[TUHS] Dave Cutler recollection about Xenix

Paul Ruizendaal pnr at planet.nl
Sun Oct 22 01:36:30 AEST 2023

An interesting set of videos indeed, although I wish they were not all chopped up in 5 minute segments.

> I consistently hear from folks the same about Bill Gates pushing for volume over anything else with Xenix.

That was his business model. His Basic for the 8080 was copied a lot (the famous 1976 open letter to hobbyists) and he shifted to selling bulk licenses to manufacturers. These could then make a bundled hw/sw sale and sidestep the copying. If I understood correctly, in the early days he sold the bulk licenses for a fixed amount, without per copy fees. I suppose this matched his cost structure, so it worked; the leverage and profit came from selling the same to all manufacturers in the market. He also used it in his deal with IBM, beating out Digital Research that wanted per copy fees. Retaining the rights to DOS also matched the business model that had been pioneered for his Basic.

It would seem that the same thinking was at play in the deal for Xenix (which I think preceded the IBM deal). He would spend money once on porting Unix to each of the various next-gen microprocessors of the time (x86, Z8000, 68K, NS32K) and sell (sub-)licenses to hardware manufacturers, who in turn had a right to sub-license binaries to end-users. The deal that he had to negotiate with Bell had to match that business model.

Beyond this, I’m sure that Bill Gates understood the strong network effects in software and the "winner takes all” dynamic that results from it -- hence his focus on volume and market share. However, I don’t think this drove the structure of his 1979 [?] Unix license deal with Bell.

> Something this brings back to mind that I always wonder about with Microsoft and their OS choices: So they went with Windows NT for their kernel, scraped the Windows environment off the top of DOS and dolloped it on top. Has there been any explanation over the years why they also decided to keep the MSDOS CLI interface?

The below site has a very nice summary of Xenix at Microsoft (I’ve linked it a couple of times before):

About blending Xenix and DOS it says: "As late as the beginning of 1985, there was some debate inside of Microsoft whether Xenix should be the 16-bit “successor” to DOS; for a variety of reasons – mostly having to do with licensing, royalties, and ownership of the code, but also involving a certain amount of ego and politics – MS and IBM decided to pursue OS/2 instead. That marked the end of any further Xenix investment at Microsoft, and the group was left to slowly atrophy.”

Probably that same dynamic was in play for the CLI of Windows NT. Moreover, as you already point out, by the time of NT there were tens of millions of users of DOS, and numerous books, magazines, etc. explaining it. Throwing away that familiarity for unclear benefits (in the eyes of those users) would serve no business purpose. In a way it is the same dynamic that kept C89 and Bash in place for so long: people know it, it is good enough and it works everywhere.


Seeing the Cutler interviews reminded me of the old joke that there are only two operating systems left: Unix and VMS (Linux being Unix-family and Windows being VMS-family). I wonder if we will see it narrow down to just one before the hardware changes so much that the concept of an OS changes beyond recognition. My hypothesis would be that an entirely new approach will come first.

More information about the TUHS mailing list