How curious.....
|
| Author | Content |
|---|---|
| Ridcully Mar 23, 2010 8:37 AM EDT |
You know........I have often thought in the past year, that Microsoft's only way ahead (well, that is if it seriously wants to get rid of the virus incubator known as Windows and puts the welfare of its users ahead of profits) is to move its kernel to a Unix based system. In his last option, Bryan Richard more or less presents this exact same hypothesis. I have mentioned this hypothesis to others during the past 6 months and the instant reaction has always been : No way; Redmond has too much too lose, eat crow like that ~ never. But I still wonder. In my estimation, Windows was, is, and always will be a "patched up stand alone, single user system" that has been "sticking-plastered up" to pretend it is something it can never be: a secure, stable OS. But if Microsoft was to bite the bullet and produce WinLinux.........the IT world as we know it would be a very different place. I wonder if this has been seriously discussed at Board levels in Redmond ? It would of course, be a rejection of their present kernel representing two decades of development..........naaaah..."profits first, customer needs a distant fourth" is always the guiding principle where Redmond is concerned; and what would such a move do to DRM and proprietary secrecy ? But what an interesting line of thought. |
| r_a_trip Mar 23, 2010 10:53 AM EDT |
I have often thought in the past year, that Microsoft's only way ahead (well, that is if it seriously wants to get rid of the virus incubator known as Windows and puts the welfare of its users ahead of profits) is to move its kernel to a Unix based system. It wouldn't solve MS's problem. From what I've gleaned form various sources over the years is that the NT kernel design is one of the most fine grained ones security wise. The problem is not the NT kernel. The problem is that MS doesn't do jack squat with the available functionality. MS's problems are mostly in their user space. The way their user land is written precludes the application of strict user separation. Most of MS's own apps assume they have system level privileges and then we have hordes of third party apps that assume the same. To accomodate all these wonderful pieces of excrement, MS keeps the shoddy non-separation in place. If MS were to clean Windows up, they could use NT as its base. They just need to completely get rid of the poor user separation and clean up win32 (remove spaghetti code and rip out the commingling with actual applications like IE). The result would be a totally new OS though, since it would lose backward compatibility. In that regard it doesn't matter if they use NT or UNIX as a base. If they practise good design it will yield a good OS, but one that is incompatible with previous Windows versions. Since Windows floats on a vast base of legacy applications, they won't do this lightly or soon. As long as new lipstick on top and a bit of duct tape and plaster of paris on the inside keeps it going, Windows won't be rewritten to good engineering standards. |
| Ridcully Mar 23, 2010 5:13 PM EDT |
Good points r_a_trip.......I didn't know that about the Windows NT kernel. Assuming you are correct, two different slants on the same problem end up at the same solution: a major OS re-write in some or all aspects. But I very definitely agree with the next bit: loss of backward compatibility. For that reason alone, Microsoft would have to be pushed very hard without any other options before such a decision was made since the disturbance to Win-based commercial/medical software would be enormous. Still, the way botnets are moving now, the options are slowly becoming fewer ~ all it takes is a sequence of massive viral outbreaks with general Windows collapse; the IT equivalent of a medical pandemic. It may be sooner than we think. |
| phsolide Mar 24, 2010 5:21 PM EDT |
I'd just like to mention that we don't really know *what* the NT kernel is like. Very little real technical docs exist about it. Sure, every major release, "Microsoft Press" hires someone to do a new edition of "Inside Windows NT", "Inside Windows 2000", etc etc. And this time, it's going to be the Bach book for NT, complete with examples. And every time, it's not. I read Helen Custer's "Inside WIndows NT" cover to cover in the late 90s, with an eye towards what really really was in Windows NT back then. Even after reading, posting to usenet, etc, I still wasn't sure if what Custer described was what really existed or not. A lot of what she described sounded like the Mach 2.0 or 2.5 microkernel that NeXT based their OS on. Other people read the book and thought it was based on VAXELN, a DEC product of the time. Since then, various people have re-written that book: [url=http://www.amazon.com/s/ref=nb_sb_noss?url=search-alias=stripbooks&field-keywords=Inside windows NT&x=0&y=0]http://www.amazon.com/s/ref=nb_sb_noss?url=search-alias=stri...[/url] Some of these might have some relation to reality, but who knows? There's no "WINDOWSIX" standard to try operations out against. The current Windows is the standard for Windows. It is what it is, to quote business people. |
Posting in this forum is limited to members of the group: [ForumMods, SITEADMINS, MEMBERS.]
Becoming a member of LXer is easy and free. Join Us!