A friend of mine recently returned the book Showstopper! after an extended loan. If you haven't heard of this book, allow me to quote the Amazon.com editorial summary:
interesting post…
Don’t want to go into a flame war or anything but surely the development of Linux would top this. Imagine a bunch of programmers communicating only through Usenet posts and FTP and coming up with that!
Interestingly, it looks the NT codebase is still fairly portable, even though x86 is now the only platform for windows.
The xbox OS is a cut-down version of Win2k - the xbox was based on PC x86 pc hardware. And for the xbox360 OS they started from the orignal xbox codebase… even though its a radicaly different architecture. With things like .net and DirectX 10 now appearing on the 360 there must have been some merging with windows/vista codebase.
Another interesting fact I’ve read is x86 wasn’t actually the original main target platform for windows NT. I forget which platform the new OS was originally planned for, but the it was only changed to x86 in the late stages of development - and Dave Cutler’s foresight in keeping the main codebase clean of platform-specific stuff allowed the change to x86 without a major fuss.
It’s also interesting to note that nobody is writing new operating systems any more.
Quite amazing how many people visit the www.osdev.org website every day then, I guess.
Quite amazing how many people visit the www.osdev.org website every day then, I guess.
And we’re seeing such WONDERFUL results from there arn’t we?
I also enjoyed this book. In many ways, it’s a paean to Cutler and his team. His Marine-like discipline and toughness of mind is a model for any project leader. The book reminds me of Microcosm, George Gilder’s paean to Carver Mead, another genius in technology.
I love that for the last few days many, many people have been bemoaning Jeff’s captcha and not commenting on his article. Good article, by the way. It seems like two seperate entries though; one on discipline (http://blog.codinghorror.com/discipline-makes-strong-developers/) and the other on Dave Cutler.
The original target for NT? Well, that would be the Prism RISC architecture, and NT would have been under the guise “Mica.” When Ken Olson (crazy as a loon) killed Prism/Mica at DEC (because the world only wants a little VAX, a medium VAX, and a large VAX), Cutler went to MS and took a lot of the team and the ideas with him. NT on DEC Alpha was, belatedly, Mica on Prism.
For what it’s worth, Cutler’s original target choice for CPU was MIPS. He didn’t care for i386 at all.
I think it’s possible we may not see another “from the ground up” OS developed in our lifetimes.
The internet 1.0, Web2.0, Google API etc. But I know what you mean, a real OS, not just an API or software development platform.
And we’re seeing such WONDERFUL results from there arn’t we?
And as was pointed out in the article, it took about 10 years to get to a version of Windows NT (2000) that didn’t completely suck. So give the new OS developers a chance. We wouldn’t want to appear hypocritical, would we?
It amazes me how everyone forgets NT started out as a co-development effort with IBM’s OS2. As an early third party developer I would get two large boxes of floppies from MS to develop with one was for NT and the other was for OS2. As a driver developer they were basically the same just substitute GDI for GPI etc. MS eventually dropped out of OS2 and focused on NT which eventually beat out IBM since it was used on third party PC’s. The IBM PS2 computers eventually lost out on cost. How much of IBM’s OS2 remains within NT has always been the question.
I was on that team. The platform was designed to run on x86 but the platform that Dave Cutler preffered was MIPS. We all had primarily x86 machines and one MIPS. Then came the the Alpha. It was always assumed the primary platform for V1 (NT3.1) would be x86. The cool thing about the OS at the time for Microsoft was the fact the it was portable across platforms. Dave Cutle was definately a personality. There were no build breaks from lousy chackins when Dave was at the helm. If you did break the build Dave would show up in your office and scare the living %$#(@ out of you. That WAS the process.
I just found your blog. Real nice article and I enjoyed it enough to go to Amazon and buy the book you discussed. I cant wait for that to be dropped off!
I don’t understand why somebody hasn’t started developing a third OS from the ground up. It makes no sense to me. Win NT has changed the landscape, but it is arguably the most frustrating OS in history. The problem I have with M$ technology is that it doesn’t work when you need it to, plain and simple. Whereas with Unix, it has no choice but to work. Apple software just works. The GPS funtions without flinching on an iPhone, but on my Win Mobile 6.1 Touch phone it may work or it may not. I say all of this to say that I just wish somebody would come up and develop a Unix competitor that works just as well and solid!