WebServices dot org

Enough About Business Value, Here's What We Really Should be Worrying About

SOA? Web services? SOAP vs. REST? Who comes up with these things? Maybe it's a time for new phrases for these service approaches.

First, let's get a grip on how the heck we pronounce SOA. Is it S-O-A (initials pronounced), or is it so-ah, as I'm hearing so many vendors say? How can we eventually figure out what something is if we can't even pronounce it?

Oracle's Mark Chapman recently brought this whole question to light at a panel discussion at a recent OASIS Symposium. Do you pronounce it S-O-A or do you call it So-ah? A report in SearchWebServices concludes that "there was no consensus among the audience and the lack of a clear way of pronouncing the acronym for Service Oriented Architecture seemed to be a metaphor for the larger problem faced by its advocates."

Well said. The article went on to observe that while every vendor is talking the SOA game, the OASIS panel -- which included architects working on standards for SOA -- "was unable to produce a final word on what SOA is." Chapman concluded that however you pronounce it, SOA is still "a work in progress."

I’ve opined in previous writings that perhaps the use of “so-ah” will evolve to a sort of secret password to signal whether someone you’re talking to is an “insider” or “outsider.” Think of the "Linux" litmus test. If someone asks you about "Ly-nix," that's a tip-off that you better keep the discussion at a very elementary level, or perhaps steer the topic to less contentious issues, such as religion or politics.

There are some other tangled semantics that are creating confusion on the market. Take SOAP , which, to the uninitiated, may appear to be SOA with a "P" added at the end. But when SOAP was launched a few years back, it technically stood for, and still is, "Simple Object Access Protocol" – whatever that means.

Ironically, SOAP came along first, before SOA was coined, so it's merely coincidence that there's a "S-O-A" (or so-ah) in SOAP. Understandably, over the years, there have been commendable efforts to brand SOAP as “ Service-Oriented Architecture Protocol .”

That almost makes too perfect sense. And indeed, unfortunately, SOA encompasses more than SOAP messages. There's all those proprietary hooks, homegrown standards, leftover CORBA chunks, and who knows what else in the stew. Oh yeah, then there's all that REST stuff, too, popping the SOA-SOAP bubble.

RepresentationalState Transfer – now, that’s a perfectly understandable description, about as simple as Simple Object Access Protocol. But, rest easy – there is a modest, informal, proposal out there to put this nonsensical acronym to rest as well. None other than Tim Bray (Sun), the father of modern XML, has proposed changing what we currently call "REST" to "Web Style."

What most people call REST is a an application style that shares basic Web underpinnings, Bray said, including, among other things, that fact that "you have a lot of things in the system, identified by URIs," and "the protocol (probably HTTP) only knows one MEP: single-request, single-response, with the request directed at a URI."

I think Web Style has a lot more style than REST, and would better define its purpose. However, quite ironically, there’s the likelihood of Web Style being acronymized to “WS,” and Bray is a publicly avowed hater of anything WS-*. We certainly don’t need an WS vs. WS catfight right now – we have enough scuffles to worry about.