I actually like what they're doing with Windows 7. The lightweight core is going to be great, hopefully. I remember hearing that they were going to implement pr

I actually like what they're doing with Windows 7. The lightweight core is going to be great, hopefully. I remember hearing that they were going to implement previous functionality via virtualization. This is also the right direction. With these in mind, I would commit to these steps: In Business: I would throw my eggs in the virtualization basket. I would be promoting the hell out of Virtual Machine Manager 2008 and reorienting the entire business strategy around it. I would make sure it trounced VMWare infrastructure and brought using virtualization over a thousand-PC cluster into the imaginations of every CTO in the country. With this, I would be building a Microsoft BSD from FreeBSD. This BSD would be marketed as a JeOS type system with virtualization in mind, and I'd be working with Oracle and IBM and every major builder to certify their products for it. This would be completely managed from VMM 2008 and would be touted as Unix for business with the ease of use that Microsoft has made with its server line. (I'd also give back to the BSD community any speed tweaks and such, and try to get as much positive press as I could from that -- and deliniate between open source good, gpl bad.) I would also extend this to development and allow a simple license for MSDN developers the legal right to pull an image from a cluster and work with it on their own machines, as well as web interfaces that allow an untrusted user (such as an outside developer) to log into a single system that is encapsulated by a sandbox. (That's a lot harder than it sounds, but I'm thinking encapsulated zones and highly granular security.) I would introduce the pricing structure as per-instance. I would drop the price to the point where it is more cost-effective to run virtualized systems than dedicated systems currently, and encourage people to get the beefiest boxes they can. On the desktop, I would introduce virtualization through the idea of a "sandbox." I would make new images as simple as a few clicks, and copying images with the same ease of use. Then, hammer in the idea, "If you want to test something, like a new program, or if you're afraid you might break something, just make a new sandbox! On the web, I would drop back to creating things that create value to the user. I think Live has lost sight of that. I think they've been thinking too strategically in the domain. Can you name a single thing on msn.com or live.com where they do things better than the competition? Yahoo doesn't seem to do things much better either, but they did it earlier. Google did it better. Picking up a few web 2.0 startups might not be a bad idea, but I think a more interesting approach would be to set up a subsidiary of 500 people or so and have them all set up as groups of 5-10. (A designer, content creation, liason, and other staff that allow a 2-3 person core to do what they're best at.) Have them throw as many things as they can on the web and see what sticks. Don't worry about monitization until later, create it as a startup culture. Things that are successful (by some metric) and monitizable get their founders a bonus where 100k is the minimum. At specified points, they have the option of moving it into mainline MSN with a larger team. They can either be a lead developer on that team or start with something new. This would flood the net with new startups that may be competing with each other, but a few sensational products would be all that was needed to invigorate msn.com/live.com.