In the spirit of typical New Years' articles on past events, this article is going to look into the fate of some Web related standards which were expected to have a big impact but which may have evolved in unexpected ways.
Whatever happened to UDDI?
When concepts of Web services advanced from the rather simple
XML-RPC, where
RPC stands for Remote Procedure Call, to SOAP, and the W3C got busy on SOAP standards, two important
related standards were born. WSDL, for Web Services Description Language, and UDDI for
Universal Description Discovery and Integration. The idea behind UDDI was that UDDI documents
(in XML) would be held in public repositories and used
to help potential consumers of SOAP Web services locate a service matching their needs.
A UDDI document would help automate the process of getting the WSDL document providing technical
details and dynamically creating client software. The designers were thinking in terms of
a sort of "Yellow Pages" for Web services.
It turns out that, although UDDI is still part of Web Services Interoperability (WS-I)
standards, the envisioned big public registries never evolved and the major players such
as IBM closed their public example UDDI services. Instead, UDDI is alive and well inside private
corporate networks for service-oriented architecture (SOA)
and has wide toolkit support in commercial and open source toolkits.
The key technology for automating SOAP style Web service clients continues to be WSDL.
Creating clients for REST Web services frequently depends on text documentation and example programs.
Generally, the process of locating REST services does not require a service registry
approach and can be as simple as starting with a Google search.
Whatever happened to ebXML?
Electronic Business XML came from a separate line of thought from SOAP and UDDI.
It descends from EDI (Electronic Data Interchange), a standard that predates the
internet from the era when bandwidth was expensive.
EDI messages are very compact, strictly formatted, and not easily read by human beings so it seemed logical
to create ebXML and take advantage of the readability and flexibility of XML.
It was even suggested that ebXML based registries could supplant UDDI and perform the service lookup
and automation tasks.
Although EDI and ebXML are alive and well, I get the impression that
the demands of SOA are causing people to rethink their requirements. Google searches for
ebXML and SOA tend to turn up articles written several years ago.
As has been prolifically documented over the past year, the growing complexity of SOAP and
related XML based process automation tools has turned developer's attention to
the simplicity of REST based architecture.
Whatever happened to XHTML?
The intent of the XHTML 1.0 recommendation by the W3C eight years ago was to push HTML markup in
the direction of a more rigorous compliance with XML style markup. The industry "browser wars" had left
the population of browsers and Web pages in an incompatible mess. Browsers generally
accept HTML markup that is very poorly formatted so Web authors have little incentive to
clean up their act. It is not uncommon to see the opinion expressed that XHTML was a bad idea and can be ignored.
I did some digging to see how much of the Web uses XHTML. A simple Google search
over the entire Web universe for "html", "htm" and
"xhtml" found a ten to one ratio in favor of html. I ran a crawler over a few thousand pages starting
with links from some popular sites and found a surprising number of DOCTYPE values in use with XHTML
"Transitional" version to be quite common. It appears that XHTML is alive and well but
nowhere near dominant after eight years of competition with other standards. The current impulse
at the W3C seems to be to create a new HTML 5 version, but if this makes a big impact any time
soon it will be a surprise to me.
Whatever happened to EMMA
When I happened on W3C draft standards for EMMA - Extensible MultiModal Annotation Markup Language -
dating from 2004, I thought I had found another standardization effort that has come to nothing.
This standardization idea appears to have started with an "Ink Markup Language" which would
be a standard for representing touch control strokes of a stylus or finger.
It turns out that the industrious folks in the W3C "Multimodal Interaction Activity" have been
working away and as of December 15, 2008, EMMA is a W3C Proposed Recommendation.
As the computer world charges in the direction of touch screens and gesture controlled
functionality (think iPhone and Wii), EMMA could have a role as a standard way to communicate
user input on a touch screen client to Web services. I have to say I don't think this is likely
to happen because it is so cheap to put substantial processing power into client side
devices and interpret user input without an intermediate representation.
Comments
It is fascinating to watch the interplay between people and organizations who try
to write standards and those (frequently the same) who work around or ignore them.
It is Darwinian evolution at its most confusing.
Resources
Don Box on the early history of SOAP and WSDL.
ebXML as alternative to UDDI in SOA projects.
Wikipedia article on the history of XHTML.
OASIS hosted site for ebXML.
Ink Markup Language working draft 23 Feb 2004.
EMMA - a W3C Proposed Recommendation 15 December 2008.
