Developing Mefeedia
Before I started this development effort on Mefeedia (we launched R1 last week, and we'll be launching a lot more stuff soon), I made a list of priorities. What makes for the ability to roll out cool stuff fast?- Short development cycles. I didn't get this at first, but mike from Blip.tv told me they use 2-week development cycles. So I started doing that, and it works. Two week development cycles rock. They really focus the mind, set deadlines. Two weeks I can handle.
- Few specs. 32 signals is famous for saying specs are bad (no functional specs!), but I disagree. The truth is, it depends on your developers, and the relationship you have with them. I have 2 developers in India, and every 2 weeks I write a 2-4 page spec with the new functionality. I also give them page mockups, already integrated in the site (ie. non-working pages). That seems to work so far. The trick is: I don't make a big spec in advance, I only spec things that we're implementing in that 2 week cycle.
- Take the time to remove functionality. The hardest part with innovating is: you have to find the functionality that rocks. Especially in a new field like videoblogging, where none of this has been done before, it is very easy to develop functionality that doesn't, well, rock. There are 100s of functions I could develop. It's not a lack of cool ideas. But some work, others don't. And you need to be lean. So what I do is rather ruthlessly remove functionality. It does take some work to remove a function, but in the future you've gained less testing, less maintenance, and more focus. So spend the time to remove stuff.
# Nov 13, 2005