Remember all the WS-* specifications that had garnered significant traction in the last decade among major software vendors? It seems that we might have seen the last of them. The Web Services Interoperability Organization (WS-I) recently announced that they had completed their work. All WS-I’s assets, operation and mission will now transition over to OASIS that will continue to drive open standards, as applicable. WS-I, which was started in 2002 had a clear goal of laying the groundwork for Web Services Interoperability and did so actively by developing profiles, sample apps and tools towards that.
As developers, how much did WS-I matter to us in recent times is an open question. While there is no doubt that the material produced by WS-I was valuable, the sheer amount of WS-* standards and the complexity that came about as a result of that, turned out to be its biggest stumbling block. Interoperability is no doubt an important aspect of any application we write today. But did the world need those complex standards for even the most basic of Integration requirements? Most likely not. As a result of that, developers have adopted leaner and simpler integration mechanisms like REST and with JSON as a data format.
In our directory, we have noticed that REST is widely being preferred to Web Services as the integration mechanism, even to the extent that many API providers are only going the REST route.
This is not to take anything away from the work that WS-I has put in. Simon Phipps summarized the WS-I in a much pragmatic way. In his blog post, The End of the Road for Web Services, Phipps mentions that the large companies came together in the name of “preventing lock-in” but ended up creating massively complex layered specifications for conducting transactions across the Internet. Another good point in his article mentions that Web Services became more of a tool for enterprise application integration and not suited to the Web.
What do you think about WS-I? Did it really matter in the last few years? Or was folding up WS-I the best decision, since it was leaner and simpler methods were the clear winners in the battle for the integration glue?