At the same time that Microsoft is taking a page from the Hugh Grant playbook of dealing with bad press, in an effort to turn around the ill will it has created with its Vista operating system, attention also is focused on what comes next: Midori.
There was a flurry of articles about this Microsoft next-generation operating system research project several weeks ago. Now, SD Times says it has more technical details and has set off a new re-examination of what is and is not known about what could be done with it once it is built. I have to imagine Microsoft might not be especially thrilled about the timing of this discussion. It's all about Vista goodness right now, remember?
Be that as it may, everyone is trying to figure out what exactly is planned for Midori, and what it means for Windows users.
Driven by the options that virtualization brings, the Web-based computing portion of its Software-plus-Services strategy, and the conundrum of how to handle multicore, some see Midori as either an impossibility or one of the most exciting projects coming out of Microsoft. SD Times' David Worthington writes that so far, Microsoft researchers see three leading possibilities for transitioning to a new operating system made up of components.
The first model would allow applications to run on both Windows and Midori "... similar to Microsoft Research's Accelerator project. Applications in the Accelerator model execute some parts of applications under Windows threads while executing others on the GPU via a DirectX device driver." In this scenario, Windows remains the dominant OS which, as Worthington points out, doesn't further the goals of the Midori OS. Plus, it strains the system too much to run a full OS as a driver.
The second option would be to try to run down the middle and "fork the executive responsibilities and require the development of an executive for Midori that is based on and would run in parallel with the Windows Executive." In this scenario, Windows isn't dominant, but the legacy code still has to handle the non-legacy "bubble."
Third, and "most radical," would be "... writing the proposed Midori Executive itself from scratch, which would transform the bubble into a truly legacy-free platform." This plan would require a hypervisor for Midori, which would bring both its own complications and its own elegance, according to this article. But this option may have the fewest drawbacks, at this point.
Surprisingly, among the other possibilities covered in the Microsoft documents Worthington reviewed is the idea of sharing subsystem source code between Midori and Windows, a concept that I would think would be banished from the table while brainstorming on how to create a totally new, modular OS that would have as one of its main goals addressing stability and security problems inherent in Windows systems.