The "U" Word

Making Windows safe for Unix people since 1995

Unixfolks Anonymous

Unixfolks Anonymous

  • Comments 5
  • Likes

Me: Hi, my name is Jason, and I'm a Unix guy.

Crowd: HI, JASON!

Me: I started using Unix a long time ago, more than 20 years now. I started out real easy, real slow; Unix version 7 on a PDP-11. But the next year, there was some new hot stuff on the street - they called it BSD, and most folks had a VAX to use it, so that's what I stepped up to. Did some playing around with an early precursor, version 6; a Unix kernel so small one person could actually hold the entire thing in his or her head.

Most people experiment with that sort of thing during their University days, then grow out of it. Not me. I loved it so much, I got into the business. I went to work at Hewlett-Packard and worked on their Unix, HP-UX (good thing Dave Packard lost the coin toss with Bill Hewlett, eh?); if one Unix system gave someone a buzz, networking lots of systems together could give'em a heck of a head rush, so that's what I did. Did some TCP/IP implementation work, some formal network standards stuff (ISO, IEEE, X/Open and Open Group), some network management.

Then I went to HaL Computer Systems in Austin, TX. Spent 2 1/2 years there. Whoa. Bad trip. Working at HaL, was, well, a four-letter word beginning and ending with the same letters as the company name. I learned some stuff:

  • If you have to live in Texas, Austin is the place to live.
  • I do not have to live in Texas.

Round about that time (1995) I ran into a friend of mine (also a Unix guy), who turned me on to this completely new thing - the POSIX subsystem inside WinNT. Now, I'd used NT off and on for a couple of years, mostly for a change of pace. I could tell there were some cool things about it, but I had such a major Unix habit I couldn't imagine just switching. When this friend gave me the chance to combine the two - make something useful out of that POSIX thing running on NT - I jumped at it. He and I and a couple other guys founded a company.

Softway Systems built a product called OpenNT, which was an enhanced version of the POSIX subsystem Microsoft had shipped in NT 3.x and 4.0. We shipped five versions of the product in the four years of our existence; changed the name to Interix, too.

Then we got acquired by Microsoft in September 1999. A Unix guy, inside Microsoft. I figured we few Softway Systems survivors would be lonely, huddling in a dark corner of a parking garage trading shell scripts and arcane API calls when the Win32 Police weren't looking. It wasn't nearly that bad. Kinda fun, even.

MS rolled the technology out as Microsoft Interix 2.2, then combined it with their existing NFS client/server product, shipping Services For Unix 3.0 in May of 2002. In January 2003, the LinuxWorld trade show named SFU 3.0 “Best Integration Product 2003”. (Can you just imagine what the typical Linux-head at the show must have thought about that?) We're rolling out SFU 3.5 at LinuxWorld this month.

Since SFU 3.0 shipped, I've been working on the Microsoft Solution for Unix Migration. “There's no one so gung-ho Windows as a reformed Unix user” may have been the thinking, but that's too-simple-minded a way of looking at things. I still see myself as a Unix guy. My mission at Softway Systems was “Make Windows safe for Unix people”; that what I do today. And I still love it.


  • Clever wordplay. I enjoy your style of writing!

  • Unix people in Microsoft? Do you recall who lost the coin toss to whom when Microsoft decided to divest itself of Xenix?

    You mention 6th edition Unix. Possibly could you confirm or refute my recollection that file naming in 6th edition was case-insensitive? At present there's a big dispute in LKML about whether and how much assistance they want to allow for Samba's need to provide case-insensitivity. Some guy named Linus something-or-other has no tolerance for case tolerance. But I'm pretty sure that in Unix's ancient history, just like with Windows and VMS and some other systems, I didn't have to memorize exactly which characters to shift when typing filenames. (Though even back then it was already necessary to memorize which characters to capitalize in C identifiers.)

    Hmm. Is SFU the reason why Microsoft paid a licensing fee to Xenix's heir?

  • 6th edition had case-sensitive files, but file size was limited to 24 bits. The lseek() API didn't exist yet; seek() took an odd data structure for a file pointer. Version 7 went to a 32-bit datatype for file offset and replaced seek() with lseek().

    SFU has existed as a product at Microsoft since 1999.

  • > 6th edition had case-sensitive files

    OK. My former recollection needs to be garbage-collected. Thank you.

    > SFU has existed as a product at Microsoft
    > since 1999.

    Sure, but in 1999 SCO/Caldera/SCO wasn't demanding licence fees from everyone who acquired Unix or Unix-alikes via other routes. In 2003 a sometime owner of Xenix started demanding licence fees from all users of Unix and Unix-alikes, and a certain famous previous owner of Xenix actually paid a licence fee. So I wondered if the existence of SFU was the reason for paying.

  • Microsoft licensed source code from SCO because SCO had some intellectual property Microsoft was interested in for future products. That's the truth of the matter, and it's Microsoft's stated answer to the question.

    To the best of my knowledge (which is pretty deep), there's nothing in SFU 3.5 or earlier to which SCO could lay claim. You can garbage-collect any assertions to the contrary. :-)