4 ms·

What do you think of Microsoft Research?

by sid0 16y ago

kenjackson 16y ago
Best CS research group in the world. Have added very small incremental value to MS as a company. The two issues, as I see them are: 1) MS is really slow productizing MS Research innovations. I think part of the problem is that MS imposes a higher bar on "product quality" for things that come out of research than things that come out of the product groups. Had the iPhone multi-touch on a phone UI come through MSR it would have taken MS 5 years to productize it. Apple did it in 18 months. 2) A lot of MS Research appears to be done in a vacuum (with respect to Microsoft, not with respect to the research community). I think Live Labs was a reaction to that. But frankly, I think having this separate research units don't work. Rick Rashid has no ability to really get Steve Sinofsky to do anything. The research group needs to be part of the product teams. The research teams should be tiger teams attacking problems with the express goal of integrating back into the product along a certain time horizon. Everyone needs to be invested in their success -- both the core product team and this tiger team (and note, the tiger team shouldn't be a distinct team... at various points, different members of the product team are a part of it, depending on the task at hand). The MS research model I think is becoming dated. The amazing thing Google did is to get top-flight researchers and faculty members working on writing code that ships, even if it is Javascript and PHP.
- sid0 16y ago
  (Disclosure: I interned at MSR over the summer. These are my personal opinions.) Note that MSR is a basic research organization, not a so-called "R&D" unit. MSR is largely focused on academic research -- there seems to be an implicit mandate that 99% of MS research output will "fail" and not be productized. For instance, I can't imagine a Visual Haskell Professional coming out in the next ten years -- or ever, really. However, there's definitely a huge value in academic research qua academic research, particularly in CS theory. MSR has algorithms, crypto/applied math and software engineering research groups that do top-notch academic work, and it would be a great loss to the CS world were any of them shut down just because they don't productize enough. An ancillary benefit of MSR is that it improves Microsoft's reputation in the traditionally pro-Linux academic world.

What do you think of Microsoft Research?

by sid0 16y ago

What do you think of Microsoft Research?

Best CS research group in the world. Have added very small incremental value to MS as a company. The two issues, as I see them are: 1) MS is really slow productizing MS Research innovations. I think part of the problem is that MS imposes a higher bar on "product quality" for things that come out of research than things that come out of the product groups. Had the iPhone multi-touch on a phone UI come through MSR it would have taken MS 5 years to productize it. Apple did it in 18 months. 2) A lot of MS Research appears to be done in a vacuum (with respect to Microsoft, not with respect to the research community). I think Live Labs was a reaction to that. But frankly, I think having this separate research units don't work. Rick Rashid has no ability to really get Steve Sinofsky to do anything. The research group needs to be part of the product teams. The research teams should be tiger teams attacking problems with the express goal of integrating back into the product along a certain time horizon. Everyone needs to be invested in their success -- both the core product team and this tiger team (and note, the tiger team shouldn't be a distinct team... at various points, different members of the product team are a part of it, depending on the task at hand). The MS research model I think is becoming dated. The amazing thing Google did is to get top-flight researchers and faculty members working on writing code that ships, even if it is Javascript and PHP.

sid0 16y ago
(Disclosure: I interned at MSR over the summer. These are my personal opinions.) Note that MSR is a basic research organization, not a so-called "R&D" unit. MSR is largely focused on academic research -- there seems to be an implicit mandate that 99% of MS research output will "fail" and not be productized. For instance, I can't imagine a Visual Haskell Professional coming out in the next ten years -- or ever, really. However, there's definitely a huge value in academic research qua academic research, particularly in CS theory. MSR has algorithms, crypto/applied math and software engineering research groups that do top-notch academic work, and it would be a great loss to the CS world were any of them shut down just because they don't productize enough. An ancillary benefit of MSR is that it improves Microsoft's reputation in the traditionally pro-Linux academic world.