[hnthrowaway0328]

I have ways wanted to understand the Windows NT kernel -- maybe the earlier versions as they are simpler.

I heard the first version is close to VMS. Is it true? Is there any material on VMS? OpenVMS is open sourced but the version is too high.

[kelsey98765431]

For a good look at the development of the NT kernel, read the book show stopper from 1996. It is correct that Dave Cutler did lead a number of projects inside DEC including projects related to VMS, the only inheritance the Microsoft New Technology kernel received from VMS was spiritual at the design and architecture levels. Dave was already in the process of a major overhaul or successor to VMS when his team was shown the door at DEC and many of these visions made their way into windows at the cost of the robustness of the minicomputing established design patterns. Cutler saw the MSDOS team as rookie programmers writing themselves into spaghetti solutions for the wrong problems but papa gates had it locked up tight with the OEM contracting agreements so Bill paid him in blank checks to deal with the kids in the room and ship a real kernel to compete with what steve was cooking at NeXT.

So no, NT is nothing at all like VMS from a featureset or likeness perspective, nothing at all from an implementation perspective (VMS had beautiful clean C standard library and compiler suite), but very much was the race to the bottom by the mini crowd to play games in the micro world.

check dave plumbers interview with cutler on youtube for more color.

[hnthrowaway0328]

I am reading the book, now that you mentioned. I only reach the part when David was hired by Microsoft though. Actually I'm super interested in his earlier works in Dupont and DEC as they paved the way.

I took him as an inspiration. His “What I really wanted to do was work on computers, not apply them to problems.” rings so true with me, and he as a natural leader also makes me look up to him.

[kelsey98765431]

Difficult realm for me to investigate myself. I have found that better than any book is to pull old VMS images out and get them up on a SIMH/Vax emulator and poke around! It's a quite laborious process and documentation of the process is sparse, last time I ended up using the Computer History Museum remote access program to interact with ancient VMS.

To actually learn how to use the thing I ended up diving deep into textfiles.com archives for references to VMS and got the hang of the basics enough to start learning via the online (built in) documentation.

I would also point you towards usenet archives if you want to learn more about the development history of VMS, what what i have been able to piece together it was really rock solid and loved and quite a shame that "open"VMS is mostly just a legacy compatibility licensing agreement and the technology is in a bit of a glacial deep freeze.

The fact that every file on the entire OS had version control - by default - before 1990... that blows me away and I feel like there's so many beautiful secrets to learn from this codebase. I always imagined it would be written in some abstract arcane language, when I learned it was in C i got a bit scared knowing what microsoft calls C has very little resemblance to how C is typically used on unix likes, and with the knowledge that many VAX admins dumped VMS for BSD I was quite afraid I may be out of my element.

I was surprised to find the C library is essentially just posix, I felt pretty much completely at home on VMS compared to any version of windows I have had to write C for. Pleasant experience, shame what happened to it, wish there was more about its story told in an authoritative context like a book.

If you stumble across anything nifty please feel free to share I may come back and read this read again someday.

Cheers

[nickpeterson]

"Work on computers, not apply them to problems" feels like the kind of mentality that gets a lot of people into trouble :)

[hnthrowaway0328]

I think what he means is system programming, not application programming. But TBH I'm so sick of my DE job that I indeed don't care at all about business problems.

Sure we all need them for jobs, hey but that doesn't mean I have to love them. Would that impede my career progression? Sure, but as long as I can get into a system programming without being too close to business, I'm happy for life. Other people can climb the greasy pole, good luck for them.

[pjmlp]

Windows Internals book series, old MSTech Journal and DDJ issues, VMS manuals on digital archives.

[hnthrowaway0328]

Thanks, I have a book about win32 programming, published 25 years ago but seems to be still relevant. I'll read the internals after I get some ideas about user space programming first.

[justin66]

Windows 2000 source code is out there in the wild (I expect newer versions have been leaked as well) and has even been used as a teaching tool.

[lproven]

WTF?

> I heard the first version is close to VMS.

No.

> Is it true?

No.

> Is there any material on VMS?

Tons. It came with metres of docs. It is still on sale. You can get a freebie hobbyist VM image.

> OpenVMS is open sourced

No, it isn't.

> but the version is too high.

Not true.

Always, always check your assumptions before asking a question. If you want to ask "because (a) can I (b)" then check A is true before asking. ALWAYS.

[hexagonwin]

Huh? OpenVMS was open sourced?