[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: is that a fork in the road?
- From: "Bullard, Claude L (Len)" <firstname.lastname@example.org>
- To: Eric van der Vlist <email@example.com>, firstname.lastname@example.org
- Date: Fri, 02 Mar 2001 12:51:52 -0600
Sitting listening to Henry's speech as I write.
"XML trajectory and SGML trajectory as ...
affected by slowness of getting style sheet
I wonder how long that trajectory would have
1. XSLT had not been primarly a means to
downtranslate to HTML (something OmniMark
already did well and a bit better); that is,
a common existing renderer made XSLT a slam
dunk compared to DSSSL.
2. XSLT had not been designed and overviewed
by many of the same people who took the decade
long march on DSSSL. That is, it is easy to
walk on water when you know where the stumps are.
At some point, the self-congratulating, "look at
what we have done" aggrandizing of XML has to
be seen for what it is: cheerleading to a
very old and very done deal by the time "we"
got here to carry a ball the last six inches
across the goal line.
Now, the deal is this: XML got complicated because
the problem scope is hard as hell and expanding. "We" are trying
to bridge systems worldwide and make that all *look*
homogeneous. Now you have to define things that
the previous decade's players left aside because
their focus was NOT worldwide hypermedia. Now you
have to agree on the much harder bits.
If you aren't up to the challenge, take the fork
out of the road, put it in the cake, and call
it done. Then some group is going to make another
cake and have at. They have to. Real requirements
for interoperability, portability and now eXstensbility
are the new goal. This goal was known long ago;
CALS defined it. It was considered unattainable
then, so people backed off and well... HTML, XSLT, XML
and so forth.
"Abstraction abstraction abstraction" and "Location
location location" have been the problems since
SGML was a draft, they are the problems now, and
we solve them by punting them away or facing them
head on. You have to decide what you want to do.
I heartily suggest that people start discussing
the differences between the plumbing and the
refrigerator and decide what level they want to
work at. We will NEVER see a grand unified
seamless net unless we straitjacket every
innovative designer on the planet. It won't
happen. Kick that one away. It IS unattainable
because once attained, it becomes unendurable
or unsustainable. All Systems Leak. Pipelines too.
You can agree to share parts of your
solutions, you can agree to processes that
enable you to choose among these, but you cannot
arbitrarily stand in front of the innovative
programmers and tell them not to use the solutions
they design or think they have invented.
Decide what you care about. Is it the infoSet,
or is it the syntax? Is it the semantic or is
the language for expressing a semantic? What
is substance? You have to decide but the decision
you make will be yours; join a community based
on the requirements you share, or go it alone.
Your choice. Choose wisely.
Simple means if possible. No simpler than possible.
Berners-Lee got the first part right and
the second one wrong, but he made the discourse
worldwide and immediate. You have to work your way
among the abstractions and carry back to your customer
the right ones. The target moves because we are
So to return to Henry's speech and inquire into
"what an XML document is": pick a simple near
target and you get a simple short trajectory.
Pick one previously bombed and you get maximum
Pick one far away and hard to hit, and you get
to wait until you over fifty to see the first
Me, I'm already home. The cake is good.
So far, we are eating our own cooking just fine.
The next course will be a lot less tasty if we
are already full or our appetites are timid.
Take a break between courses and let the cooks
know that is what you are doing. Otherwise,
they bring you food as fast as a Sorceror's
Apprentice brings water.
Infoset pipelines look good from here. One will
want to know where to put the tap to get the water.
Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h
From: Eric van der Vlist [mailto:email@example.com]
I think that the main issue is that both the perimeter of "XML" as a
whole and the XML community is expending very fast.
It's easy for a small number of people to agree on a very focussed
It's more difficult when you increase the number of people or the
perimeter of the spec and with XML we are seeing a growth on both axes
The only way I believe this can be dealt with is through a real
modularization of the specs.
Otherwise, I think you are right, a fork is very likely to happen later