The Future of Sexxi

What a difference a week makes. I started work at Tangerine expecting to spend the rest of my summer working on what we conceptualized as a “next generation content management system.” It was called Sexxi, for those of you who didn’t catch the site while it was up, and it’s aim was to be Social, EXtensible, and leXIcal. And it would have been, too. It may still be.
The general idea was that an author using Sexxi would tag keywords in her posts. Sexxi would index these keywords and automatically link them up with other content on the local blog, then head on out through the author’s blogroll and/or FOAF files to see if the people she knew and read were writing about the same stuff. Pretty social, huh? Lexical, even? And when conceptualizing it we recognized that it could probably just be a Moveable Type plugin – MT has extensibility out the wazoo. But was this really Sexxi enough?

We wanted to shake things up. We initially ignored the inclination to work our idea into a tool many people (ourselves included) happily use, deciding that we could make sweeping under-the-hood improvements to the way blog tools worked. It was a grand design: 100% XML on the backend, coded in totally object-oriented Ruby, open source, built to just be uploaded and run, no messy configuration. It was even to do things like Google and Technorati integration out of the box. Who wouldn’t switch?

Then problems arose. Ruby, it seems, needs to mature a little before you can think about shipping a complex Ruby application without requiring the end-user to have root privileges on their host for library installation. So we looked at other languages and decided on PHP. This quickly lead us to Pivot: it’s got the backend improvements and just about everything we wanted going for it. But, to be perfectly brutal, who the hell’s using Pivot? It certainly got me thinking.

Obviously people will adopt better software when it comes along. But would Sexxi provide anything markedly better for the user? Other than the keywording and remote indexing, not much that would be readily visible. And was the keywording and such that unique, or even in demand? Clay Shirky told us in person it sounded kinda neat, and we’ve gotten some similarly reserved interest from others. Nothing that we could see motivating users to leave their favorite blog tool – which I can sympathize with, as Moveable Type treats me pretty well, by and large.

So Sexxi won’t be a full-fledged CMS. It’s gonna be an MT plugin, and even then it may rely on some related and prior art: Adam Kalsey’s work on Related Entries (manual keywording? pish-posh!), Ben n’ Ben’s More Like This From Others, and so forth. The truth is, I’d rather implement this as a plugin for software people actually use then reinvent the wheel just to add one feature, even if it would’ve been one badass wheel. I still think the blog world could use this breed of truly social social software, but we need another CMS like a hole in our collective heads.

I’ll have it working sometime not so far away.