The Excruciating Slow Pace of XML Standardization

It's now the fourth week at my new job. I'm again taking a good look at the state of XML-based Business to Business standards. This stuff was part of my research way back in 2000 right before the dot-com bubble burst, now I'm back to looking at it in detail.

I've got to admit, the current state of affairs is a complete disappointment. It's been almost 3 years since I looked at it even though things are a bit different today, the pace of development has been excruciating slow.

Let's just take a look at what's been announced in the last month or so. First there's the announcement by WS-I of the Basic Profile. Finally a serious attempt at supporting better interoperability for Webservices. However, I've got to ask, "Why only now?". I mean, isn't the entire rationale behind XML as transport is to support better interoperability? What have the SOAP, WSDL, UDDI crowd been thinking this past 3 years? Does it take that long to realize asynchronous document passing supports interoperability better?

Then there's the UN/CEFACT announcement saying that they've finished the ebXML specification. This should be good news, however, nothing has changed much in the last year and the only truly solid contribution is ebMS which defines essentially reliable messaging using SOAP with attachments.

Speaking about reliable messaging, it's unbelievable that it's only now that the vendors have decided to tackle this problem. However, they can't agree, so there are now two groups competing to address this glaring omission. This leads me to wonder why XML standards are being reinvented every two years.

B2B interoperability however is an even more difficult problem than deciding on standard messaging headers. In addition to standardizing the payload syntax, you have to define a mechanisms for conveying semantics, be it in a generic business sense or it a particular domain's terminology.

There are several approaches to this problem, from defining common vocabularies (i.e. RosettaNet), defining all possible business documents (i.e. OAGIS), defining core business concepts and defining standard rules for extensions (UBL, CICA). The last approach although promising is moving forward at an excruciating slow pace.

Witness the Walmart announcement mandating suppliers to migrate their communication channels. The press release mentions EDI-INT and AS2. If you do a bit of googling, you'll quickly realize that it's all about transporting plain old EDI over HTTP! The one bright spot though is the use of UCCnet, an XML standard for synchronization of product identifiers.

What appears to be happening is that vendors are doing their research and experiments using standards bodies. It would be bearable if the work was innovative, unfortunately, it's like re-inventing the wheel by committee. Buyers are at a complete loss, if you don't have a single story for semantic inteoperability then what's the point?

I'm wondering, am I asking for too much too soon? I guess I'm tired of all this XML interoperability hype that has lead to nowhere in the last 3 years.