SPML, love it or ignore it

A couple of years back, for the sake of some books I had published, a colleague pressed me very hard to create a MySpace page. I really hated having to accommodate the latest web fad, but I did it, and there was some value in it, I’ll admit. But then along comes Facebook, just a couple of years later, and MySpace is considered an anachronism (unless you’re selling music).

It’s tough, being the Next Big Thing, because it automatically nominates you as Last Big Thing. What’s even worse is when you’re the heir apparent, and you never quite get there. For years, the Cubs (and it hurts to even type their name) had a shortstop that everybody kept saying had “so much potential.” He never quite sprouted, but man, he had potential, right up to the day they dumped him.

I have started to wonder if this isn’t what SPML is. For years we’ve heard that anything with “Markup Language” tacked to the end of it would run the world. SPML was going to make provisioning executable via a standard protocol. But it just ain’t happening.

In my book about designing an identity and access framework (shameless plug coming)

http://www.amazon.com/Designing-Framework-Identity-Management-Osborne/dp/0071741372

I even mention how, despite some corporate cheerleading, SPML isn’t gathering the following it was expected to. A lot of vendors still don’t support it. No business support means weakening intellectual support. And by intellectual support, I mean the many do-gooders out there, like those found at universities, who develop standards with fancy acronyms all the time, trying to save the world from its own boobydom, only to see their babies coddled only at seminars and focus groups, while never seeing commercial adoption.

Some people complain that SPML is too complicated, and point out correctly that there’s no agreed-upon user schema. I’ve often thought that a souped-up SAML was a better candidate for this work anyway, since it HAS been adopted.

Oracle supports it, although it’s interesting that THE big name in database hasn’t allowed for provisioning of DB accounts via SPML. One of the role mining vendors (with whom I’ve competed), trying to dive into provisioning, was pushing it, but perhaps only until they got their connector story together. “You don’t need that connector stuff, we can do SPML for you. Oh wait, things have changed.”

If anybody can get the thing moving, it would be Oracle. Unlike Sun, Oracle has always liked to make money off their software, which means driving demand, which in turns means commercial adoption. SPML is so very close, but it could end up being the Almost Next Big Thing.