On 09/22/2012 06:53 AM, Alister wrote: > On Fri, 21 Sep 2012 18:47:57 -0400, Dave Angel wrote: > >> <SNIP> >> >> That's not true at all. You'd re thinking of Windows 3, Windows 95, 98, >> and ME, which were hacked on top of MSDOS. But Windows NT3.5, 4, 2000, >> XP, Vista and Windows 7 have an entirely different bloodline. >> >> NT 3.51 was actually very robust, but in 4.0 to gain better performance, >> they apparently did some compromising in the video driver's isolation. >> And who knows what's happened since then. > Although NT upwards has tried to introduce
Your wording seems to imply that you still think NT was built on some earlier MS product. It was written from scratch by a team recruited mostly from outside MS, including the leader, a guy who was I think experienced in VMS development. The names escape me right now. But there were a couple of books, by Helen someone, I think, which helped us outsiders understand some of the philosophies of the development. > user-space requirements the > need to maintain backwards compatibility has compromised these efforts. > it is not helped by the end user's (just look at what happened to Vista's > attempt to make users authorise any changes to the system) > > I don't see any connection between memory address space user models and user security models. -- DaveA -- http://mail.python.org/mailman/listinfo/python-list
