Little noted but of major interest: At last months Microsoft MVP Global Summit, Microsoft Chief Software Architect Ray Ozzie spoke about how he approaches his role as leader software technology steward at Microsoft. The session provided an inside view of how this famed technologist operates.Ozzie, who worked on the influential ‘70s Plato project, led Iris Research, invented Lotus Notes and oversaw development of Groove, said that over the years at small startups and at large companies, he has run across different challenges in terms of how to get feedback into the software development lifecycle.
He said, “With small products, with low volume products and early stage products it actually tends to be pretty easy to get feedback. You bring something new to market, some early stage people self-select and they jump on this thing, and you know who to listen to…But with high-volume products there are a tremendous amount of challenges.”
“The biggest problem is that you don’t really know who to listen to. You get so much feedback, you can’t hear everyone, or find out what is really as important as easily. So the development team isolates themselves, instead of listening to feedback all day. But once isolated, who should they be listening to? The executives who fund the project? Family and friends? News stories? The real question is how do you find a way to select participants that “reflect just the right set of feedback that helps you use your constrained resources very wisely in order to have the greatest impact moving forward,” he said.
Ozzie observed that with Wikipedia 1% of the contributors are responsible for 50% of the contribution, and with Digg, the top 100 contributors are responsible for 56% of the site. Likewise, of about 100 million participants in online tech communities surrounding Microsoft products, 4,000 MVPs submit 70% of all the postings that are online. They also report about 20 to 30% of the bugs that actually get fixed.
When he goes to meetings, he asks the development teams to tell a story about some relevant audience on the outside that might care about what they are working on. “In the mid ‘80s you could pretty much build anything … interesting or innovative and new … But the environment today is far, far, far different. It is not supply constrained by any stretch of the imagination. “We have an abundance of amazing choices of technology out there, products and services for enterprise, for IT pros, for developers, for individual end users, and we have to really deeply earn the attention of each one of those audiences, that you and they have a world of choices.”
In this realm statistical measures break down, and what becomes important is how this is going to affect people. Ozzie characterized Internet centric products into three principles, which, depending on which team he is talking to, he will concentrate on one more than the others:
1) The Web is a hub.
2) Server service symmetry – the choice enterprises have to host locally versus the compute cloud.
3) Small pieces loosely joined – where both the front-end and back-end software is being fragmented in some meaningful way.
Read the full transcript of his talk to find out more details:
Ray Ozzie at the Microsoft 2008 MVP Summit – Microsoft.com