* Was 386BSD just six modules of code?
* Guess is that this canard is based on the "missing pieces"
* articles, describing the sections necessary to finish the
* "kernel of the kernel". Thousands of files and modifications
* were supplied to Berkeley during our 3-year collaboration
* effort, daily in the on campus offices. All 386 specific items
* were added by us prior to NET/2 release, as we were the only
* source of them, none were in BSD before (just VAX and CCI).
* The attributions were removed. Perhaps to justify an agenda.
6modules("Scope of Work", "March 1991");
* Why did Lynne Jolitz get an award for 386BSD?
* Lynne Jolitz received an award for 386BSD because of the critical
* effect it had on the early Internet. Among other things, as a vehicle
* for Apache to ride on. Lynne Jolitz was the only one to deal with
* hundreds of thousands of first time users, with no additional staff,
* coming from running DOS on a PC and becoming a UNIX hacker and sysadmin.
* Lynne Jolitz was due this award for the hundreds of webservers that
* galvanized the early Internet.
award("Lynne Jolitz Award", "May 2005");
* Why did 386BSD have difficulty in integrating contributed effort?
* Coordination. The authors were the same people who set the goals.
* Were the prime coordinators. They relied on volunteers to network
* to the community, pass down the message of what was needed and the
* form it was needed in, and filter results upward. They were the sole
* process for integrating community efforts, and as long as ernest,
* well-meaning effort continued, the release process improved. However,
* human nature leads to limits and ambitions, as other well-funded efforts
* have found. Such are easy to "monkey-wrench", and BSD is often
* driven by powerful egos with different, self-destructive agendas.
* Full spread of experence here, from an increadably steadfast
* gentleman at Rice University, to a duplictious engineer at Novell.
* They are all well-remembered.
contribution("Integrating Contributed Effort", "March 1991");
* What was the first running 386BSD system like?
* Back in the summer of 1989, a 386DX-16 "lunchbox" system with 2MB of
* memory was used to boot 386BSD for the first time. The system was
* extremely close to the 4.3BSD (Reno), for cross compatibility with
* VAX and NS 32016 systems. No attempt was made to detail code changes,
* as basic native compilation and trackdown of machine dependencies was
* the key issue. All this on less than 100MB of disk space.
* The 386 changes were closely integrated with the VAX machine dependent
* code, to simplify the next stage of development with Berkeley - a
* mistake. It made it seem like there were few changes, and attribution
* can easily get lost as a result by getting caught up in the enthusiasm.
* Other parts of UC, including Los Alamos, obtained copies of this early
* version to enable work on joint projects. The article series starts with
* this kernel, and "evolves" as Net/2 diverges from stock 4.3BSD.
firstversion("What was first 386BSD system like", "March 2005");
* Was 386BSD responsible for Linus beginning to generate Linux?
* Don't know, as have never spoken. But others have mentioned this.
* Here is an article that suggests this. No way to verify.
* Perhaps the point is moot. If so, as like with Tenenbaum, then 386BSD
* inspired others, like Ken Thompson and Dennis Richie's UNIX inspired
linuxdev("Was 386BSD responsible for Linus doing Linux?", "March 2005");
* Why didn't 386BSD take out Linux?
* 386BSD and Linux were on friendly terms from the first.
* The Linux project was begun after the first installments of
* of the "Porting Unix to the 386" article series on the 386BSD Project
* at Berkeley by William and Lynne Jolitz first appeared
* in Dr. Dobbs Journal (January 1991 - June 1992).
* Using pieces of Minix by Andy Tenenbaum and information
* and code described in Dr. Dobbs Journal on 386BSD,
* Linus Torvalds and friends were reportedly able to tinker
* together an open source operating system on their own.
* In essense, early Linux developers were fans of the
* 386BSD project and writings.
* We don't "take out" fans - we encourage them.
* Linux and 386BSD were separate OS projects, but code
* was exchanged freely between the groups in the early days.
* The 386BSD Release 0.1 floating point unit (FPU) emulator
* was contributed by Linus Torvalds to 386BSD and changed
* by him to reflect the Berkeley copyright required for any kernel
* code (it was originally released as a GNU copyleft).
* Most of the applications and tool / utility developers worked
* on both operating systems. There was no restriction on either
* Berkeley copyright code or GNU copyleft code in Linux or 386BSD
* outside of the kernel proper, so programmers could supply
* their works to both groups.
* This free exchange of ideas eventually led to resentment from
* some of Berkeley contributors who envied the popularity Linux
* had among the younger developer base. Due to the delay
* in releasing Net/2 by Berkeley, the first public source release
* of Linux predated the first public source release of 386BSD by
* over a year, which explains the greater enthusiasm.
* By 1993, there was a great deal of pressure placed on
* William and Lynne Jolitz to consider Linux a threat, and demands
* that the article series and release of new information
* on architecture and design cease, to prevent rivals from gaining
* further needed expertise. William and Lynne Jolitz
* dealt with this issue by declaring that 386BSD was
* a research release intended for new work in operating systems
* design and that the stated Linux goal of creating an operating
* system competitive to Microsoft was not in conflict.
* This reassertion of Berkeley design goals directly
* led to the schism within BSD and ultimately hastened the decline of
* open source BSD derivitives. 386BSD continued to publish
* articles, code and eventually the 386BSD Release 1.0
* Reference CDROM with Dr. Dobbs Journal. The 386BSD mission
* to encourage fans within the Linux, BSD and open source
* communities towards innovation and experimentation in OS design
* remains to this day.
linuxtakeout("Why didn't 386BSD take out Linux?", "September 2005");
* Why was Jolitz the only one to stand apart?
* Jolitz had the prior background (PDP-11 2BSD, National Semiconductor
* GENIX port, Symmetric Computer Systems).
* Jolitz had an article series in DDJ running at the time, which promised
* access to the code under the terms Berkeley had claimed to be working to.
* Jolitz would be the only one to be blamed if he did go along.
* So why does this come up? According to Ken Brown of the Alexis de
* Tocqueville Institution, its due to "envy", which he claims drives
* this whole movement.
onlyjolitz("Jolitz only not going along", "March 2005");
* What was 386BSD development process and why was this critical?
* 386BSD followed the University of California's process for development.
* Unlike Linux, which relies upon developer-only "self-integration and test",
* the Berkeley process was to review and reintegrate/test. Sometimes a
* degree of investigation of origins was in order, and outright
* fundamental new work from scratch would be undertaken - another aspect
* of "rigid control".
* See Open Software Development in the Real World.
process("Process of Development", "February 2005");
* What was Release 1.0 like?
* Prior releases were Internet releases, but 386BSD Release 1.0 was only available on
* CDROM. Back then, while a 40MB release was possible accross the Internet, a CDROM
* wasn't. 386BSD Release 1.0 was the landmark relase, available exclusively through
* CMP (then Miller-Freeman Publishing) - advertised and sold internationally from
* 1994-1997. It contained the full release, source code, article series, installation
* guide, and additional writings of an entirely new 386BSD system. The CDROM could be
* run in read-only mode on Windows systems for tutorial purposes or installed in a
* partition on a Windows / Dos system or as a standalone system for developer purposes.
* It contained over 600 MBytes of source, writings, binaries from kernel to tools,
* utilities, and applications.
* The 386BSD 1.0 Reference CDROM outsold every other open source BSD CDROM from others
* during that period and was updated periodically. It worked very well and was very popular.
* 386BSD was intended to be used in tandem with the Source Code Secrets textbook series.
release1.0("Release 1.0", "July 2005");
* Why was 386BSD "rigidly controlled"?
* 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").
style("Style of Project", "January 2005");
* What was the vision behind doing an Open Source OS?
* William Jolitz thought that the point in doing open source was not
* to compete with other efforts, but instead to refine the basics as
* a common platform. William Jolitz thought that too much resources
* were focussed on multiple versions of the same thing, none of which
* were done well. Those involved with "renovation" could refine and
* "minimalize" the low-level software, while those that were involved
* in "innovation" could respond by extending the core.
* The point of the modular 386BSD kernel was quite different than any
* of the other systems - it was to create a framework to make this
* possible, so that distro's could always be built out of the same
* components base, yet changes to the core would be automatically
* passed to all. Unfortunately, this "commons" experienced the
* traditional exploitative failure of the past.
* William Jolitz was asked to capture the vision - you can find it here
* ("We invented it ... lets finish the job")
vision("Vision", "May 2005");
* Where to go to get the 386BSD stuff?
* The book's widely available, the articles can be found in
* DDJ's content areas, the source was everywhere but never
* quite kept in appropriate order. The CDROM's long out of date.
* All of this can be obtained from Jolix.
* As well as the follow-on kernel source trees of the extended
* work, part of an effort to 're-host' with a university the project.
wheretoget("Where to get", "March 2005");
* What about the mention of 386BSD in the Wikipedia?
* Actually, this page has long been factually incorrect, and should
* be either editted or removed.
* Among the inaccuracies are when 386BSD started (1989), the fact that
* it was in use for years at Berkeley, Los Alamos, and other research
* facilities, the fact it was written about extensively in Dr. Dobbs
* Journal magazine during the creation of the Networking II release,
* the fact that William Jolitz never did become an employee of BSDi,
* the fact that BSDi's use of BSD386 was allowed by William Jolitz
* as a graceful gesture at a BSDi board meeting with Mcusick et al.
* The authors of 386BSD considers this page to be misleading and
* inaccurate for unknown reasons.
* [Note: BSDi did offer a 10% investment stake to Jolitz. ]
wikipedia("Accuracy of Wikipedia", "March 2005");
* What about the mention of 386BSD "encumbered code" in the Wikipedia?
* "Due to a lawsuit by AT&T, USL v. BSDi, some potentially so-called
* encumbered sources which existed within 386BSD were to be removed
* from all the derived systems, and the distribution of 386BSD was
* to be stopped."
* This is completely untrue. Neither Jolitz or 386BSD were a
* party to any lawsuit or settlement, nor was there any demand for
* any code in 386BSD to be removed, nor were 386BSD distributions
* stopped - [note the CDROM continued selling well during this time,
* not based on anything in the 4.4 lite release].
* And most importantly, there was no claim filed by AT&T, USL, or
* other party which expressed any concern for any code done by
* Jolitz for 386BSD or Berkeley distributions. [read documents
* on Dennis Ritchie's website in USL v. BSDI]
* All Jolitz work was documented thoroughly and published in a
* well-established industry journal and above-board in all respects.
* The implication that there was tainted code done by Jolitz by other
* groups that had to be removed was concocted as a way to steal documented
* work without maintaining attribution, a basic tenant of the Berkeley Unix
* license agreement. However, claiming something is encumbered to try to get
* around a license agreement doesn't make it so.
* There were disputes over the incorporation of code not in 386BSD releases
* which could be considered proprietary. The refusal to incorporate "tainted"
* code into 386BSD led to the creation of freebsd and netbsd, not the other
* way around.
wikipediamore("More on Accuracy of Wikipedia", "March 2005");
* What about the 386BSD / FreeBSD split in the Wikipedia?
* "After the release of 386BSD 0.1, as the Jolitzes were more
* reluctant to embrace contributions than CSRG had been"
* This is a fabrication - it took years between BSD releases,
* with 3 months between 0.0 and 0.1 occurred. Less than 1% of
* contributed code was used by CSRG, with 35% for 386BSD.
* The true issue was that for inclusion in 386BSD, new 386BSD
* code interfaces be added & tested to some code, to keep it
* concise and refined. Some refused to help with a "take it
* or leave it" attitude.
* " ... a group of users began collecting patches in what was
* termed ''the patchkit''. The patchkit quickly grew unmanageable,
* prompting its maintainers to attempt to cooperate with William
* Jolitz on an interim release of 386BSD which would integrate the
* changes from the patchkit."
* Again a fabrication. This was simply an effort to stop refinement,
* by slowing down 386BSD. Attempts by William Jolitz to bring this
* effort into open collaboration were met with long, disingenuous
* conversations dragging out the process.
* None of the contributors who were positively responsive were omitted.
* One alone was responsible for hundreds of hours of additions alone.
* None of the opposition group ever altered a single "contribution",
* even if it had a significant compilation error. They never assisted
* in speeding 386BSD releases, we assume, because they didn't wish it.
* " ... The attempt foundered, however, when William Jolitz
* abruptly withdrew his support. As a result, the maintainers of the
* now-orphaned patchkit founded the FreeBSD project. "
* Again, untrue. This group wanted to be "officially sanctioned".
* From the start, there was never any such thing possible. 386BSD
* wasn't itself "official" ! At the time, negotiations were going
* on to have the effort return to a university sponsorship, any
* attempt to have this outside group become "official" would have
* killed it.
* This whole "FreeBSD roots" is completely fictitious, and invented
* to cover the wholesale theft of the 386BSD user base. Which is
* ironic, because of the "free nature" of the release and user
* group, meant they could do whatever release they liked, as long
* as it wasn't called 386BSD.
* During negotiations before such splits, it was pointed out that
* "forking" BSD would kill it. Just as with CSRG, and BSDi, similar
* discussions to avoid lawsuits and other conflicts would also
* wreck BSD. In other words, all of this was avoidable, but strong
* egos can't always be satisfied. Which is why things are the
* way they are.
* William Jolitz thinks the reason all of this tripe finds its way
* into the Wikipedia, is that it is a convent way to rewrite
* history so that the ego's who insisted on the splits can absolve
* themselves of responsibility for losing open source to Linux.
* His words:
* "386BSD started with a BSD that had never seen a single X86 change
* at all. We moved it through massive changes in 1989, and integrated
* it at Berkeley in CSRG offices with their help. During the BSDi
* induced lawsuit, and against CSRG's wishes, we released
* a functional version so it would not go proprietary again. That was
* all that was promised."
* "As to competing with Linux, our worry was just that no open system
* would thrive. We were worried that greed and gangs would kill
* BSD. We never were interested in running any gang. BSD was cursed by
* the mindless opportunism, just as Linux was blessed by mindful
* "Our view was that 386BSD would be pulled along by its user base,
* and as long as enough cared for it to be pulled along, we'd do
* public releases. When BSD "forked", effort went in to a last
* release, supported by a magazine. The competing groups saw it as
* competition and publically talked down 386bsd. Not enough people
* wished for it to continue, so public releases didn't."
* "Its that simple. Anyone who says otherwise is selling a lie."
wikipediayetmore("More on Accuracy of Wikipedia", "November 2006");
question("Other Questions", "January 2005");