Having had the opportunity to use some of the existing technology (which shall remain nameless) developed around web services over the last year, I've noticed that most of it only supports a subset of what's in the WSDL 1.1 specification (published almost exactly 9 years ago) and even less supports WSDL 2.0 (published in 2007). In 2005 the large companies providing public UDDI registries discontinued them. The best alternative (that I know of) of something UDDI-like, seekda's web service search engine, seems to be inactive since 2009. It's not surprising to me that it's hiatus correlates with the SOA obituary written by Anne Thomas Manes earlier that same year.
I more or less agree with her conclusion, but not exactly the cause. Of course, the economy impacts the development of technology at all levels. What's missing from the analysis is the impact event of XMLHttpRequest in 2005. XMLHttpRequest does something very similar to web services but with much less overhead and formalism. It's true that the two technologies are not mutually exclusive, but one is focused on accomplishing a specific goal and the other requires a "redesign of [an organization's] application portfolio" as Anne puts it. That kind of work takes time, and if it does pay off, it happens much further down the line. XMLHttpRequest fits much more with the iterative model that has worked well for web applications. That's why truly disruptive and successful companies like Facebook move fast and break things, rather than spending inordinate amounts of time mulling over their application portfotlio.
The Audio Prof, Digital News Test Kitchen, CU Libraries News, CU Money Sense, sciencegeekgirl, Exemplary Support (Chris Bell), Jeeg Salbian, Paul O'Brian University Communications CU ATLAS CU Boulder Career Services Prof2Prof Danny Caballero
This weblog is not meant to represent the University of Colorado in any respect; the information and opinions contained herein are solely my own.