One of the more enlightened things that Microsoft does to to maintain a really serious research effort, employing really first-rate people and giving them great freedom. This report of a dinner conversationw ith CEO Craig Mundie suggests that the policy will continue.

The centrally funded model for Microsoft Research is still right: This is especially interesting to me, because I have written two books on corporate research. I can tell you that funding a corporate research operation straight from central corporate coffers — as opposed to via contracts with various business divisions for work they want done — is almost unheard of in this day and age. That’s because most companies believe that in order for their research groups to have good ties to their business needs, labs need to get all, or at least a major part of, their funding from business divisions — for work the business divisions want done. Microsoft’s approach is to let MSR work without those strings, in order to keep researchers more unfettered and open to new things. "I still think that that was and remains a good strategy," Mundie told me. Over the years, many have questioned whether Microsoft’s investment in MSR has been worth it, and everyone from Gates to SVP of research Rick Rashid on down has steadfastly maintained it has been — with contributions to just about every Microsoft product. The way Mundie put it to me: MSR is “just becoming more and more integral.”

(Note: I am talking about research, the ‘R’ part of R&D. The vast majority of Microsoft’s roughly $9 billion annual investment in R&D is for development, with the research labs getting only a fraction of this amount.)