W3C home > Mailing lists > Public > www-tag@w3.org > February 2002

RE: [namespaceDocument-8] 14 Theses, take 2

From: Simon St.Laurent <simonstl@simonstl.com>
Date: 18 Feb 2002 19:50:52 -0500
To: "'TAG'" <www-tag@w3.org>
Message-Id: <1014079852.948.264.camel@localhost.localdomain>
On Mon, 2002-02-18 at 17:29, David Orchard wrote:
> For point of information, I did a quick survey internally and every single
> developer that I talked in my company found option #1 to be correct, and
> option #2 to be a non-helpful definition.  The re-inforcement of this
> position was pointed out in the very name of "XML Schema", which talks about
> syntactic constraints.  A developer asked about the difference between
> definition #2 of schema and the term metadata, to which I didn't have an
> answer.  I venture that every developer in my company and probably almost
> all of our > 10 000 customers consider #1 to be the correct definition.

I'm not sure that I trust your survey to produce the right answer (and I
believe it hasn't).  The problem isn't with your estimations or your
projecting the answer to more people, but with the kind of people you
spoke to.

Given what little I know of BEA and BEA's customers, I'd wager that
nearly every one of the developers you spoke to was a programmer, likely
a programmer used to working with OOP methodologies, tools, and
languages, and/or relational databases.  I have no doubt that they are
excellent at what they do and have a thorough understanding of what they
do.  (It's entirely possible that I'm wrong about this, and I don't mean
to challenge anyone's competence at their job.)

I worry, however, that the understandings they've developed in their
work have only an imperfect relationship to work with XML and "markup
technologies" more broadly.  I'll be amplifying on this theme with
papers and presentations over the next month or so, but I suspect that
programmers, and perhaps especially good programmers are not
well-equipped to deal with the kinds of problems markup presents.

The simplest case I can bring up is that of encapsulation - a best
practice in OOP and often either impossible or catastrophic in markup. 
Modularity is popular, but opacity is not.  I'll be exploring other
cases in writing and code.

I say this as someone who is currently refactoring an interface because
I failed to appreciate how substantial the divergence between objects
(in my case) and markup was, as a programmer who went looking for OOP
solutions to markup issues.  My focus, however, isn't on the OOP - it's
on the markup.  In my experience, I've found that most programmers are
still thinking in terms of objects.

I'd suggest that your poll, however useful as a data point, is hardly a
final answer.  This paragraph in particular suggests to me that your
respondents are programmers, and not very well attuned to the different
potentials and liabilities of markup:

> I briefly thought that we could define 2 sub-types of schema 
> languages, something like: syntactic schema and interpretive/
> semantic/? schema.  I also ran this separation by our 
> developers.  The response was one of disbelief that the TAG 
> was even spending time on this when the "real" answer was so 
> obvious and one developer joked about counting angels on pin 
> heads.  All this reaffirmed my earlier position on the 
> definition of schema.

I hope TAG has a lot of thinking yet to do on these issues.

(And I think Sean McGrath has an answer regarding angels on pins.)

(And I'll shut up for now before the programmers put me in the funny
farm.  More coming soon, but in other forums.)

-- 
Simon St.Laurent
Ring around the content, a pocket full of brackets
Errors, errors, all fall down!
http://simonstl.com
Received on Monday, 18 February 2002 18:46:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:04 GMT