[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Why 90 percent of XML standards will fail
- From: "Bullard, Claude L (Len)" <clbullar@ingr.com>
- To: Mike.Champion@SoftwareAG-USA.com, xml-dev@lists.xml.org
- Date: Wed, 28 Feb 2001 11:35:44 -0600
Title:
Actually, we have talked about those subjects on this list. If
he is
claiming the so-called XML vocabularies are standards, he
is
probably right about the failure rate per se. It has been restated
here
before that the big committee-driven efforts like 28001
traditionally had limited success and that the keiretsu nature
of the
business supply chains would make similar efforts
dicey. Their big success will be where they are built by
the
industry itself, or are mined for good design. In both
cases,
they don't fail, precisely. They get reused. That
is
precisely why XML was applied. I'm surprised, Mike,
that
as long as you have done this, the article was needed
to
make you consider it. We have a lot of markup experience
and we
know with some certainty that the commitment
to the
vocabulary and its definitions is all that makes
it
valuable. Success varies by how you measure it.
Pirating DTD and schema elements is not only the usual
thing,
it is often, the best thing. No size fits all
comfortably.
Academic or not, he needs to make it clear that
failures
are not the same as 'reuse'. His numbers go down if
he
makes that point. His article doesn't make me
uncomfortable because it is clear he
doesn't understand
the industry he purports to consult for or
is being
disingenuous, and therefore, just another
bear.
FUD is not a way to make people
think; it is a
way to make them react in fear, therefore,
become
enamored of guidance.
As to the business model, a protocol for interface
and a protocol for performance are not the same
thing. Interfaces are discoverable;
performances
are scriptable. We may find that B2B
standard
performances emerge, and these we can safely mark
as genre: expected transitions that reward by
meeting expectations or coping with
surprise.
We accept the challenge to "make the strange
familiar and to make the familiar
strange".
It's a trade paper think piece, not
an academic article!
The gist of it is the piece seems to be:
The specification initiatives (aka
"standards") most likely to fail are those that are not aligned to the real
business needs of major companies, that over-promise, that take on an already crowded space, or that try to
dictate business processes rather than accomodate existing practice. Since 90%
of XML initiatives fall afoul of more than one of these, 90%
are likely to fail.
Is this
really unreasonable, or inconsistent with the experience that we have had
with various "standards" efforts? I don't think so ... I find this
list potentially useful in predicting which efforts will succeed and which
won't, so that I can ration my scarce attention span on the ones that might go
somewhere.
The last one
about standardizing business processes is the most controversial, I'd
guess. Does ebXML's inclusion of a "Business Process and Information
Meta Model" fall afoul of this? If so, is this really reason to
worry? Not sure ... but at least it's an interesting question that would
not have occurred to me until reading the article in question. Of course
it would be nice to have more examples to come up with a credible answer, but
that's not what think pieces are supposed to do. They're supposed to
provoke thought!