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


Shared Source vs. Open Source: Craig Mundie and Michael Tiemann

08/09/2001

Tim addressing the OSCON attendees
Tim O'Reilly addressing the OSCON attendees.

 

Shared Source vs. Open Source Related Links

Technetcast of the debateWith audio

Part 2: Panel Discussion

OSCON Conference Coverage

Comment on this articleCraig Mundie presented a compelling argument for Microsoft's business model and their position for sharing source code. Michael Tiemann says that Microsoft isn't accurately portraying their intentions. Where do you stand?

Post your comments

Editor's Note: On July 26, 2001, Microsoft Senior VP, Craig Mundie addressed the open source community during the morning keynote at O'Reilly's Open Source Convention in San Diego, CA. Then Red Hat CTO, Michael Tiemann presented his counter argument to many of the points made by Mundie. Here is the transcript of both presentations.

Tim O'Reilly:
As my friend Doc Searls wrote in The Cluetrain Manifesto, "markets are conversations." And one of the most important conversations going on in today's software market is the conversation between proprietary software — whose most successful example is Microsoft — and the open source movement (I should say the open source and free software movements just to be clear... [applause]), whose work is really transforming the software market and changing the rules that Microsoft has worked under, and been so successful under, for so long.

I think Microsoft has displayed an awful lot of courage and vision over the last couple of years. I know people — in this community particularly — like to beat on Microsoft, but I really think that they have been extremely bold as they've really embraced very boldly this change from a desktop-centric architecture to an Internet-centric one. And they've been thinking long and hard about what that means.

And now they're doing the same thing with the lessons that they're seeing out there in the world coming from free software and open source. They're looking at what's necessary — or (asking): Is there something out there in moving from proprietary software development models to ones that are more open and collaborative?

Right now they've released some preliminary thoughts on shared source licenses. They might not like to think this way, but I tend to think of the current shared source licenses as Windows 1.0. They got that there was a revolution happening with the GUI, and they jumped on it and everybody laughed at their first attempt.

People may complain about things in shared source, but my belief is that Microsoft is really good at learning from the competition. And they're learning from us. They're saying, "Hey, you guys have something. We're going to figure out what's good about it, we're going to figure out how to keep what we also believe in." They're going to keep working at it until they get it right. So maybe they got it right this time; but even if they didn't, I think they're going to keep at until they do. A lot of [what] I'm really wanting to do is to help them get it right. And I want you to hopefully help them get it right. And help them learn what it means to be a participant in our communities.

Anyway, I'm really pleased that Microsoft Senior VP Craig Mundie has agreed to come down today and engage in this conversation about software development methodologies and business models. It's really clear that Microsoft is willing to learn from open source and free software, and I'm hoping that we're able to listen with an open mind to what Craig has to say and to learn from him as well.

I talked about markets as conversations, and the format of this keynote is also a conversation. We're going to hear briefly from Craig. Then we're going to hear briefly from Michael Tiemann of Red Hat. And then we're going to invite up several additional panelists to join us in what I hope will be a very,very thought-provoking conversation. And I hope you'll enjoy it also. As we get towards the end, we'll invite the audience in to participate as well. And when we do that, I really hope that you will be sort of as respectful and thoughtful as we're trying to be ... [audience laughter] ... in this conversation. All right?

With that, I'd like to welcome up Craig Mundie of Microsoft.

Craig Mundie:

Tim addressing the OSCON attendees
Craig Mundie giving a historic keynote address at OSCON 2001.

Thanks Tim, and good morning everyone. It is indeed a pleasure to be here. I almost don't have to give my speech because I think Tim summarized some of the key points for me, but I'll go ahead and try anyway.

Let me just begin by explaining a little bit what my role is at Microsoft, so you can have some perspective on the comments that I offer and the role that I play. I've been at Microsoft about nine years, and originally went there to start all the company's activities in non-PC computing.

So I mean, as Tim said, Microsoft is a company that tries to learn and tries to move ahead. And as far back as 1992, we had decided that the world in fact was going to change substantially — that we wouldn't just find computing in the things that we call computers. And my job was to try to move the company in that direction. And so, many of the things that you see today in terms of Windows CE and Pocket PCs, auto- and television-based products, are things that I did in the first six-plus years that I was there.

Three years ago, those things came to a point of maturity where Microsoft realized they needed to be blended together, and we needed to address the marketplace by thinking in a more holistic way about computing and how we offer those products, not only to enterprisers but to consumers. And we merged all those together. At that time Bill also — Bill Gates — had decided to focus his entire energy, going forward, on questions of strategy and architecture for the company, because we recognized that the market ahead of us was probably more turbulent than any market the company had known in its twenty-odd years of history.

It was at that time that Bill asked me to go work for him, and I've done that for the last three years, working with him almost exclusively on prospective issues of architecture and policy and strategy for the company. And in that role I try to support Bill and other executives, to take on policy issues and to be a speaker for the company about things that we're concerned with. And usually that falls in the realm of things that happened in the capitals of the various major countries of the world, but in this case there are a set of issues here that are quite profound and [we] decided that we should speak out about them and try to at least provide leadership.

In the discussion that has ensued since I gave the first comments about this at New York University Business School the third of May, our goal all along has been one of creating an environment around informed choice. So my role in coming here today and in speaking out about these issues for the last several months has not been about one of trying to legislate anything to anybody, but rather to create the dialogue and make people understand what we think some of the long-term implications of their choices are.

If you look at the reportage that we have had over a lot of my comments — and as I thought about coming to this conference — go to the next slide — we wondered a lot about how I hould actually introduce myself.

[The slide superimposes the faces of Craig Mundie and Dave Stutz over those of Dr. Evil and his Mini-Me.]

[Audience laughter] David and I came to this conference — Dave, of course, can be here for the entire conference, I can only stay for part of the day — we wondered, What is it that people think we're really trying to do? Why are we taking on this battle? It clearly, certainly has been reported at times that Microsoft doesn't like open source. So let me be clear. One of the reasons I came to this conference is to speak to the open source community directly.

Microsoft has no beef with open source. In fact, we think it's an integral part of an ecosystem that has fueled such tremendous success and growth around the world in the software and information technology businesses. But there are aspects of this general movement — even as Tim himself pointed out — there seem to be really two large camps in this environment: the free software movement and the open source movement per se; but largely the press has certainly been confused, because my comments have been largely directed at some of the issues that we find around the licensing regimes, around the free software movement. And yet it's most frequently reported as "Mundie attacks open source."

So let me be again quite clear. Open source isn't the issue. And as I go through the rest of these comments, hopefully you'll understand specifically the concerns that we have; why we offer these for your consideration and those of everyone else in the world, and why we are happy to have an opportunity to maintain this dialogue directly. Next slide please.

Shared Source vs. Open Source Related Links

Technetcast of the debateWith audio

Part 2: Panel Discussion

OSCON Conference Coverage

At the end of the day, the biggest concern that Microsoft has is about the long-term preservation of what we think of as the software ecosystem. This ecosystem, like those in the physical world, evolved over quite a long period of time, and they reach usually some state of equilibrium where people are happy with the results for a long time. But as we've learned in the physical world, certain actions that people can take on or certain things that they do can have some long-term effects on that ecosystem. And we deal with those in our daily lives as we think about endangered species, as we think about having enough electricity, and a lot of other things. And Microsoft, while certainly a company that has benefited from the commercial software environment, recognizes that we only play a small role in this large software ecosystem. And that role may be significant in some dimensions, but there are many ways in which you can measure it and in fact find that it is in fact a relatively small part, certainly by total dollar volume or any of a variety of other measures.

And so, if we think about this ecosystem — or at least, the way we think about the ecosystem — there are four different components to it that we think have a strong, symbiotic relationship, and which has led to the kind of productivity that we have in the software industry today.

[The slide shows Intellectual commons -> Industry -> Customers -> Government in a circle with arrows pointing from one to the next.]

Clearly the industrial companies are a big part of it. If you can go back in time to the beginning of computing and also to the beginning of software as we know it today, it was really a contribution that came through the intellectual commons, broadly defined, and in particular university efforts. And in fact a lot of the things that are the most popular today probably wouldn't have had their beginning without specific government funding. Where, for example, the original Arpanet, funded through DARPA, was a key part of creating the technology that we all enjoy the benefits of today in the basic forming of the Internet. And of course the customers themselves are part of the ecosystem, not only in terms of consuming the efforts of the companies and other institutions in this environment, but more and more customers are finding that they have to do some of their own software development. And the reason is that more and more, the base business process is becoming captured in the software that defines the business. And it's our belief that that will in fact increase over time, not decrease over time.

And so it isn't really surprising that some aspects of the open source movement speak directly to some parts of this ecosystem. The whole idea of having an intellectual commons, outside the academic environment, seems to be a key underpinning of what people like as one aspect of the open source movement. The ability for customers to believe that they can have access to source codes at times certainly has merit, and it's one of the areas where Microsoft has learned from this community and recognizes that there are things that we too can do to improve in this area.

As Tim said, I think Microsoft is a learning machine. It's got quite a few smart people. We really try to listen to customers in the marketplace and to things going on around us. And many people I don't think stop and think, but the Microsoft we know today is a company that has in fact morphed probably five or six times in its 25-year history, in order to adapt to the changing conditions around it. And certainly we find ourselves in a time of flux as great as any that we've known in the history of the company. And so we find it incumbent on us to pay attention to what's going on that engages all the members of this ecosystem and try to figure out what is our way forward. Next slide, please.

[The slide is an outline:

At the end of the day, it is all about choice. When I started giving these, people said, "Well, hey, it's consumers that want choice." The reality is that everybody wants choice. As an individual developer, you want choice. As a customer, you want some choices. As governments, ultimately, you want choices. So I want to talk a little about some of the choices I think we all face together in the course of the next few years.

When we talk about development choice, there are a number of different things that immediately come to mind.

The question of language. And language itself in the computing environment has been one of continuous evolution, and we expect that to continue.

Community. Clearly a very important aspect of the open source environment — I mean, this group calls itself a community. I think community is not [a] unique notion. One of the things I tried to highlight in my comments in New York was that in fact Microsoft has always had a developer community itself. In fact, it numbers in the millions. But we in fact haven't related — for a variety of reasons — to that community the way that you relate to each other. And by observing that and how that works, we find a lot of merit in the aspect of what you do, and are trying to emulate that in some things.

The source model. In the sense of, "How do you develop? What are the tools you use for development?" The platform that you target these things to. All are choices that developers make, whether they're individuals or corporate members.

Then you have the choice of distribution. There's been lots of experimentation in the last three decades as people have done more and more work in software. There's the open source concept, which is the hallmark of this conference. There were the shareware notions and the freeware notions, which to some extent I think were a progenitor of some aspects of open source. And then there were what you might think of as the commercial or mixed models, where companies like Microsoft choose to provide some technology on a commercial basis and some we actually contribute through either standards efforts or other mechanisms — for example, academic programs — largely to try to continue to bootstrap the ecosystem and help it move forward. Because we recognize that we benefit from that ecosystem, and we want to continue to have the opportunity to benefit from it.

People have licensing choices. There's the Berkeley licensing. I mentioned that in 1982 I started a supercomputer company, and we were designing machines that had no previous architecture that was really quite similar, and so we had to build our own tools, our compilers, and in fact our own operating system. You know, I took in my company a Berkeley license and we built an operating system — a UNIX operating system — based on that. So I have personal experience about the benefits from being able to stand on the shoulders of the great work that was done by people like Bill Joy and others in their academic careers.

There's the shared source license, which is essentially the umbrella name that Microsoft has chosen for a family of software products that we offer to our customers. We've offered a small number of these than in the past, but we do find legitimate reasons to offer a wide array of customers a quite varied array of shared source licenses. And in fact, I totaled them up last night. We probably had two or three of these in effect when I gave this speech in May and said we would do more. And today we have eight. And so we're quite aggressively pursuing providing source access to people, but in a very specific licensing regimen.

There are the traditional commercial licenses, which cover most of the products that people buy today. There's the GPL, and there's the public domain.

Then you get down to business choice. What is the business that you're going to be about? You can be in the business of offering services. And in fact Microsoft has recognized [this], courtesy of the Internet revolution, but more and more there almost has to be a service component to every software product in the future. That wasn't true in the past, but we think it will be more true in the future. And we clearly have begun to adjust our business in order to support that.

There's the packaged software model. There are different forms of distribution. There's embedded software and appliances, and specific products that are shipped with hardware. There are many different business models that people confuse.

[Slide:

We happen to like, and will continue to pursue, a commercial software model.

Each of these choices ends up having some long-term policy implications — either in terms of its effect on the ecosystem, the sustainability of your investment, or ultimately, the form of the business that you take on.

So software as a business is a model that Microsoft believes in. It's one we think is important in this ecosystem, in terms of having adequate research and development spending over a long period of time. And we will continue to have this be our choice. I guess ultimately the market will tell us whether or not that choice is a good one, or whether it's to be supplanted by some other predominating model. But right now, we think it will be important option for people, and we will continue to pursue it.

The software industry is a critical industry in this country, and increasingly a critical industry elsewhere in the world. If you just look at the U.S. economy, it's an integral part today. There are 148,000 commercial companies in the software industry today, and they provide two million jobs. These are companies who actually pay people to develop software, and then sell that software. And we think that that is going to continue. It won't go away overnight.

The net result is this feeds the ecosystem, because through that employment and through the corporate profits, this industry in the United States last year paid 28 billion dollars in taxes — which, in our society, is the way that we redistribute the wealth that's created in these companies — in order to feed the ecosystem, through government programs, through academic programs, and things that generally support our society, of which we all benefit. More to the point, this industry exported 121 billion dollars worth of products out of this country, and that clearly has been an important part of the economy. The net effect of that is that it's taught every country that in the future they're going to have to have a software industry.

And so this questions of the ecosystem. This questions the licensing regime and its impact. And ultimately, [it's] even an important policy matter for heads of state, because if they want to have a software industry, we think that they have to in fact foster an ecosystem and a culture within their society that says that people benefit from the fruits of their labor. And where those things are intellectual properties — that too should be respected and paid for. Next slide.

[Slide:

So Microsoft certainly is attempting to learn from the open source movement at large. We've expanded, in quite dramatic ways, our community program. We've recognized that people do in fact seem to relish the opportunity to interact with each other when they have a shared interest, and in fact interact directly with our developers. And while that hasn't been our practice in the past, since May 3rd we've actually begun to foster the involvement of our developers in a variety of the access and community programs, and we'll continue to expand that.

We've expanded our source access, as I mentioned earlier, to have a broad range of licenses.

One thing I think is hard for some people to understand is in fact the scope or scale of the business that Microsoft is in today, on a worldwide basis. We do 65 percent of our business outside of the United States. Every time we ship a product, it's pretty much demanded that we deliver it instantly in about seven to nine languages and within ninety days in as many as thirty-five or forty languages. And it takes a huge effort to be able to do that on a worldwide basis. We have a huge range of customer requirements that range from military requirements of the United States and others to, just traditional, individuals and consumers at home.

And so we recognize that we need to have and will continue to expand the range of licenses, both for the product as product but increasingly for the source access or tools that depend on source access for these products in the future. But at the end of the day, we think that we can offer some of these things and still do it under a commercial model.

So those are my remarks I'd like to offer today to the community. Make sure that you understand what it is that is really driving us to have this concern about your choice, in particular around a licensing regime and what the long-term implications of that are for this community as a part of the overall worldwide software community.

Thank you for your attention.

Tim O'Reilly:

From the nunber of red hats out in the audience I think you probably know who our next speaker is. Michael Tiemann was the founder of Cygnus Support, and then it became Cygnus Solutions, and then became part of Red Hat. But Michael was one of the very first open source — actually, he was a free software entrepreneur, I should say. He started with GPL Software and he's basically been on that path ever since. And I think Michael is probably the best exemplar there is of somebody who sees that even with the GPL, there is an enormous potential for a software business. And so he's here to tell us about why that's so.

Michael Tiemann:

Thank you very much.

Tim addressing the OSCON attendees
Michael Tiemann speaking his mind on the idea of shared-source software and Microsoft.

Esse quam videri is the motto of the state of North Carolina, and for those of you where were hacking instead of studying Latin in school, it means "to be rather than to seem." I claim that to build an architecture of trust, it is better to be open than to seem open, better to be trustworthy than to seem trustworthy. [applause]

Such an architecture is vital to creating and enabling and governing the technologies which are going to enable our future, and that a lesser architecture will ultimately crash and burn, stunting economic opportunity for all.

This debate is important because it's about the future of software and the increasing substance of technology, its importance in our economy, and if Lawrence Lessig is correct, the way in which code functions as law and how that ultimately governs us.

As long as Microsoft insists on writing code, we as technologists — and we as a business seeking fair and equitable competition — and we as citizens — want Microsoft to do what is right.

Now, of all the choices, open source makes it easier to be rather than to seem. For example, from this perspective, there is no reality to Microsoft's shared source license in the sense that, as [with] many well-spun phrases, it's not unlike the alternative minimum tax: it is neither alternative nor minimum. [laughter] There's no reality beyond it being a proprietary software license — albeit one that seems to offer something new.

Now when preparing for this debate, one question kept coming up, which is why Microsoft would try this new and very high-profile deception when the sum total of its prior deceptions were earning a billion dollars a month. What were they really trying to fix? The answer goes back to October 31, 1998, and the Halloween documents.

Now there are a lot of smart people in the open source community, including — because we include the free software community — at least one genius. [laughter, applause] But I must concede to my worthy opponent that there are a lot of smart people at Microsoft. I do not know who is the first one to see, as I did, that open source software, including software covered by the GPL, could be the basis of a business model powerful enough to legitimately compete with Microsoft. But she was probably one of the smartest.

Then a second smart person, and then a third, clued in to the Halloween documents. And those documents, illuminated by Eric Raymond, showed that a fair number of people within Microsoft began to get it. That open course was a better model, delivering advantages and benefits that Microsoft could not achieve with proprietary software alone. Period. So: hats off to the smart people at Microsoft.

There are many open source software projects and many licenses that govern them. If we look at them as a body, the GPL is the spine. In the licensing debate, many focus on the free versus the proprietary, but they often miss the second dimension of strong versus weak. Microsoft writes strong proprietary licenses. The GPL is a strong free license, much like the first amendment is a strong law protecting free speech in the U.S. [applause]

Microsoft has benefited, albeit illegally, from the application of strong licenses governing its own software. Red Hat has benefited, as has its customers, from strong protections of the GPL, which ensures that our investments, our participation, cannot be used in a way to exclude us from competing in the market. If the GPL did not provide the strong protections of freedom and the guarantees of freedom, we could not have made the investments that we made, for fear that somebody else with more money and more market power might grow to embrace and extend and extinguish us. Instead, we are healthy. We hit our earnings number each of the eight quarters we've been public, and we've announced a profit one year ahead of schedule. Who says that the GPL is bad for business? [applause]

Now back to 1998 and the Halloween documents. 1998 seems like such a long time ago, and three years is a long time in the open source world. While we were participating in a revolution that resulted in unprecedented adoption of open source and free software, I think there was another revolution going on inside Microsoft. A revolution fueled by the technical superiority of the open source model and by the recognition of large, economic considerations by smart people inside of Microsoft. In fact, in his comments, it sounds like Craig Mundie is one of those people, and I very much appreciate that.

When Microsoft bought Hotmail, they not only became one of the largest free email service companies, but they also ran one of the larger FreeBSD server deployments. [laughter, applause] When they tried to switch to Windows, the inevitable occurred. [laughter] FreeBSD works, Windows crashes. FreeBSD works, Windows crashes. Somebody else clued in. The light bulb goes on. Open source software is better. Do you think for a minute that the people who administer those systems are saying, "Gosh, I wish I could dump this FreeBSD crud and run Windows?" I don't. [laughter]

But the revolution was not confined to what Microsoft consumed. They began to produce. Microsoft eventually started to write code and specifications that were actually useful, XML being one of the shining examples of that. And we thank you for creating an architecture that made it simpler and easier to exchange data among programs in a consistent fashion. So smart people inside of Microsoft are getting it, which is great.

Now, as an entrepreneur and an executive, I know what it's like to run a company. And this is both the good and the bad: Before Cygnus was acquired by Red Hat, we had our own revolution to contend with. Outside investors who did not understand our model — and that's further proof that some people have more money than brains — tried to inject proprietary software into our company. We tried to rationalize it. We set up internal barriers. We partitioned the company. We crafted corporate emails with executive-speak to make everything seem consistent. But esse quam videri. It is better to be than to seem. Having read such documents, I can recognize their smell when somebody puts them under my nose.

This shared source thing may be "It's only 1.0." — [but it] has nothing to do with community outside of Microsoft. It is not so much a license, I think, than a treaty, crafted by executives trying to buy time while they quiet the internal rebellion that is Microsoft's own civil war.

History has taught us the dangers of getting too involved in another civil war, and it is a very delicate issue — and I think one that we should debate. What is the appropriate way to make way for the coming revolution of open source within Microsoft?

Shared Source vs. Open Source Related Links

Technetcast of the debateWith audio

Part 2: Panel Discussion

OSCON Conference Coverage

Craig, I think you were brave to come here [applause], and I think you can report back that when Microsoft is ready to sign off on the GPL, and encourage its use to help us build a better, more transparent, more trustworthy architecture for computing — one that empowers individuals, promotes fair and equal competition, and enables freedom at higher levels — we will welcome you to this party as a first-class citizen. [applause]

And you can bet that on that day there will be plenty of both free beer and free software. When that happens, we can be glad that we replaced the winner-takes-all mentality with an everybody-wins model, and that's really what developers and customers want.

So, thank you very much and let the real debate begin.

Copyright © 2009 O'Reilly Media, Inc.