Washed Clean, Washed Up

July 18, 2001

Edd Dumbill

Welcome to the first installment of my new biweekly column, <taglines/>. I'll be examining recent events in XML and providing an alternative spin on the industry. If you think you've got a story or a juicy piece of gossip, feel free to send me mail (put "taglines" in the subject).

Slippery SOAP

Earlier this month, the W3C's XML Protocol Working Group released a Working Draft of SOAP 1.2. This specification is notable for several reasons. First, it's the first version of the protocol produced inside the W3C, rather than being a submission from member companies. Second, the name "SOAP" has been retained, rather than changed to "W3C XML Protocol", as one might expect. To some people this will be welcome news. Not everyone agrees with the W3C choosing generic names like "XML Schema" and "XML Protocol" for their technology standards, as if there could only ever be one.

In fact, there seems to have been a bit of battle over the new specification. SOAP 1.2 is not completely backwards compatible with the previous SOAP specifications. In certain quarters previous versions of SOAP are seen not merely as submissions to the W3C but as working, informal standards. One might expect, then, more than a little resistance to issuing a new specification called SOAP that wasn't backward compatible. The problem has been overcome, however, by changing the namespace, so applications will be able to distinguish between the different versions of SOAP.

The third item of interest is that the name has ceased to stand for anything. In previous versions, "SOAP" was an acronym for "Simple Object Access Protocol". For the many who thought this was stretching the definition of "simple" to impossible boundaries, this will come as small relief. Doubtless there are some who are wishing that the specification had changed to match the name, rather than the other way around.

The biggest question surrounding the active development of SOAP must still be "why?" As pointed out by Tim Bray in several forums of late, SOAP has massive deployment as-is. Microsoft and others are behaving as if SOAP is a standard, so the effort involved in carrying on with its development seems a little pointless. We know from bitter experience that the complexity of an XML standard is like entropy; it can never decrease. With a large body of implementation experience, surely the time is ripe to slim SOAP down, throw away the bits nobody can understand, and be done with it.

Conference Capers

Attendees and speakers at the XML Devcon series of conferences will remember the confusing and acrimonious parting of ways of SYS-CON Media and Camelot Communications earlier this year. A rather public fight ensued between the two companies that managed XML Devcon, a fight conducted in emailed statements to anyone who had been associated with the conference. Since the parting of ways, Camelot kept the Devcon brand, and SYS-CON set up a competing "XML Edge" series of conferences. It appears now that Camelot has felt the sharp edge of the economic situation and has closed shop. This is sad news. Their conferences were generally well run and interesting.

Meanwhile SYS-CON is promoting its conference keynote speaker, Charles Goldfarb, the inventor of SGML. Their latest promotional material shows Goldfarb in a peculiar movie star pose and have rather unfortunately given him the title usually ascribed to Jon Bosak, "Father of XML". I guess "Grandfather of XML" doesn't play all that well in conference brochures. Given my own involvement with XML conferences, it would be unfair to comment further, but I do hope this isn't the start of a trend of depicting XML geeks as pinups.

Schemas with Character


This cute looking character is the Schematroll. Not a collective noun for the activities of James Clark and Murata Makoto, but the mascot for Rick Jelliffe's Schematron XML schema language. Rick started development of Schematron in 1999, and he's recently launched a company, Topologi, whose first offering is based around the technology. The Topologi Schematron Validator is a free tool for Windows that permits validation of XML documents against Schematron schemas, DTDs, or W3C XML Schemas. Schematron's distinctive trait is its ability to produce flexible and detailed analysis of errors in a document.

Rick is well-known to many in the XML community for his entertaining, refreshing, and often frank views, and there's certainly a lot of goodwill toward his new venture. A free validation tool doesn't seem like a product line in its own right. The Topologi site says they're "using it to protoype new ideas", but details aren't yet forthcoming about what the new product might be. However it's a fair bet it will be colorful and innovative.


  • Nothing like a good flamewar to make an open source project worthwhile. The Apache XML Project may produce some popular XML software, but if its code organization is as messy as the recent discussion over how to produce its web site, I'll be looking elsewhere.
  • OASIS is broadening its empire, opening a European office in the Netherlands. Though OASIS still seems like a collection of hopeful beginnings looking for a happy ending, this expansion looks like a confident move. As XML standards development moves further into the application and vertically-oriented areas that OASIS targets, its star may rise over the W3C's, which appears more and more mired in complex design issues. All OASIS needs now is an Asian outpost.
  • Matt Sergeant, one of the leading lights in the Perl-XML world, has started to look at Microsoft's .NET architecture. His observations, the July 16 entries in his use Perl journal, make for interesting reading on how implementing web services could be made easier in Perl. On a similar theme, news that Python has added XML-RPC support to the standard library has prompted discussion about whether Perl is getting left behind in the new web services world.