This week was all about Microsoft's current and future operating systems (August 4-10 2008):
- The much touted security of Windows Vista seems to be worthless, according to findings by the two security researchers Mark Dowd and Alexander Sotirov. SearchSecurity writes:
By taking advantage of the way that browsers, specifically Internet Explorer, handle active scripting and .NET objects, the pair have been able to load essentially whatever content they want into a location of their choice on a user's machine. [...] That's completely game over.
The researchers seems to have found a way to compromise .NET objects, which IE assumes are safe, effectively using the objects as stepping stones for other attacks.
- Dave Winer has a piece about why he thinks Vista may fail, comparing the operating system to two other failed OSes of the past: OS/2 and Apple III. "Vista has the smell of death," writes Winer, outlining a few other reasons why he disapproves of the OS. The ones I particularly agree with are "5. Everything is happening in the web browser now," and "6. [...] there is no demand for new operating systems." No one actually needs grand new operating system, just keeping tuning and tweaking the existing OSes should be enough. That's one thing I like about Linux (the kernel specifically), it's (almost) from the beginning been called Linux and will probably be called that forever. It just has a version number (currently 2.6.26), which is incremented as bugs are fixed and new features are added. The only reason I see for Microsoft to release a new OS is if they come up with a complete rewrite, with all junk code discarded that has piled up through the years (cf. this highlight mentioning the current problem with Windows core architecture). This could be the promise of Midori, which is pointed out as Microsoft's future non-windows operating system, see the next item below.
- Ars technica reports on recent speculations about Midori, a coming non-Windows operating system from Microsoft. Key features of Midori are that it's based on managed code, .NET presumably, which takes the load of memory management off the developer. Further, the OS should be easy to virtualize, even allowing it to run as a separate Windows process. This would facilitate backwards compatibility, as the OS could initially be run inside a legacy Windows OS. Midori is also assumed to be well prepared for cloud computing, allowing for asynchronous API calls and making it easier to program for multiple processors. This last part is important from a performance and efficiency point of view, as CPUs get multiple cores. Multithreaded programming is a difficult and error-prone task, not well suited for the everyday programmer.
- On a lighter note, Lifehacker featuring the How-To Geek, debunks a number of Windows performance tweaking myths. Like registry cleaning and memory optimizing, as is offered by several tools, and disabling of Windows services (except for possibly a few ones). Personally, I use CCleaner to remove temporary files, and I defragment my hard drives a few times a year, which seems to be enough to keep Windows XP reasonably fit.