I had the privilege to work on the windows kernel team in the NT5 then XP days. I really wish they could share some of his code. it was the cleanest, well segmented, and commented code I've ever seen. it made the system much more maintainable and understandable, in areas that are inherently complex. great interfaces with a clear understanding of what was going in and coming out. and it helped all the other devs raise their game.
At Microsoft the joke, which was funny but also pretty much true, is that when MS invents a new "top of the ladder" engineering title (Distinguished Engineer, Fellow, Senior Fellow, etc.), Dave Cutler always gets it first. His influence on Windows is huge. As just one example, there basically wouldn't be a 64-bit Windows if he hadn't personally forced it into existence.
there is exactly one "senior technical fellow" at the company - davec.
while is wasn't entirely a solo effort - folks like forrest foltz, landy wang, etc were involved - it was amazing coming in on a monday morning and seeing the email that said "sync your enlistments and rebuild, you'll see a new amd64 version" and 95% of the system was now building and running for x64. it was crazy.
I was at Microsoft for over a decade and had access to the Windows NT source throughout.
I couldn't agree more about Dave's code and the code of the other early contributors to the system. Extremely clean, clear, and simple code neatly partitioned into small understandable functions.
I have been wondering the same thing. Also, Larry Osterman had hinted that Gary Kimura (one of the core 12 who joined Microsoft with Cutler) had an interesting coding style, but I don't think he ever got around to writing about it. I have always been curious to know what was so interesting about it.
Quick warning to people wanting to take a look: This code is part of a famous code leak that happened 12 years ago, in which large portions (a 200mb rar file) of the Windows NT/2000 source code found its way to the internet thanks to a company named Mainsoft being hacked. You'll notice that there's a bunch of .eml files.
There's actual kernel code in here, reading this code could most likely prevent from working on WINE or ReactOS and who knows what else could apply under your jurisdiction.
It's funny that his code is right there for us to see, but we shouldn't see it!
Publicly available to me means: "for free, but in the binary form". Open Source = can have some constraints like CDDL license for example, but I can at least read the source code.
It's not the first time I'm mentioning this around here, but the WRK 1.2 is quite easy to find online. Surely, there are some license issues with that, but with a simple google query you run into something like this : http://gate.upm.ro/os/LABs/Windows_OS_Internals_Curriculum_R... .
Here you've got most of the kernel code of the Windows Server 2003. It's clear that Dave Cutler had a great influence on it.
If you're a university student, Microsoft has an academic program where you get access to the "Windows research kernel", which is the NT kernel.
Also, the Windows 2000 code was leaked back in 2004; there's a torrent floating around in your nearest pirate bay. Disclaimer: I'm not promoting doing anything illegal.
I heard a lot of times that he wrote bulk of the NT kernel himself, is that really correct? I feel like a lot of people in his team are not given enough credit, and I also feel like he is being pushed as another tech hero.
OpenVMS is closed source. The Open there refers to DEC's (then Compaq, then HP) use of pen standards like TCP/IP, instead of proprietary standards like DECnet.
I might be mistaken, please correct me if I'm wrong, but I think he is more of an designer and architect than someone who writes a lot of code himself.
If you want to understand his designs it might help to read the book Windows Internals.
IIRC, he wrote a lot of original NT code himself. But you're proposing a false dichotomy. The core NT team consisted of people who designed and architected their code, and then wrote it. A programming technique sadly rare in today's world!
For me this sounds like saying, "he's more of a painting designer than someone who sketches or paints himself". I mean sure, artists often have plenty of people help with the work. [1] But still, I don't think you get to be a good software architect without writing a lot of code. Even if we imagine that minions did every brushstroke on Michelangelo's famous works, it'd still be worth looking at his sketches, which are amazing. [2]
