xml-dev - The failure to communicate XML - and its costs to e-business

[AndrewWatt2000@aol.com]

I have titled this essay "The failure to communicate XML" but could equally 
accurately have called it "The great XML confidence trick".

I am not saying that to provoke but to emphasise the seriousness of the 
misinformation and misunderstanding about the XML family of technologies 
which I believe exists and needs to be corrected. Confusion and lack of 
clarity exist at many levels, of which I mention only a few. There is a 
substantive problem there which needs to be fixed.

The multiple existing failures to communicate and comprehend XML technologies 
lead to unnecessary inefficiencies in the e-business setting. If e-business 
is to be measured in the near future in trillions of dollars can it be 
claimed that misunderstanding and misapplication of XML technologies is 
without major and increasing cost to e-businesses?

XML (Extensible Markup Language), so it has been claimed, is a "subset" of 
SGML (the Standard Generalized Markup Language). SGML, for better or worse, 
is notorious for its complexity and lack of accessibility. XML did, at least 
initially, remove some of the complexities of SGML but to promote XML as a 
"simple format" is profoundly misleading.

If XML ever was "simple", can it seriously be suggested that that remains 
true after the addition of SMIL, XSLT, XPath, RDF and XHTML and the soon 
emergence of SVG, XPointer, SMIL 2.0, SMIL Animation, CC/PP, Canonical XML, 
XML Digital Signatures etc?

XML is not, for the vast majority of humankind, a "simple" format although it 
may be for some. Yet, even if the simplicity of format of XML were conceded 
that in no way means that there is simplicity or efficiency in usage.

Computing is founded on the simplest of numerical formats ... zeroes and ones 
... but who would claim that using that ultimately simple numerical format 
has led to computers which are simple, efficient or reliable?

XML, so it has been claimed, is the "next generation HTML". Yet more 
misinformation or misunderstanding. It would be fairly accurate to claim that 
XML is the "next generation SGML" but that is a far less cosy image given the 
unfamiliarity and notorious complexity of SGML, as perceived by many Web 
developers. Much more convenient to put forward what is, at best, a half 
truth but one which has media impact and an appearance of familiarity.

XML, like SGML, has no pre-defined vocabulary. Each is a "language without 
words" ... or a "meta-language". It would have been a significant improvement 
if SGML and XML had been explicitly identified as meta-languages. Potential 
students would have immediately recognised that such meta-languages raise 
issues not found in everyday communication. Perhaps the Extensible Markup 
MetaLanguage, XMML, although a more accurate term was thought not to have the 
same instant appeal.

The problems of terminology and comprehension go deeper.

For example, the notion of a "root" (with or without some qualifying term) 
appears in many XML-related Recommendations or drafts. Yet there is 
significant inconsistency in the usage of terms between XML technologies 
which inevitably leads to wasted development time and consequent (avoidable) 
costs to e-businesses.

But other issues also arise. Not least is the sheer volume of material which 
needs to be mastered.

XML on its own does, essentially, nothing. Let's add the approximately 90 
pages of the XSLT Recommendation to express our XML as HTML or XHTML. Then we 
can add the 500 pages of the XSL-FO draft if we want the potential advantages 
of Web and paper output from the same source data. And if we want to 
illustrate our pages with Scalable Vector Graphics (SVG) images we can add 
another 492 pages of reading. But, let's not forget that SVG has dependencies 
on Cascading StyleSheets, CSS, has a variant Document Object Model, and is 
dependent on the SMIL Animation drafts for animation, which in turn has 
dependencies on the SMIL 1.0 Recommendation and the SMIL 2.0 Working Draft.

Now, just who was it that claimed that XML was "simple"?

The failure to communicate XML technologies as they actually are is impeding 
the W3C's declared objective of leading the Web to its full potential.

There is a need for misunderstandings of XML technologies to be dispelled. 
XML is not, was not nor ever shall be "simple". XML is not the "next 
generation HTML".

I firmly believe that XML technologies do have a potentially major role in 
e-business. It is time to move on from a stage of partisan enthusiasm to a 
mature, objective evaluation of the complexities and potential of XML 
technologies. XML technologies have, I believe, the strengths to withstand 
such scrutiny.

It is sometimes suggested that W3C documents target only implementors. When, 
as XML technologies are increasingly adopted by e-businesses, the health or 
survival of a business depends on efficiency online ... a survival which may 
depend on savvy usage of XML technologies ... a wider, wiser more 
business-orientated approach is needed.

Communication, at all levels, needs to be improved. For example:

1. Improve the expression of concepts in W3C documents. Take steps to improve 
accessibility for those not in the current elite.
2. Improve the integration between W3C activities
3. Move forward from hype about XML technologies to realistic appraisal of 
strengths and weaknesses
4. Initiate remedial action to improve communication of past W3C documents

I am aware that W3C has begun to take steps in at least some of these 
directions. The issues are too important for tentative, partially funded 
initiatives, to suffice.

XML technologies are too important to be allowed to be confined to an elite 
ghetto, as happened to SGML. Let's all work together to understand, 
communicate and apply the enormous potential of XML technologies and assist 
the W3C to lead the Web to its full potential.

Andrew Watt