It's been a few weeks since the dustup over whether SOA is dead or not (consensus: not), a tempest in a teapot kicked off by Anne Thomas Manes claiming that it was the recession that did it, in the study, with a pipewrench. This, of course, garnered immediate responses of "nuh-uh" and "it's just a hoax!" and hilarity ensued.

It turns out that SOA hasn't been quite the failure that Manes portrayed it as, at least in the minds of those who have been deploying it, and that even when it has failed it has been relatively bloodless, at least in traditional multi-million dollar IT failure terms.

Manes later clarified her point by saying that yeah, she wasn't kidding, the concept itself is really dead and should be left that way, but also said, "My point is that IT people should stop using the term "SOA" when talking to business people. They should not try to sell architectural concepts to business people. They should focus on selling solutions that deliver value to the business."

I can get behind that, at least.

But I think that this was a debate that happened largely inside the cloistered community of pundits and SOA insiders, and perhaps missed out on some of the larger concepts and implications of SOA in business. As everyone from all sides readily admitted, service-orientation isn't going away anytime soon, and the acceptance or denial of an acronym doesn't make or break a concept. The sweeping SOA approach that Manes envisions as integral to the concept may only be something that ever existed inside that community; and her point that what is important to the business is what is most important only tangentially acknowledges that it's the business that we should all be talking about, not our own perception of an architectural concept which will by definition apply differently to almost every business.

As Dana Gardner points out, the real issue isn't so much whether SOA has succeeded or failed, but the fact that legacy IT is falling by the wayside and you're going to need to find something to replace it. I disagree with his assertion that legacy IT is entirely dead, or will be, incidentally; there are niches where it may be the most effective means of delivering services still, and there may always be. IT is a large field with many permutations, and while I'm as big a fan of the sweeping generalization as the next guy, it's worth keeping that fact in mind when we have these debates. Still, there is little doubt that the larger part of the field is indeed in the midst of a major transformation, and it's something that CIOs need to seriously consider even if they don't hop on the bandwagon. Gardner is absolutely correct that this is the time to do it; the broad strokes necessary to transform your organization will be easier and more palatable in the midst of this crisis than they were last year, or will be two years from now.

• Is enterprise CRM Dead?