Even More Extensible

August 2, 2000

Alan Kotok

An Updated Survey of XML Business Vocabularies

In February 2000, XML.com published a survey that found 124 different XML business applications at various registries and reference sources. By August 2000, the number of XML business vocabularies listed with these sources appears to have doubled to over 250, putting even more pressure on standards bodies to develop a way for these applications to interact.

As in the first survey, a majority of the vocabularies (146 of the 251) represent one-industry applications, called verticals. Some 95 other entries express business functions, defined as common operations that apply to a number of industries. The fewest number of vocabularies are frameworks that offer interoperability among the business languages, but frameworks provide the most functionality and have received the most attention.

Many of the vocabularies uncovered in the new survey tend to represent much more specialized uses of XML. The new entries include more vocabularies for the exchange of scientific data, including a Physics Markup Language and languages for genetics. The increase in specialization is also expressed in single company vocabularies designed for doing businesses with those companies. For example, the DSL service provider Covad has an XML vocabulary for handling various customer service functions.

Despite the proliferation, some consolidation among vocabularies has begun. The Human Resource Management Markup Language now comes under the HR-XML consortium. Likewise the Hospitality Industry Technology Integration Standards and Open Travel Alliance announced plans in June to merge their specifications.

Since February, a new framework appeared, the Small and Medium Business XML (SMBXML) vocabulary that joins the nine others already listed in the survey. It covers purchasing, selling, banking, and accounting. NetLedger Inc., a California-based application service provider, announced SMBXML in May 2000.

As in January, the survey covers a number of registries and XML reference sources. They include leading directories of XML applications collected by XML.com, OASIS/Robin Cover, Schema.Net, and IBM's alphaWorks. The survey also covers XML vocabularies registered with OASIS's XML.org and Microsoft's BizTalk.org portal, as well as those managed by Data Interchange Standards Association (my current employer) and other standards services.

In reviewing the lists, it pays to remember Mark Twain's admonition about three kinds of lies -- lies, damn lies, and statistics. In some cases the growing numbers represent nothing more than the method of registering the vocabularies. In BizTalk.org, for example, each schema is recorded separately. Thus a vocabulary built on equivalent request and response messages will have two separate entries -- one for the request and one for the response -- even though they work as a matched pair. By comparison, the Open Travel Alliance specification has one document type definition that covers request and response messages for creation, reads, updates, and deletions of traveler customer profile records, but has only one entry in the survey.

Another caution is the apparent lack of progress made by several of the vocabularies. The Bank Internet Payment System Specification still has its April 1998 public review draft posted. The Steel Markup Language also appears to have not ventured any further than a press release in January 2000.

The new survey includes an entry for recipes, called Document Encoding and Structuring Specification for Electronic Recipe Transfer, or DESSERT for short. Be sure to save room.