ONLamp.com    
 Published on ONLamp.com (http://www.onlamp.com/)
 See this if you're having trouble printing code examples


A Look Back at 10 Years of OSI

by Federico Biancuzzi
02/12/2008

The end of the '90s was a great epoch for the software world. We lived a revolution without being aware...

The spirit of open source and the importance of free software spread worldwide, and ten years later we take as habit things that were completely unthinkable ten years ago.

Sun buying MySQL? Microsoft releasing open source software? $200 Linux PCs at WalMart? Governments that switch to Open Source systems? PDAs, phones, and consoles running free software?

If you want to learn more about those years, you should read the open book Open Sources: Voices from the Open Source Revolution.

During February 1998, Eric Raymond and Bruce Perens founded OSI, the Open Source Initiative, with the goal of promoting Open Source culture, especially in the business world.

After 10 years of activity, the foundation has reached many unbelieveable goals, and it has a great future ahead.

To celebrate the moment, Federico Biancuzzi interviewed the two founders, Bruce Perens and Eric Raymond, Brian Behlendorf (a member of the first Board of Directors, the others were Ian Murdock, Russ Nelson, and Chip Salzenberg), and Michael Tiemann, the current President.

Let's start with Perens' interview...

Federico Biancuzzi: What dreams and goals (about open source) did you have when you co-founded OSI?

Bruce Perens: Dreams? What happened was far greater than I dared to dream. Open Source is a key part of enterprise computing, and government computing, and it's in very many people's homes, mostly playing roles they don't notice. It's taken over the biggest computer firms in the world. I talk with a lot of companies, because I make a living helping them make corporate policies and processes for working with Open Source. We find that there are two kinds of companies: those whose executives know they're using Open Source, and those who are using it, but their front office hasn't come to terms with the fact yet. You don't have to sell anyone on Open Source any longer, they already have it.

One of my projects today is working on Open Source and Open Standards for the AMSAT satellites, primarily an upcoming geostationary satellite. What could be more fun than having your work go in space? Talk about deams, I'm living them!

Biancuzzi: Why did you leave OSI?

Perens: You are the first person to ask that question in at least five years. Rather than cry over spilled milk, let's be constructive: it would be good for OSI for me to be on its board now. It doesn't seem to be hurting me that I'm not on it, though. I get the same opportunities to promote Open Source whether or not I'm on OSI's board: meeting with national government leaders and top CEOs, addressing the UN, etc.

Biancuzzi: What is your point of view on OSI activity since you left?

Perens: I'm concerned that they hold the line, rather than allow the meaning of Open Source to be diluted any further than they already have allowed. The most dangerous area they've tread recently is the difference between fair attribution of the software creator and "badgeware." Socialtext's license was as mild as possible and there was precedent for it in the attribution requirements of the GPL, etc., but it potentially creates a slippery slope to much worse restrictions in licenses that would effectively handicap the Open Source software under those licenses for commercial use, necessary forks, etc. That would hurt all of us.

I'm especially opposed to licenses that place any legal burden on a mere user, rather than someone who modifies or redistributes. One of the goals for Open Source was that users should be able to just use it, without reading the license or having to hire a lawyer. To some extent, software patents break that for us, for some businesses, but that's not our doing. We shouldn't create additional hurdles to simple users if we can avoid it.

And I regret that Open Source has sometimes been used to deprecate Richard Stallman and his promotion of Free Software. I have always maintained that Open Source and Free Software are two ways of talking about the same thing, to different kinds of people. Obviously Open Source arguments are tailored to business people, while other folks, for example programmers, have been more receptive to Richard's arguments about the importance of Freedom a priori. Many people have started to apprehend the arguments of Open Source first, and have been led to an appreciation of their Freedom and Stallman's philosophy once they understand why Open Source makes business sense.

Biancuzzi: Is the distinction between "free software" and "open source" still meaningful?

Perens: It was always a mistake. Open Source and Free Software are two different ways of talking about the same thing. One is written to appeal to programmers, the other to business people. Once they are into Open Source, the business people start to appreciate Stallman's ideas.

However, Richard feels differently about this issue.

Biancuzzi: Do you still see the same excitement towards open source present as the end of the '90s? Or it went away with the bubble? Or maybe it evolved in something different (business, for example)?

Perens: There's tons of excitement and it's nice that people who want to write Free Software can get jobs doing that. What is reduced today is the perception of an ethical and public-benefit side to writing Open Source: we're sharing our work with the public, that is a good thing which should be recognized. Too much of the focus today is on the folks whose main interest in Open Source is to cash in. That's wrong from both a social and an economic perspective. Most developers of Open Source work for companies whose main interest is to use it, rather than to make money off of it. They're happy to share, because they never expected to make money on their enabling, infrastructure software. They just want to make better software, and more of it, with less expense. My paper The Emerging Economic Paradigm of Open Source goes into that.

Biancuzzi: Has the funding process evolved since 1998?

Perens: The problem with making your money from support is that early adopters are self-supporting. So, selling support didn't work for a whole bunch of early Linux companies. It works better today for Red Hat and MySQL, for example. But there are many other ways to make money. Most folks just use Open Source as a tool, they pay for their share of its development with their IT budgets, and they make money in some way that isn't connected with sales of software at all. Dual-licensing is a popular way to make money from users who aren't willing to accept a Free Software license like the GPL for their own work.

Biancuzzi: What differences do you see between late '90s and today in the way corporations interact with Open Source projects/developers?

Perens: Gosh, it used to be hell convincing some corporate folks that they had to listen to the Open Source developers at all. I remember one day explaining to a marketing person from Corel (remember Corel Linux) how she needed to track the developer community and how they feel about Corel, and she looked at me as if I was suggesting something really gross. Her, tarry with programmers? Well, I'm glad that time is over. Companies recognize the value of their communities.

Biancuzzi: Do you see any difference in the way Open Source evolved in U.S.A. and in the rest of the world?

Perens: I think it has been easier to be listened to in Europe, where they have parliamentary governments that represent a broader set of interests. We got the ears of the Green Party in many nations before the others. Now that we've gone so mainstream, people from all parties listen.

Now some Q&A with Eric Raymond. He is the author of The Cathedral and the Bazaar...

Federico Biancuzzi: Is the distinction between "free software" and "open source" still meaningful?

Eric Raymond: That depends on what your goal is. If you're primarily interested in moralizing and conducting ideological warfare, then there is a theoretical distinction that matters a lot. If you're primarily interested in either producing superior software or describing the collaborative behavior of people who produce better software, then the distinction is practically meaningless.

I'd say "completely meaningless," except that there are two rarely-used licenses (the NASA Open Source Agreement and the Reciprocal Public License) that OSI certifies "open source" and FSF says are non-free. I think this difference is due only to minor drafting errors and technical differences in interpretation. Recently RMS and I agreed in principle that OSI and FSF should jointly approach the NASA-OSA license owners to fix that problem.

Biancuzzi: What is your point of view on licenses proliferation?

Raymond: A lot of my recent work with OSI has been around attempts to slow down -- even end, if possible -- the flood of new licenses. I think it was necessary for the business world to experiment and explore the license design space, but I also think that period of exploration is largely over now. In fact, I expect the number of licenses in use to contract sharply during the next few years.

Biancuzzi: What is your opinion of the Creative Commons project and its licenses?

Raymond: A noble effort. I think there's a serious flaw in some of their licenses, though, in that they rely on being able to make a bright-line distinction between "commercial activity" and "non-commercial activity." In fact, at least under U.S. and British law, there's no reliable way to do that; you run into edge cases like fundraising activity by non-profit organizations. That's why the OSD doesn't depend on making any such distinction -- it's too easy to end up with either exploitable loopholes for non-cooperating commercial entities or nasty legal risks for nonprofits.

Biancuzzi: I read different explanations about the link between GPL and LKM for Linux. Do you think that LKMs for Linux should be automatically covered by the GPLv2-only used by the kernel itself?

Raymond: I'm outside that debate. First, I think it's Linus's policy decision to make. Secondly, I don't care to get involved in theological disputes about GPL variants or its scope because I no longer think we need "reciprocal" or "infectious" licenses at all.

The GPL made sense for a community that was small, frightened, and believed that economic incentives would push corporations into taking code without giving back improvements. Well -- today the community is large and it frightens corporations. Most importantly, we've learned that defection is its own punishment, because the day you defect you lose most of your developers and your cost structure goes way up. Eventually, your competitors who didn't defect eat your lunch.

So, why GPL? The answer I think is that it's mostly a historical habit. And a lot of hackers like to see themselves as persecuted rebels, so clinging to GPL is psychologically rewarding and they don't really want to notice that it's no longer necessary.

Biancuzzi: What about GPL3 then?

Raymond: It's another reciprocal license. Better written in some ways than GPL2; I had a long email conversation with Linus in which I tried to persude him to adopt it, my goal being to reduce the amount of political crap flying around in the community. But functionally I don't think we need this class of license any more.

Biancuzzi: Quoting from http://opensource.org/history

Our first President, having studied the history of reform movements, was much concerned that the open-source community needed leading institutions not dependent on the charisma or talents of their founding members. The single most important fact about the history of OSI as an institution may therefore be that as of 2007, the offices of President and Vice-President have rotated three times each and the Board Of Directors is about to inaugurate its fourth slate of new members.

Would you like to elaborate on these studies you did?

Raymond: There is a pattern that one sees over and over again in failed political and religious reform movements. A charismatic founder launches the movement, attracts followers, and enjoys significant successes; then he dies or leaves or attempts to name a successor, and the movement disintegrates rapidly.

One of the classic, much-studied cases is that of John Humphrey Noyes and the Oneida Community, 1848-1881. It was especially clear in that case that its succession crisis and eventual collapse was due to over-reliance on Noyes's personal leadership. At the time I co-founded OSI in 1998 I judged that FSF would very likely undergo a similar crackup if it lost RMS, and was determined to avoid that if possible for OSI.

Biancuzzi: Had you ever considered to use a democratic process (elections) to involve more people and make OSI fans more active?

Raymond: Yes. I initially rejected that idea because I felt OSI needed to be a very lean, low-overhead organization. I believed that a democratic membership structure would give us only a marginal gain (if any) in energy or public legitimacy, at the cost of more chronic internecine political squabbling than was really worth it. I was influenced by the history of SPI (Software In the Public Interest, the parent organization of the Debian project), which has had exactly this problem. In 1998 I considered SPI a grim warning not to go there.

Biancuzzi: Do you still see the same excitement present as the end of the '90s? Or it went away with the bubble? Or maybe it evolved in something different (business, for example)?

Raymond: I don't think I have a meaningful answer to that question. I don't think my excitement about open source, or that of the people I knew, had a lot to do with the dot-com bubble. (Well, except that at least some people think I helped cause the bubble -- but not me, I think Bernie Ebbers lying about a broadband demand explosion that wasn't really happening was what fueled most of the hype.)

Biancuzzi: Bill Gates is leaving Microsoft... I guess you still are their "worst nightmare," right? :)

Raymond: I was Microsoft's worst nightmare (e.g. an articulate, marketing-savvy ambassador from hackers to business) in 1998. Ten years later they have far worse nightmares than me... Google, the open-source movement I helped create, and an increasing number of vendors shipping low-cost PCs with Linux preinstalled.

Biancuzzi: Do you think this event (Bill Gates leaving Microsoft) might affect in some way the open source world?

Raymond: Unlikely. I think peaceful coexistence with open source is impossible for Microsoft as long as they're stuck to their current business model. And I can't see that business model changing -- there aren't any plausible alternatives that preserve their monopoly profit margins.

Biancuzzi: Could a deal between Microsoft and Yahoo change this?

Raymond: I don't see how. It's more likely that if Microsoft acquires Yahoo they'll just impose their anti-open-source norms on the place. That's what happened at Hotmail.

Biancuzzi: Is the patent system broken (and needs fixing) or useless (and needs termination)?

Raymond: I'd say extremely broken, so much so that it might be necessary to terminate it and start from zero. I don't think it's useless, but I think the combination of high process costs and patent terms not adjusted to the short innovation cycle times of today are a deadly, innovation-suppressing combination.

Biancuzzi: If I think at the major commercial OSes (Microsoft Windows, Apple Mac OS X, Sun Solaris, IBM AIX, etc) I see that they are developed by corporations in U.S.A. Maybe they are all US based because of some laws, such as software patent laws, that provide them advantages.

I don't know, but at the same time I see that some popular companies that develop free software (Novell/Suse, Redhat, Mozilla, etc), and some no-profit umbrellas (OSI, FSF, various *BSD Foundations, etc) are US based too... I start wondering if these laws actually damage the software development environment, help it, or are irrelevant.

Raymond: On balance, I think they probably damage it considerably.

Biancuzzi: How do you explain that all these initiatives related to software, yours included, are created mostly in U.S.A.?

Raymond: Culture and capital. The U.S. combines several advantages that are hard to beat. One is that we have huge amounts of capital sloshing around looking for high-risk but high-gain investments, and historically computers and software have been capital-intensive industries. That era is gone now, done in by cheap PCs and the Internet, so some of the U.S.'s competitive advantages have gone with it.

On the other hand, the U.S. has cultural traits that are very good for incouraging "initiatives related to software" and that are harder to duplicate elsewhere. Part of the emotional engine that drives really creative programmers is a secret (or not-so-secret) belief that they're smarter than almost anyone else. A culture that smashes that kind of egotism in the cradle (like, say, Japan's) can produce many programmers that are skilled but few or none that are superbly talented and driven. The U.S., on the other hand, is more supportive of that kind of individualism than anywhere else in the world. Which is why Linus Torvalds lives here now.

Biancuzzi: What are the biggest successes of OSI during these 10 years?

Raymond: We sold open source to corporations with lots of money. We taught the hackers how to get along with profit-seeking businessmen and the businessmen why they shouldn't try to change the way hackers behave or organize themselves. We taught our geeks to love the free market and the free market to love our geeks right back.

As a result, you can now get a good job doing open source. And you can buy a $200 Linux computer at WalMart. And we have powerful allies against enemies like Microsoft, the RIAA, the MPAA, repressive national governments, and anyone else with a desire for locked-down control of computers and the Internet.

OSI did that. And I think we have reason to be proud of it.

It's the turn of Brian Behlendorf of Apache fame...

Federico Biancuzzi: What was your role in 1998 as member of the Board of Directors (of OSI)?

Brian Behlendorf: My only contribution was to participate in the license approval process. Best would be to look at archives from the list around that time frame to remember specifically what I did.

Biancuzzi: If I remember correctly you worked with IBM to help them include Apache. What differences do you see between late '90s and today in the way corporations interact with Open Source projects/developers?

Behlendorf: To be clear, I was external to IBM -- still working on my day job at Organic Online -- the main thing I did was travel for one meeting in Raleigh, and then help them figure out how to be a contributor just like anyone else to Apache. But I also spent time with IBM execs explaining the business models that other contributors used, why they didn't need to fear the license (so spent much time with their lawyers on this), that sort of thing.

Back then, no major companies wanted to publicly admit they used such software, let alone publicly support it. Today you have a clear separation between two groups: companies that are supportive and let their employees contribute publicly, perhaps even driving new open source projects, the way CollabNet does with Subversion - and other companies that are still either too conservative or fear a backlash from proprietary software companies. I think the former group has a competitive advantage when it comes to recruiting and retaining talent.

Biancuzzi: Is the way open source developers collaborate changed over time?

Behlendorf: Some things have not changed; emails on developer mailing lists are still the lifeblood of any well-run project, creating the conversational space that the other tools simply hang off of. The quality of those add-on tools has improved tremendously though - from CVS to Subversion, from gnats to Bugzilla, from Emacs (which wasn't bad) to Eclipse, that sort of thing has made a difference. IRC and instant messaging is used perhaps more than before, but it's usually used to simply educate and maybe do a bit of brainstorming. The real meaty back-and-forth of development still takes place via email, which provides the ability to participate asynchronously, and creates a history of that conversation in a readable form. Ever tried to go back and read an IRC log? :)

Mr. President himself, Michael Tiemann, answers a few questions!

Federico Biancuzzi: What is the role of OSI today?

Michael Tiemann: Concretely the OSI grows and serves the open source community as the authoritative steward of the Open Source Definition, the authoritative body that discusses and approves licenses as open source, and as one of many resources for open source reference material, best practices, policies, and strategy. The board meets monthly to discuss administrative progress (such as license approval) and to plan and discuss things we can do to help nurture and expand open source throughout the world.

Some recent, major concrete milestones on the licensing front were:

Each of these achievements was the result of months of work, hundreds of emails, several face-to-face meetings, and a lot of careful reading, thinking, and writing. And each of these achievements resulting in large measures of both praise (for our evenhandedness) and scorn (for making a decision that went against one or more minority opinions).

We launched a new website (affectionately called opensource.org version 3.0) in April of 2007, and in early January our homepage logged its 2 millionth hit. Our license listing page generates nearly 100,000 page views a month. We have hundreds of blog postings, some of which have generated more than 50,000 reads. All of this speaks to the community's interest in having a strong, independent voice that is neither aligned with a commercial company nor any specific software project.

Biancuzzi: Who is financing OSI?

Tiemann: The OSI accepts money from a variety of sources, including small donors. Our policy is to acknowledge all donations larger than $5 via electronic mail. We do not accept money with specific quid-pro-quo stipulations, but only money that is unrestricted (including no requirements to advertise the source of funds). Last year at OSCON, the OSI raised funds from over 400 individual donors, which made us very proud.

Biancuzzi: What are the goals of OSI for the near future?

Tiemann: The OSI will be holding another Board election in April, and we will continue to travel the world and speak at conferences and community events (schedule and budgets permitting). We have been talking about changing the structure of the OSI from being a self-selected organization to becoming a membership organization, but we have not made enough progress to promise any concrete changes in the near future.

If you believe in Free Software and Open Source, you can do something concrete for OSI: donate :)

Federico Biancuzzi is a freelance interviewer. His interviews appeared on publications such as ONLamp.com, LinuxDevCenter.com, SecurityFocus.com, NewsForge.com, Linux.com, TheRegister.co.uk, ArsTechnica.com, the Polish print magazine BSD Magazine, and the Italian print magazine Linux&C.


Return to ONLamp.

Copyright © 2009 O'Reilly Media, Inc.