|
|
Is XML Doomed ...? (Was: Polemicism)
|
[
Lists Home |
Date Index |
Thread Index
]
In a message dated 09/05/2003 01:10:33 GMT Daylight Time, Matthew.Bennett@facs.gov.au writes:
Most of the truly insightful software I've
encountered was authored by one or two; most of the lousy software was
written by corporate teams. Isn't the plethora of inane 'standards' around
XML evidence that it's doomed to rest in the latter category?
Matt,
The plethora of standards is an interesting phenomenon.
What I suspect it does mean is that XML is doomed .... to repeat the simplification process that went from SGML -> XML (pre-PSVI).
Timescale? I don't know but the dissatisfaction that led to Relax NG being developed at OASIS isn't a unique phenomenon. We have seen on list in the last 24 hours hints of the unrest that exists in some quarters in relation to the complexification (sorry for the neologism) of XSLT and XPath. [In fact I did wonder if your post was an oblique reference to that very point.]
The focus of much of this dissatisfaction is W3C XML Schema. In part the problem is the complexity of what the XSD specification attempted, in part it is due to a failure of the specifications to adequately communicate. Formalism attempting precision in communication when little is actually communicated has a signficant component of futility. The fundamental concerns regarding W3C XML Schema might not have developed so far, if it weren't the case that W3C have progressively imposed W3C XML Schema in important follow-on specifications and are continuing to do so.
A significant number of people don't want or need the complexity that W3C XML Schema brings. If the W3C specifications were well-layered then those who wanted XSD could have used it in scenarios where they felt it was helpful, while those implementors and developers who had no perceived need for the complexity and overhead of XSD could produce software based on their needs. To impose XSD lock-in is, in my view, inevitably going to lead to a new cycle of simplification, when software developers vote with their feet and develop son of XML family, with better layered specifications.
Many look at vendor lock-in as the intellectual property equivalent of sending young boys up chimneys to sweep them. Standards body lock-in is no more enlightened. The W3C has no divine right exclusively to develop specifications. The W3C is operating in a global marketplace in which other standards bodies exist. W3C is in competition to meet the global market's needs. If W3C produces specifications which meet the needs primarily of only a few, influential parties then it will find that new organisations spring up or develop to meet a different, possibly more focussed need.
In retrospect I wonder if the development and, arguably premature, release of W3C XML Schema and its subsequenct imposition on an unprepared and, to a degree, unwilling developer community will come to be seen as a defining moment in W3C's history.
XSD Schema is an odd phenomenon. I can see why people like it. I can see why people loathe it. Personally, I am much more comfortable with it in XSLT 2.0 / XPath 2.0 than I was 6 or 9 months ago. I wouldn't say it was an approach I would have chosen, if the decision were mine. But I can see that the guys on the WG have, since that fateful decision was taken to base XSLT 2.0 / XPath 2.0 / XQuery 1.0 on XSD Schema, worked enormously hard to produce something that will get close to working for many.
My speculation is that all those efforts won't negate the progression towards a new cycle of simplification. Time will tell, I guess.
Andrew Watt
|
|
|
|
|
|