UDDI's Slow Adoption and Ad Hoc Approaches to a Services Registry

I've received a lot of positive feedback on this article. Many asked me why I didn't come out stronger against UDDI. I've noted since then that there seems to be a lot of activity around using RSS for a registry. I think it can be used as is today for small numbers of services but some enhancements are required in order for it to scale. More specifically: standardization, tooling, and how to allow it to scale (federation).

But it does raise the questions: Why is UDDI not just killing off this sort of ad hoc approach? Where is UDDI in terms of adoption? Are people buying it? One would think that it would dominate the market around services repository but it hasn't. It may dominate the commercial repositary space but there are a lot of ad hoc approaches going on in SOA implementations.

Some industry leaders I've spoken too simply don't like UDDI. "It's too hierarchical", "It's too much like a CORBA Naming Service", "I need something that's more flexible in terms of lookup". For many, as I mentioned in that earlier posting, it's just too big for the few services that they have started with. Others seem to be aligning with my view of using technologies like RSS to provide a cheap and easy to use registry. (And if the hits on my original post are anything to go by we should see much more.)

In my original post I tried to stay neutral on UDDI saying:

"I am not trying to replace UDDI. UDDI is the right and standard approach for discovering Web services. In fact I think that my idea can compliment UDDI ..."

Though I still maintain that the RSS approach can compliment UDDI, I'm leaning more towards the "who needs it" (UDDI) approach. This is based on the response I've received. And I'll need stronger convincing that UDDI is necessary. Perhaps UDDI needs an overhaul?