Open Source and Russell’s Paradox – A New Commons?

Jesus Villasante, a senior official at the European Union Commission in charge of Software Technologies in a spontaneous panel discussion on open source decided to shine a light on the lack of coordination, the influence of commercial interests, and the inability to evolve beyond current corporate paradigms. “Companies are using the potential of communities as subcontractors–the open-source community today (is a) subcontractor of American multinationals.” Mr. Villasante is completely correct in his analysis, although I doubt he will find many who will agree in either the corporate camp or in the open source camps.

Well, I do have extensive experience in this area. I co-pioneered the first Berkeley open source operating system using our own personal resources and with the support of the editor and staff of one of the most popular American technical magazines at the time – Dr. Dobbs Journal. Along with doing a complete novel port to the X86 architecture and new architectural design, we documented the porting process, kernel and architecture. We did not believe that simply reading the source was sufficient to understanding, and it required extensive documentation and open design review to provide releases that were reliable enough for researchers (much less consumers). This was the Berkeley way, and we cleaved to it with the support of the technical press.

Even with many releases, there was extensive documentation and careful attention to design trade-offs and discussion – for example, “Is it better to commit time to a patch if the Berkeley architectural goals from 1982 from Dennis Ritchie et.al. intended a new subsystem?” and “Are artifacts such as brk() wise in perpetuating so that some legacy programs can be run while impeding evolution in modularity?”. We addressed these and many other issues.

However, what we found is that various interests in the open source and business communities did not wish any new paradigms or evolution of the operating system if it required any changes to legacy applications (some suspiciously acquired). And this was the beginnings of a lot a bad blood in the Unix community.

You Never Know Who’s Watching You

Declan McCullagh of Cnet posted an item last week about Maureen O’Gara and Groklaw which spilled over into the bizarre world of open source paranoia. According to McCullagh, “Maureen O’Gara, a freelance writer who pens the weekly LinuxGram, alleged that Groklaw blog author Pamela Jones is a ’61-year-old Jehovah’s Witness with religious tracts in her backseat.’ O’Gara said she personally visited what appeared to be Jones’ apartment and Jones’ mother’s home in the New York City area.”

While that is mildly amusing, it’s not really surprising. The net allows people to assume, let’s call them “avatars”, that mask the real person with all their consequent flaws and frailties. But anonymity isn’t a Constitutional right, especially when you take center stage in a legal battle, as Groklaw has done. In fact, why be anonymous at all? Since Ms. Jones has lots of supporters who like her work, what’s the problem?

Sun Pats Rump SCO – Tarantella Cashes Out After Lots of Agony

A teeny tiny acquisition announcement brought back a lot of memories today.

Remember Santa Cruz Operation – no, not the SCO you read about fighting IBM and Novell, but the “old SCO”? Bob Greenberg and friends did a very brain-damaged version of Unix for the PC (originally derived from Version 7 and System 3) way back in the dark ages. Bob had done a Version 6 Unix derivative for the RAND Corporation called BobG Unix. The group was spun (thrown? or maybe walked?) out of Microsoft because Microsoft really didn’t want a Unix system if it wasn’t written in BASIC.

In 1982, Intel offered Symmetric Computer Systems CEO and Founder William Jolitz a great deal on 286 processors when he was deciding on processor bids to use in their new workstation funded by Technology Funding Partners (Symmetric’s lead venture firm). There were lots of problems with the 286 and Unix: 1) the instructions were not restartable, so if the operation could not complete (like memory wasn’t loaded) you could not reliably reload the instruction (there were steppings that supposedly could, but you never knew what you’d get), 2) the only way the address space was made large was by the reloading of segments – we’d encountered this problem before with the PDP-11 (William Jolitz work as an undergrad was on overlays for the PDP-11, so he was very familiar with this problem) – performance goes to hell when you move data from overlapping 64 kbyte segments to other segments, with all bets off when you hit an exception during that time, and 3) the calls within the segment and intrasegment calls made for variable sized stack frames, and the way Intel, Microsoft and others agreed on stack frame layout required a major rewrite in Unix – ironically, this made it difficult for early Windows programs derived from DOS as well. From these issues, we knew 286 Unix would never be a successful product, because there were too many compromises to move too many software packages from architectures like the VAX to it. SCO went ahead and made a Xenix based on the 286 – took them three years and a lot of work – and it was still a disappointment.

Open Source – The Times They Are A Changing

Three very interesting little open source stories passed my desk recently that I found shone facets on open source issues.

Last week, the Industrial Commercial Bank of China has signed a deal with Unix-clone Turbolinux to run open-source software in all of the bank’s operations. “Linux deployment is growing in China, with software makers targeting segments such as banking, insurance and wireless applications. Intel last year began a program to boost sales in China of desktop computers based on Linux.” Perhaps Microsoft shouldn’t look to China for much market growth.

Meanwhile, domestic open source companies have also been pursuing revenue through maintanance fees. According to Martin LaMonica of Cnet “In the absence of software license fees, open-source companies are adopting a services-intensive business model, accelerating an industrywide shift toward ongoing, rather than up-front, revenue.” Why would anyone want to pay up-front for software, when you can try it out and pay as you go?

Finally, open source companies, the pariahs of venture capitalists, are finally beginning to get some respect. When William Jolitz and I released to the public after 3 years of work in 1992, there was no serious business model for open source companies – especially an operating system. Companies bundled Unix on their brand of hardware, like Sun or Symmetric Computer Systems. Microsoft had the X86 desktop “locked up”. No one could compete with Microsoft by “giving away software”!

But the times, they are a changing. According to Gary Rivlin of the New York Times, “The first time Marc Fleury tried to raise money for his technology start-up company, in mid-2000, a venture capitalist told him that he didn’t have merely a bad business plan but a terrible one. Not only was Fleury planning to compete against the likes of IBM, but his product was open-source software, which he would give away. Four years later, he tried again. His business was still based on the free distribution of code, yet now there was a dogfight among venture capitalists competing to finance his company, called JBoss.In February 2004, JBoss received a combined $10 million from two prominent venture capital firms: Accel Partners in Palo Alto, Calif., and Matrix Partners in Waltham, Mass.”

Fun Friday – the Curse of BSD and the Four Mistakes

My discussion earlier this week on inaccuracies in papers discussing the evolution and history of resulted in some very interesting questions (see Oh, Goodie! Another Academic on 386BSD…). And a really nice question from suresh at Berkeley: “I’d like to hear your opinion: why did BSD lose to Linux in the battle for OSS hegemony..? How was the BSD release architecture (e.g. what was the political process) decided on? Some friends of fine [sic] swear by the FreeBSD operating system, but they are a minority.”

So Fun Friday is answering the really simple question “Why Did BSD Fail?”. Oh, this is a big one, but I’m game if you are…

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 ?” attached to the email. With the John Adams philosophy that “facts are stubborn things” firmly in place, I perused it, leaving errors outside of for others to find.

Oh, boy. I found it to contain serious inaccuracies with respect to the history of – which is absolutely amazing for an academic paper since 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…

Fun Friday: What If We Built an Operating System, and Nobody Came?

Well, Red Hat put lots of time and money into creating a professional developer version of Linux, put it on the market at $2,500 per “computer”, and in two weeks a clone of it called CentOS done by a squad of open source developers was put on the net for free. It’s hard to compete with “free”.

According to Stephen Shankland of Cnet “It’s clear, however, that many Red Hat clone users aren’t likely to embrace the original anytime soon. ‘I don’t pay for Linux, and I have absolutely no need for a Red Hat-style subscription (for) support,’ said Collins Richey, a Denver Linux enthusiast who uses CentOS on his personal computers to keep them compatible with work machines. ‘I’m considering recommending CentOS for limited use as a trial project…at work’.”

Red Hat tries to put a positive spin on it, saying “If they try versions that are not supported or supported inadequately, they will get a hint of the value propositions that are available for Linux and ultimately turn to a company that can support their businesses,” (Leigh Day, Red Hat spokeswoman). Some most assuredly will. But if my bit of experience in this area is any indicator, I believe that customers will wait until it is hacked, cudgeled, and otherwise moulded until it becomes good enough to be supported in-house. And still remains free. It may not be profitable to Red Hat, but it is “free enterprise” at its finest.

Fun Friday: Jimmy Dolittle and the BSD Bomb

BSD (Berkeley Software Distribution) has had a long and somewhat checkered history full of avarice and heartbreak. While revered (and still used) by many, BSD releases have been 1) obviated by “better” proprietary systems (e.g. SunOS to Solaris), 2) licensed to death (AT&T / USL / whoever), 3) unlicensed and released to great acclaim and even great expectations (386BSD), and 4) once some flaw is found, dispised, derived, hacked, and then poorly marketed against juggernaut Linux and Microsoft (NetBSD, FreeBSD, YourNameHereBSD,…). Unlike , which stayed focussed on the BSD research goals and writings which are Berkeley’s best quality (see 386BSD Release 1.0 Reference CD-ROM: Essays on Kernel Design), ‘s many “commercial” derivations never achieved the kind of monetary success expected after The Fun with 386BSD we all had.

Why was this? Perhaps William Jolitz chatting with Tom Foremski of SiliconValleyWatcher (see How I learned to love Linux and profit from it — Wind River turns from Linux basher to religious zealot) may provide a bit of insight as to why the Curse of Commercial BSD continues: “Back in 2001, met with Wind River’s co-founder and board member Jerry Fiddler, along with John Fogelin at their office in Alameda, about the BSD purchase. It wasn’t a confident feel in the room, and they had no interest in putting any more “wood behind the arrow”. Just then, several B-25 Liberators flew overhead, commemorating Jimmy Dolittle’s raid on Tokyo. As they rumbled past, I recall thinking that the bomb Jerry bought was going to be bigger than the ones Jimmy dropped. Ironically, I was an executive at a Japanese company at the time. But such has been the BSD karma. You can’t say it doesn’t have its humorous side!”

Would You Like Your Disk Drive with Extra Aggregate?

Matt Marshall mentioned an old defunct company that I was rather fond of – Miniscribe. Now, Miniscribe in the 1980’s went from nothing to making and selling quite reliable 85 MByte drives (full size) at what was then a really great price (around $850 in quantity). It was the most common disk drive we shipped in our Symmetric 375 computer running the Symmetrix operating system (Berkeley Unix 4.2 derivative) from Symmetric Computer Systems.

Of course, having bought, installed, and supported so many, I just couldn’t resist putting in my two cents with Matt as well: “Ah, Miniscribe. I recall their 85MB drives well – we shipped many Symmetric 375‘s with those drives. They were quite cheap and quite good. But I also remember how they tried to claim we signed for a pallet never delivered. Turned out the signature was an obvious fake, and Miniscribe dropped the claim. We never did get those bricks…. We did however have one of their drives catch fire in testing – came into the office to the smell of phenols and a “squeek, squeek” sound of a slowly (for a disk drive) turning disk. I bet the engineers at Miniscribe spent a lot of time on that RMA!”

That Miniscribe disk drive failure was about as memorable as the time we tried a new “under the chip” capacitor on several 375 motherboards (I still have a few around if you’d like to see them). The motherboards were very tightly designed, and the majority of boardspace was taken up by DRAM. They were really cool – until one caught fire under the chip! Of course, surface mount technology caused much of this stuff to fade into obsolescence.

The League of Extraordinary OSes

Every now and then I get handed a paper and asked for feedback. Most of the time, these papers have long boring titles and lots of funny charts with red error bars on them. I’ve even written a few of them myself, so I guess that’s why I get handed more and more. But go ahead – I enjoy another article on clustering (Gordon Bell handed me one a while back) or TOE’s (that was one of the SiliconTCP boys). I’ll even look at the “let’s drop TCP and make it fast” papers, because sometimes the authors are actually pinpointing a real-world problem even if I don’t agree with their proposed solution.

But I was recently referred to a paper “Open Innovation: The Paradox of Firm Investment in Open Source Software” by Gallagher and West precisely because it related to the evolution of open source, and since I’m often referred to as a “Pioneer of Open Source” with , this is another topic on which I get requests for feedback. So I read the paper discussed, and found their discussion of proprietary work quite good. However, I also found that their isolation of time in studies (1998 onwards) actually missed the primary evolution of every single open source model cited in the paper, which misses the point of the exercise, likely due to ignorance of the topic. So perhaps an examination of Berkeley’s influence in this regard would be a valuable addition.