Why 90 percent of XML standards will fail

Those who are making XML standards are reliving the mistakes of past standards bodies, says John Rymer, consultant and founder of Upstream Consulting.

Those who are making XML standards are reliving the mistakes of past standards bodies. I can see what's coming and it is a whole lot less than any of us would like or need. I think 90 percent of the current activities will not produce meaningful technology. In my view, that's failure.

Pardon my skepticism, but I've lived through too many can't-miss, can't-live-without-it standards efforts. There was the gargantuan effort to create an alternative to TCP/IP by the International Standards Organization (ISO), the tortured efforts to standardize the Unix operating system, the Open Software Foundation's DCE debacle, and the gun-to-the-head tactics of the Object Management Group (OMG). Of these, only the OMG's CORBA can be called a commercial success.

Will XML truly take off? Discuss! YES

Each of these efforts suffered from one or two mistakes that doomed it to failure.

Mistake #1: Nonalignment
A key benefit of standards is vendor-neutrality. Standards organizations will tell you that vendors are responsible for implementing neutral standards in products that are fast, reliable, and scalable. Experience says you can't assume vendors that matter will get the job done unless the standard is aligned with their competitive needs. A standards organization has to align with the real strategic imperatives of major companies if it hopes to see useful implementations of its work. I see very little of this in the XML efforts underway.

Mistake #2: Over-promise
XML standards are the latest in a series of great hopes in IT. XML standards will provide users with vendor independence. XML standards will strip all of the latency out of intercompany operations at a low cost. XML standards will create a single global electronic market enabling all parties irrespective of size to engage in Internet-based electronic business. XML standards will provide for plug-and-play software.

Does any of this sound familiar to you? It should because we've heard promises just like these for standards in Unix, objects, and various network protocols. These promises are the marketing, not the reality, of XML standards. Early experience with RosettaNet and Microsoft's SOAP indicates that XML standards provide some leverage for some problems in small-scale systems. The backlash is inevitable, and can be fatal even to well-considered standards efforts.

Mistake #3: Overdo it
XML standards-making is at a fever pitch, with continual announcements by a range of standards groups of intentions, specs, proposed specs, selections of specs, and so on. There are now dozens of XML standards efforts underway--far too many to be practical for user organizations to consider, much less adopt. A winnowing process will ensue, eliminating most of the wanna-be standards announced during last year. Even big organizations, such as the United Nations' Trade Facilitation and Electronic Business (UN/CEFACT), the Organization for the Advancement of Structured Information Standards (OASIS), RosettaNet, and the Open Applications Group (OAG) face a struggle for relevance and survival.

There are only two abiding sources of XML standards. The first is the World Wide Web Consortium (W3C), which is responsible for defining the base technology of XML. The second source is credible vendor that creates publicly available XML formats and protocols as part of meaningful products. Ariba and Microsoft are in this category at the moment.

Mistake #4: Overreach
Many XML standards efforts aim to standardize business processes. RosettaNet is the leading example. I can't think of too many efforts to standardize business processes that have worked. Business processes are too specific to individual companies to standardize them. The most successful IT standards address protocols and formats. Some of the XML standards groups are following RosettaNet down this path. They may as well not. RosettaNet is operating in a unique industry context (electronics/high tech supply chain), which does not resemble other industry categories.

Pardon me for being cranky about this, but the net effect of XML standards has been to slow adoption of XML products and technology. There's too much noise, too much hype, too many promises--too much risk. Shouldn't we know better by now? Let me know what you think about XML standards in the TalkBack below.

John R. Rymer is principal consultant and founder of Upstream Consulting, a strategy-consulting group in Emeryville, Calif. Upstream has developing business strategies for several XML technology companies.