[TUHS] Early Xenix source code (was: What UNIX Artifacts Are Still Missing?)

Paul Ruizendaal pnr at planet.nl
Thu Dec 7 22:09:21 AEST 2017

> They might actually.  Gates isn’t in charge, and there has been a major effort to being Linux compatibility into the Windows 10 kernel.

I agree that they might. Once there was strong commercial logic to disown their Unix history; that commercial logic may have reversed in the last decade.

> The biggest issue will be the never ending tangle of licenses, if they had other stuff integrated into there.

Let’s analyse that bit:

- They could pick the Nokia solution, i.e. to simply make an undertaking not to sue and thus avoid taking a position on Unix ownership etc.

- It would seem that Xenix 2 (no apparent version 1?) was more or less V7 and for internal use only (lacking a binary redistribution license, as Clem pointed out). Very little chance of 3rd party source code in there, but also of little interest for the historical record.

- The first real ports occur with Xenix 2.x in 1981-83. This would appear to have been based on System III with the PDP-11, Z8000, 68000 and 8086 as targets. Considering the size of MS at the time and how busy they were with IBM and DOS I don’t think they would have had much time to do more than a basic port: they contracted out much of the work to SCO, a two man shop at the time. It would seem that MS owned the IP with SCO being a contractor & part-owned subsidiary.

- This Altos manual https://web.archive.org/web/20170222131722/http://www.tenox.net/docs/xenix/Altos_Intro_to_Xenix.pdf actually says that Xenix 2.x was still based on V7 and would seem to be a vanilla port, i.e. unlikely to include other stuff than V7, some 2BSD and of course MS' own stuff.

- The next release, Xenix 3.0 in 83-84, still appears to be SysIII based and would remain easy from that perspective. However, given two years of polishing it may have picked up bits and pieces from the outside and the tool chains probably started to include MS’ own compilers, it would be harder to figure out what could be released. However, looking at this highly interesting leaflet http://www.os2museum.com/wp/wp-content/uploads/2012/10/IBM-Seminar-Proceedings-Vol2-No9.pdf it would seem that it is still all SysIII, BSD and Microsoft code.

In my opinion the real hurdle is finding a retired MS VP who’s interested in knocking on doors and making the case for this public relations move.


PS: There is scope for confusion over version numbers. It would seem that MS never directly sold Xenix, only via OEM’s. For example, IBM PC Xenix 1.0 would appear to be the same as MS Xenix 3.0

More information about the TUHS mailing list