Oh, Goodie! Another Academic on 386BSD...

Another paper handed to me, this one on "open source governance" (isn't that a bit of a oxymoron?), with the usual "Isn't this wrong about 386BSD?" attached to the email. With the John Adams philosophy that "facts are stubborn things" firmly in place, I perused it, leaving errors outside of 386BSD for others to find.

Oh, boy. I found it to contain serious inaccuracies with respect to the history of 386BSD - which is absolutely amazing for an academic paper since 386BSD was extensively written about in one of the lead trade magazines of the time - Dr. Dobbs Journal - in a 17-part series Porting Unix to the 386 documenting it's evolution, and also distributed through the magazine, and had multiple releases via the net per standard Berkeley Software Distribution methods. So it's not as if one can't find lots of source material from the authors. But this paper is riddled with errors with respect to release governance, intentions and motivations, and control - and that pretty much covers everything in "governance", doesn't it? So here's the real story...

It is true that 386BSD predates Linux, and I have heard from many that Linus used the article series and Andy Tenenbaum's work in his effort. In fact, there was much cross-pollination between groups early on - Linus himself contributed the floating point emulator to 386BSD and changed the GNU license to the BSD license so it could be used (we followed the Berkeley license in the kernel, as all prior Berkeley releases had done).

His stated goal at that time was that of competing with Microsoft. I believe he has succeeded.

The stated goal of 386BSD from the beginning - published and documented - was to continue the Berkeley tradition of innovation and new work in operating systems design. We began a publishing relationship with Dr. Dobbs Journal under the encouragement of the editor-in-chief, Jon Erickson, and wrote extensively about kernel architecture and design from 1991 onwards. We also conducted several releases via the Internet.

Unlike Linus, we were working with a publisher and focused only on documenting kernel architecture and design and we kept our word with the publisher, editor, and readers of Dr. Dobbs Journal. Development, especially outside the kernel, was not rigidly controlled - I think the 386BSD FAQ discussion is quite apt here:

"* Another canard often wondered about. If anything, the style of
* 386BSD comes from university projects and memory of
* the Homebrew Computer Club. Would guess this is again to justify
* another agenda, by taking the tightly focussed architectural
* objective and unfairly slanting this as "rigid control".
* Often times we communicated back to our coordinators that we
* didn't care how something was done, so long as it followed the
* objectives, goals requirements("original work owned by them")
* and process("self-integration and test"). "

A good academic always questions the agendas of rivals and opportunists as well those of the creators of the work. Our work, goals, and opinions were and are extensively published at the time 386BSD was discussed in reputable journals. We stand by our published work, and would wish they be as commonly cited as postings on the Internet in a discussion of our release methodology and governance.

The culmination of our work was 386BSD Release 1.0, a comprehensive work in OS design and architecture, including kernel source code annotations, discussions of trade-offs in design, history and legacy issues, and future perspectives, sold 1994-1997. It was very popular. This publishing project end-to-end was five years - a not inconsiderable period to dedicate to new design and published works.

However, 386BSD was *not* intended to be a competitor to Microsoft, nor were we interested in legacy support issues or other political agendas. 386BSD was a fun project and came out of years of experience with Berkeley and Unix. Having been a part of a number of startups, we did not see (nor did any investor or VC at that time) any business opportunity.

I find, in particular, Eric S. Raymond to be an unreliable source with respect to 386BSD. I don't know him, he's not a Berkeley alumnus and was not part of earlier BSD releases, and he has never been part of the 386BSD development team or publishing efforts. There are many other people of reputation who were personally involved in the decisions at the time, but he was not one of them. Since we all are still very much alive and involved in the computer industry, perhaps it would be of more value to "talk to the source"?

If any reader would like to read more about some the experiences, the 386BSD FAQ is most useful, as it also addresses several other erroneous claims with respect to 386BSD that seem to circulate on the Internet. As I'm sure everyone is aware, the Internet is full of hawkers and gawkers, and you can't believe everything you read simply because someone posts it. The best research materials are considered published works from primary sources who were actively involved at the time. There was considerable momentum to publish, and these works, while very technical in nature, are still available (and still quite popular - you have to pay Dr. Dobbs Journal to get them).

Well, that's about it. I wish that these phony papers dressed up as "research" would get the kind of professorial pushback we got a Berkeley. But I suppose that's what a first rate university like Berkeley is all about - telling the truth, and making it stick.