Blogging Roller

Nick Bradbury: when I'm coding away like a madman, I don't blog (or do much of anything else!)

That's what happens to me too. For me it's this -- do I blog about the blog software I'm developing or do I continue to code? If I blog about those cool new Roller 3.0 features I'm working on, perhaps I won't finish them in time to get them into Roller 3.0. Fortunately for my blog, a mega-thunderstorm just rolled in and it's bad enough that I've shutdown my w2100z, unplugged my UPS and moved out to the porch to catchup on email and blogs.

So regarding Nick's topic of coding like a cowboy I'll say this: it's a lot more fun to just start coding and sometimes it works best to lay-down some code up-front, but if you want to work with others you need at least a little process. On Roller, we've got a pretty light process. We require that folks write proposals for features they want to add and each must include a brief requirements statement and an outline of the design. But the purpose of a proposal isn't to carve the design in stone, it's to provoke discussions, reviews and arguments about the best way to do the work. So you can code like a cowboy, but you'd better coordinate with the other cowboys so you don't run the herd off of a cliff. You know, like the lazy-consensus cowboys of days gone by.

If you want to see what those Roller-dev proposals look like, take a look at the three we're doing for Roller 3.0: New "Atlas" front-page, New URL structure and New Atlas Macros/Models. And if you want to learn more about Roller features, architecture, project status and future plans, then attend my talk next week at ApacheCon EU next week in Dublin.

The storm has passed by now. Back to work.