Shared Source vs. Open Source: Panel Discussion
Pages: 1, 2, 3, 4, 5, 6
Tim:
[Let's] come back to a point that Mitchell made,
and that's really the health of the overall ecosystem and I think
some of the concerns that people have. I think one of the big
differences between Microsoft and the rest of the world is that you
guys [Microsoft] see yourselves as still a relatively small player in
a much bigger world that you want to be part of. You see a lot more
markets where you don't have that power.
For most of the people in this room, you are the universe. You are, to use Scott McNeely's phrase, "the top predator." And in this ecosystem, there is a concern that there really is not enough to go around.
You mentioned the size of the software industry earlier, and you mentioned the number of companies. I remember an SPA study a few years back — you know, the average software company with ten or fifteen employees, which if you do the math is most of those that you talk about — does not make money, because a few big players are making all of the money in that industry.
And we've basically created a situation where it is very, very difficult for new entrants to compete in the traditional software industry. [applause] And I think a lot of the attraction to open source is people saying, "We want in. We want an opportunity as developers to have a chance to have our ideas flourish." And you guys have been so successful that you've created a market where it's difficult for developers outside Microsoft to have a chance to succeed. So I'm kind of wondering if one of the issues behind this whole proprietary source versus open source debate is really, "How do we get to a health in the ecosystem?" Is it "What's good for Microsoft is good for the software industry?" at this point? The two may diverge.
|
Panelists: Clay Shirky Michael Tiemann, David Stutz Mitchell Baker Ronald
Johnston
Craig Mundie, Brian Behlendorf Tim
O'Reilly |
Dave:
Tim, don't you think that the reason that —
one of the reasons that it's hard to enter the industry now is that
the expectations of the consumer have been brought up by the quality
of software that's out there? I mean we've written — all of us
have written very good, high-quality software, and it's continued to
raise the bar over time.
Tim:
My experience is that it's actually pretty easy
to enter in new markets that are outside the scope of what people
have thought about at Microsoft. It's not very easy to get
very far in those markets, because then Microsoft comes in and uses
its current market power to basically take over those markets. So you
see cases again and again where somebody comes up with a good idea
and they don't get very far.
Dave:
Well, perhaps I've got a unique viewpoint, but
I've certainly been inside Microsoft and seen repeated failures. It
is not the case that we automatically are granted the
franchise.
Tim:
Oh, I understand that.
Mitchell:
But one of the things that Microsoft has is
plenty of ability and money and backing with which to fail.
Craig:
Sure.
Mitchell:
And also — I'll finish in a minute —
the other thing that Microsoft has, which is pertinent in this open
source debate about making money — is a very efficient system
for taking money out of certain aspects of the software market. It
can do that by giving products away for free — like the browser,
so there's no money in the browser any more. Or by rolling it into
the OS or into Word or whatever. And so in that system, where
Microsoft has, rightly or wrongly, as a business matter, the ability
to extract all of the money out of given areas of software
development, I'd say it's open source — where people join
together and voluntarily produce something more than any one of them
can do [alone], and jointly create a work without direct payment
— that is the only way to get innovation in these areas.
[applause]
Tim:
Dave?
Dave:
So I think one of the points that Craig was
making a minute ago, that I'll come back to... We recognize that we
are in a position in which we have a lot of resources and that people
are sensitive about this, and we have started saying, trying to be
very clear about what our intents are, and trying to work with
people in the community to carve out places that are safe places,
such as the ECMA standard, for example. And so we need, as a
community, as a whole, to continue to develop ways that enable
businesses based on free markets, which is where we are, to continue
to exist — but also for ways to foster trust and innovation.
Tim:
Well, let me ask this question: Would it be a
fair analogy to say that Microsoft is shifting from a hunter-gatherer
economy to a farming economy? And that part of the goal of shared
source is to say, "Hey, we've got to breed innovations, we can't
just sort of expect them to sort of happen out there?"
Dave:
Well, you know, I am a farmer, actually — as
Tim knows, I have a farm. So I'd like to think yes. But Craig, I
don't know, what would you say to that?
|
Shared Source vs. Open Source Related Links |
Craig:
I actually think a little differently.
Microsoft would be nothing if it wasn't for the fact that millions of people wrote apps and deployed them on PCs. I mean, just offering an operating system is not an interesting business per se. And so— [audience and panel laughter] —It's true. I mean, it's a fact. If you don't have a symbiotic relationship between the platform and the users of the platform — both the ultimate consumer and a development community for it — the thing will not sustain itself. So in a way, maybe we already were a farming economy, in that we depend on having a lot of people both develop for and use this thing. And in a way, as this world has changed, as people have diversified both the number and type of platforms that are interesting, and you get this incredible device diversity, I think we've been forced to recognize there's no direct transference. And again, I reiterate I can personally speak to our inability to transfer any of our PC franchise to any of these other ecosystems.
Tim:
Is that part of why the GPL bothers you so much?
Because it's getting so much traction, and Linux is getting so much
traction on embedded devices?
Craig:
No. No. In fact, our concern about the GPL is
strictly the fact that it creates its own closed community.
Tim:
But so does Microsoft. [applause]
Dave:
No, no, it's not true.
Tim:
Yes. [applause]
Craig:
I mean, that's fine. If the GPL community wants
to explain how people get to stand on their shoulders in both the
commercial world and the academic world, fine.Then that's what this
debate should be about, to elucidate that.
Brian:
[Let's] compare the GPL to the shared source
license. The GPL tells me as an entrepreneur, as a businessman, the
terms under which I can use that software. Right? It sets boundaries,
it says what I can do, what I can't do, how I can incorporate, how I
can't, that kind of thing. The shared source license, at least the
one for the Windows CE release that I saw, says "This is the
noncommercial use license. Contact us for terms under for commercial
use."
Craig:
That's right.
Brian:
As a businessman, it doesn't tell me how I can use it.
Craig:
Call us up.
Brian:
Well, let me ask the audience here. How many
people here write software completely, only, for hobbyist purposes?
Not for their employer. Not for some commercial purpose. [pause] A
few, I'd expect. But certainly the vast majority here write software
[that has] some other commercial purpose. So in the spirit of helping
Microsoft come up with a good shared source license, perhaps —
maybe I shouldn't do this, but — you know, if it even said,
"A dollar per CPU that it runs on," or something like
that.
Craig:
Anybody who wants to know, call us up, allright?
And we'll figure out, well, what is it you want to do? We'll give you
a price. Just call us up!
Brian:
Granted. That sets up a different set of
relationships, though. The shared source license itself does not
define a community of people who are part of that community the same
way that open source licenses do.
Dave:
Actually that's not true. Specifically, in the CE
community, not only is there the noncommercial license, which is the
one I was waving around, but we also have a community-based
source-sharing effort with people that build devices commercially. It
is in fact set up around the notion of everybody sharing changes
—
Tim:
—I saw Ron trying to get in there for a
moment and he hasn't had a chance to say anything.
Ron:
I'll give the suits' — as the lawyers are
known — perspective on a couple of these questions. The legal
business is just only recently getting into debating the
interpretation of these licenses and the enforceability of these
licenses. But we've been advising companies on drafting these
licenses for a long time, and most of the commercial software
companies draft licenses tailored to their distribution, tailored to
their parsing of intellectual property rights, tailored to their best
commercial interests. Now that is not just a Microsoft problem. That
is the problem for virtually every large software company that at
least I'm aware of, save a couple. The trouble with the GPL, to
companies who are doing commercial development, is in fact it doesn't
at all cause clean lines.
The language in the GPL that tells you what you can do without being subject to the redistribution provisions in the GPL is expressed in about four different ways in the GPL. None is any more specific than the basic derivative work standard of the copyright law. Some of the others' expressions of that standard are inconsistent with the derivative work formulation under copyright law. Now we've been litigating that standard in the software business for over twenty years, and there's been very, very little guidance after twenty years of litigation.
I've tried cases on the meaning of that standard. There's been very little guidance from the courts as to what constitutes a derivative work. It's clear that it goes beyond code. It's clear that it goes into the levels of abstraction above code. But beyond that, there's been no articulation of a useful test. The test that's most often used by courts is in fact dangerous because it makes the court feel, when it gets through the test, like it's accomplished something — when in fact, as any programmer would know — it's just starting the analysis.
So the trouble with integrating any kind of commercial software with the GPL is going to be in part that there's a huge uncertainty as to whether you can do that — and where you're going to end up five years from now.
Brian:
If the uncertainty was clarified — I mean,
Stallman's working on Version 3 [of the GPL] — if Microsoft
wanted to participate in that, I'm sure he'd welcome their
feedback.
Craig:
We already are.
Ron:
One trouble is that —
Craig:
—One sec. We put up all these questions and
say: "Here. Here's twenty questions. Ask yourself these
questions." You can't answer these questions by looking at this
license. I'm mean that's pretty clear guidance where the problem
is.
Michael:
What I just heard is that the derivative work
clause in the Microsoft shared source agreement is something to be
feared because in fact it has stipulations about what is covered by
intellectual property rights concerning derivative works of
authorship. And if derivative works of authorship extend beyond code
into abstractions — if I look at Microsoft software, I'm
infected. [applause]
Ron:
The language that has classically been chosen by
companies depends on the kind of software. For example, the —
Michael:
— The language is derivative work of
authorship. It's the same language in both licenses.
Ron:
In some context, that makes sense as a standard.
In other contexts, where you're trying to establish a generalized
standard for all software, it doesn't. Now Microsoft obviously has,
like most other software companies, a whole series of license
agreements that include different standards and different language
and grants of rights depending on the nature of the software
—
