Nov 2

Nat Torkington

Nat Torkington

OSCON: Jonathan Schwartz Interview

At OSCON in August, I interviewed Jonathan Schwartz. Here's a transcription of that interview, the audio of which is available on IT Conversations. Jonathan commented on it in his blog. Don't forget that OSCON 2006 will be July 24-28 2006.

Some highlights from the interview:

  • Anyone who tells you that open source isn't safe is flat-out lying to you
  • I don't think there's one license or one hammer for all nails in the world
  • Every product at Sun will at some point be free or open sourced
  • I think patents have been misused to stifle innovation rather than promote and protect it
  • "So if the Linux kernel were to implement DTrace, Sun wouldn't employ the patents against them?" "Knock yourself out."
  • "It's no longer about competing against a social movement as the big evil company that has proprietary products that they're trying to withold from you
  • I think we're moving away from simply free software to free services
  • Not to mention the running cellphone gag and the part where Nat tells him that OpenOffice is slow and buggy.

Nat: I'd like to bring up our last guest for the morning keynotes. He's a very interesting and very controversial figure. He hasn't yet been fired for being a blogger, but several times I guess it must have come close. I'd like to welcome one of the most prolific commentors on open source and the man who heads some of the projects that have been bitterly fought over and fought for in the open source community. I'd like to welcome Jonathan Schwartz.

Nat: This is your first time at OSCON, so I have a list of tricky questions for you. But it's polite to start with a nice soft question, so I'd like to start by finding out about Andy Bechtolshein.

Jonathan: -sheim.

Nat: -sheim. Sorry.

Jonathan: You should probably talk to Andy about Andy.

Nat: You sure?

Jonathan: I'm not sure I can give you any personal insights.

Nat: I'm not looking for psychoanalysis.

Jonathan: He's very tall.

Nat: Well, that was the easy question over with! So, open source Java. The feeling from Sun for a long time was that you didn't want Open Source Java, Java didn't need to be open sourced, and the community didn't need or want it. Now Apache has started work on an open source Java stack, so I'd like to know what your reaction is.

Jonathan: Choice is a good thing, and I think that the code to Java being available is different to the license to Java allowing someone to fork it. I think that unlike several technologies out there [GSM phone noise]--is that me or you? Excuse me. [Puts phone on speaker] Don't let me forget that.

Nat: So, one of the prizes for filling in your evaluation forms ...

Jonathan: [laughter] In general I'm thrilled to see open source implementations of all technology in the world. We will continue to drive and be at the forefront of making open source safe for enterprises and continue to drive the evolution of technologies as diverse as Apache, OpenOffice, Firefox, and Mozilla. We're contributors obviously to all of that.

So I think the concerns around whether Java should ship under a license deemed open by OSI is separate and apart from "is the code to Java available?" Obviously the code to Java is available, just go take a peek at it. We have concerns on the desktop obviously about forking. Why? Just look back a few years and see what one of the biggest threats to the viability of Java was, it was Microsoft forking it.

Just as we've seen most open source projects tip towards a dominant provider. I think we have seen that, though it's not something we talk about here. There's a reason that when you go into the enterprise, or go into Yahoo!, you see a lot of Redhat and not a lot of anything else, it's because the world can't afford lots and lots and lots of different things to go do the same task.

So in general I'm thrilled to see more open source innovation. I'm thrilled to see open source implementations of Java. I think the jury's not back on how we will continue the future evolution of Java. We're obviously paying attention, we're aware of the issues, and maybe there's a new world emerging and maybe there's opportunity for change.

Nat: Will Sun be participating in the project? I believe IBM announced support in one shape or form, be in moral or bodies on the ground.

Jonathan: There are over 900 members of the Java Community Process, so to ask whether we're contributing to the evolution of Java ... yeah, I think we're going to be pretty invested in the ongoing evolution of Java.

Nat: I meant the Apache open source Java project.

Jonathan: I'm sure folks at Sun will be contributing to Apache as they do today and to other projects at Apache, whether it's Derby or Harmony or any of the variants and offshoots. We don't spend a lot of our time trying to figure out what to stop Sun employees from getting involved in, just look at our blog infrastructure.

Nat: I'll leave open source Java. The other big project I've seen come out of Sun lately is OpenSolaris. How is that being received?

Jonathan: I think it's been received fabulously. We've had 2M license downloads over the past six months or so, and a huge community of folks building around it. I think Andrew [Morton, a keynoter from earlier in the morning] was a little bit nervous because I'm about to meet with his boss about starting a Solaris tree, and I can certainly understand that there's a concern there, but we see the more participants and the more innovation happening, it's a good thing. Competition's a good thing.

So I think where we look at the future of Solaris and frankly where I we see the future of LInux, neither are going away any time soon, both are wonderful and safe. Anyone who tells you that open source isn't safe is flat-out lying to you. Just go ask General Motors that just signed a worldwide site license for an open source operating system (happens to be called Solaris 10) and I think there's more to come. More innovation and opportunity everywhere.

Nat: One of the big sources of conflict around OpenSolaris, or at least sources of comment around OpenSolaris, was the licensing terms which essentially prevented co-mingling of Linux code--

Jonathan: No, that was the GPL that prevented comingling. The CDDL is a license that permits anyone to intermingle.

Nat: The CDDL is the license you chose, so in some sense Sun chose to prevent comingling.

Jonathan: Let me give you a couple of customer stories. I've been with several of our OEMs, Original Equipment Manufacturers, who embed Solaris into their products and ship them. They made statements to us like "if you use the GPL, we can't use Solaris" and so that was one set of choices we had to go grapple with.

Secondly, we didn't own all of the technology within Solaris to be able to dictate to anyone who mingled with Solaris that they change their license terms. I think the reality is that, unlike Eric Raymond who believes the GPL should go away, I'm a big fan of the GPL. OpenOffice obviously ships under it, and a huge spectrum of work we do at Sun obviously ships under it, whether it's GNOME or any of the extensions onto it. But I don't think there's one license or one hammer for all nails in the world. Choice is a good thing, and just like if you turned on the TV set and there was only one channel, that's not a world I'd be interested in living in. I think there's going to be diversity and competition's a good thing. We should allow a thousand flowers to bloom and not just say "that's the flower that we have chosen from on high, that's the one that's gonna bloom" because the odds are it's going to be kinda an ugly flower.

Nat: It was actually Chairman Mao who let Chairman Mao say "let a thousand flowers bloom".

Jonathan: Which is curious, isn't it?

Nat: Yes, one of the more misquoted pieces of his work.

Jonathan: I think the selection of the CDDL and the selection of a license that gave developers choice is something that's obviously pretty important to us. We want people to have choice and for those that aren't familiar with the CDDL license, it's a file-based license, it's a derivative of the Mozilla license. The Mozilla license, last I checked, was the most popular open source license in the market place, just given the volume of software that's been delivered out from under it. I think that's different from the open source operating system world, but nonetheless that freedom that that conveys to a developer, that BSD-like freedom that says "hey, it's your business, you figure out how you want to run it, if you would like to contribute back to the community, we'd like to invite you to do so" but there's an element of self-confidence that we had in using the CDDL that said "we believe there's going to be a lot of contributions from folks who want to make contributions back" as opposed to telling them "you must" and dictating the terms under which they can do so.

Nat: Have there been a lot of contributions coming back?

Jonathan: Absolutely. There's a huge community, I think there's probably 7 or 8 thousand folks registered worldwide and user groups popping up all over the world. And again, two million licenses had to go to someone. And there's a curious thing about those licenses being downloaded. There's 100% attach rate to a developer and a computer, and last I checked that was an important demographic for us.

Nat: So I guess the important figure for you will be the ongoing use of the software they've downloaded. Is there any way to measure that?

Jonathan: For me I think the important thing is that we have more innovation in the marketplace, more awareness that open source is safe and can be used to run businesses as diverse as General Motors or Yahoo!. I would be thrilled to see a thousand distros of Open Solaris propagating in the world. I'd be thrilled to see cooperation and collaboration on up-stack components where we don't necessarily have to get into a debate on license religion. And I think we're going to see a growing number of registrants and participants.

At the end of the day, the most important thing for an open source project isn't the number of community of contributors, it's the number of users. I think the number of users that will be served by Solaris is obviously: how many users will there be of GM. This won't just be about folks and their IT shop. It'll be about those of you who drive your GM cars being able to ping back through OnStar--you're hitting Solaris. Or those of you who are in the Apache community and you're hitting the Apache servers, which are also running OpenSolaris.

Nat: That's interesting. I'd heard that Sun in the early days of the Internet gave away a lot of hardware to try and seed the spread of the Internet. Now I hear that's starting to happen again with the open source communities. Can you tell us something about that?

Jonathan: We will to continue not to simply be a user of open source technology and put up a poster showing you all the products that we're harvesting from the marketplace to use. We're going to continue to make contributions back to that. Every project we have at Sun, and it may not be there now, but every product at Sun will at some point be free or open sourced. Every one.

Which means that we will be taking the seven or eight thousand software developers at Sun and they will be working on open source and free products.

Nat: That's including Java?

Jonathan: In one form or another, yes. As you well know, Java's already free which is why we're downloading about 20M runtimes a month. And for those of you that missed that: 20M downloads a month.

Nat: Firefox just broke the 70M mark.

Jonathan: Cumulative.

Nat: Yes.

Jonathan: We're on a race. Mitchell and I were conspiring about how we can escalate each others' numbers. So I think all of that's goodness. That's just creating more goodness and more opportunity and at the end of the day we're as focused on getting users to adopt as we are about bragging about it on a slide.

Nat: So glad I don't have your network bills. 20M a month.

Jonathan: It's a big bill, trust me.

Nat: So what are the things the Sun was recently involved in with the European software patent fight. What's your take on software patents?

Jonathan: It's a complex question and I think --- [audience laughs]

Nat: Can you give us a simple answer?

Jonathan: I think software patents are largely inapplicable. I think patents have been misused to stifle innovation rather than promote and protect it. But I think the reality is they can serve both roles. [applause]

Look, we've been a victim of patent litigation. Just go back and look at our filings. We paid the Kodak corporation ... when you go buy a camera, just remember that Kodak sues us for intellectual property over a single patent which they acquired for the purpose of litigating. I don't think that's consistent with the objectives of the original patent philosophy. [applause]

We are not a troll. We're not out there finding people to sue. We're trying to protect the community. We say "if you pick up OpenSolaris, and going forward with all products at Sun, to the extent that if you're a licensee of any of those products then you've inherited the protection of all of the 5,000 patents we have related to that technology", which again continues to drive the umbrella forward. Just as Red Hat is picking up patents, just as SpikeSource is picking up patents, just as lots of open source community companies are picking up patents, we've got a huge warchest. Our intent with that warchest is to use it to protect the innovation which ultimately creates market opportunities for Sun, not to go through it and say "ok, we've got a patent here, who do we nail with it?" Not going to be the most productive way of driving innovation.

Nat: So if the Linux kernel were to implement DTrace, Sun wouldn't employ the patents against them?

Jonathan: Knock yourself out.

Nat: Wonderful to hear, thank you. Well, so much for that curve ball.

Jonathan: If you haven't seen DTrace, please go download Solaris 10 and check out DTrace.

Nat: So what are the reasons that people should use Solaris, given that Linux has been the open source poster child for many years?

Jonathan: A couple of things. There's no such thing as Java in the sense that you have to look at a product, there's an ideal called "Java". There's no such thing called "Linux". Linux is an ideal, a wonderful ideal, and then there are products. So if you compare Solaris to a product, let's engage in a discussion about the diversity of the way we approach a problem: "How do we deal with DTrace and diagnosing production systems" and those are the things we are focused on.

I think in respect to how do we compete, Open Solaris vs Linux, there's multiple products just like there's multiple distros out there. Individuals are going to make choices about the licenses they want to use, the conventions, the communities they want to be a part of, and they're all going to make individual choices. The good news is that unlike a few years back, and I think we need to update Andrew on what's happened with Solaris. Given that it's now available on every x86 and x64 computer made in the marketplace, given that the code is completely available, now we're not engaging in a religious or philosophical discussion of Solaris v Linux, we're saying "here's your product, here's our product, let's go compare them. Let's go run performance benchmarks and talk about functionality and diversity". That's what we can have the discussion on.

It's no longer about competing against a social movement as the big evil company that has proprietary products that they're trying to withold from you. Now it's about "look, let's get the politics out of the way and talk about innovation". Where should Red Hat run well, where should Solaris run well, and how do we may collaborate and cooperate in ways that we haven't historically to go put pressure on the real company who has a problem open sourcing their code. It's not just Yahoo! with Messenger and Flickr, which aren't open source, it's also a few other companies that build operating systems that are yet to make their code available.

Nat (coyly): I'd like to know which one you're thinking of ...

Jonathan: (pause) I really believe that HP should open source HP-UX. [Laughter] And particularly on Itanium. I believe it's important for my home server that I get that.

Nat: I'm not sure they're prepared for the 20 downloads/month they'd get.

Jonathan: Yeah, they'd have a hard time keeping the Itanium infrastructure up.

Nat: So what's next for open source. You're a big thinker in the area, you blog regularly about what catches your mind. What do you see coming?

Jonathan: The first thing I see happening, and I think Andrew and I and Jeremy [Zawodny, a keynoter from earlier in the morning] would all agree with this: there is just a tidal wave washing over the marketplace that suggests "if you're really committed to interoperability and choice for consumers, you must deliver your code into the market place". Customers are demanding that. That's number one.

Number two: the price of delivering software is going to zero. Remember back 10 years ago, if you wanted a product you had to go to a retailer. Spend some money, they gave you a box wrapped in plastic which you threw away, take the CD and then you'd throw that away. That was paying for access to products. That's obviously going away, now everything will be freely available and distributed over the web.

Innovation is increasing. Look at Firefox as opposed to IE. Which would you pick? I know which I'd pick. (That's Firefox by the way)

Nat: IE on Solaris is still lagging?

Jonathan: It is. We've been working with Microsoft but ... anyways. It's driving innovation because you're not going to compete on the basis of how much you cost. Both are going to be free, both are going to be available, both are going to be open source. Guess what? You're going to have to start innovating, or you're going to be ploughed under.

Comments I've made recently: why did NetBeans get so good? One word: Eclipse. Period. Having competition is a good thing. Our showing up in the marketplace to go compete against some of the OSes that historically were the only ones available under open source is going to put a little pressure on those other operating systems to get a little more innovative. Certainly we've seen that pressure, we've felt it.

I think the combination of those three things: more free products available in the market place, the cost of those products going to zero, and the services. My favourite examples are Yahoo! and Google. My favourite story about open source: I was on a panel with Dan Rosensweig who is the President of Yahoo! (and a really good guy) and John Markoff who was doing the interviewing. He totally threw me a softball. He said, "I was just up with Bill Gates, and Bill Gates says that open source and free software are communism. How do you feel about that?" I said, "interesting. Yahoo! is a free search service. Has anyone ever called you a communist?" He said, "no!" They'd just announced a billion in earnings.

I think we're moving away from simply free software to free services. If you look at the adoption around the world of the number of people coming online, why is it accelerating? Because it's getting so cheap. That's a wonderful thing, that creates more opportunity and more marketplace for all of us in the room.

Nat: So if there's free software and free services, where is there money to be made?

Jonathan: You have only to look at Yahoo! and to look at Google ...

Nat: Advertising?

Jonathan: Well, I think it depends on what you're going to do. I think that Kim [Polese, a keynoter from earlier in the morning] with SpikeSource has probably found another mechanism. "Free software", all it really talks about is that the price of software has gone to zero. That doesn't mean there's no value in it. The value may be in service and support, the value may be in an update network (as with Red Hat), the value may be in ads (in the case of Yahoo!), it may be in location services if you're Google or Vodaphone. I think there's more and more opportunity.

The perfect analogy, and I know it ruffles some feathers to say it. How many of you have a mobile phone? [audience raises hands] Leave mine alone, by the way.

The most popular handsets in the world are free. Why? Because they're free! It's not because the code is available to them or you can disassemble the phone and reassemble it on your own, it's because they're free. And that's a good thing.

Does anyone wonder how the carriers are going to make money off of free handsets? No. The ones that have to wonder about it are going to go out of business quickly. The ones that are monetizing it, there's a huge amount of money for them to go after. I think it forces us to have a different conversation. It's not about how much you pay when you go to Egghead to buy a product, now it's "what's the value that this provider will offer to me that's above and beyond what the rest of the marketplace offers."

Nat: I think there's an important distinction between open source software and cellphones. Cellphones are fundamentally opposed to being open source, the mobile networks are fundamentally opposed to being open sourced.

Jonathan: How many of you believe this is a sustainable strategy? I don't see a lot of hands.

Nat: So you're saying that the cellphone is a model for the new economy, but that it's also unsustainable?

Jonathan: No, I'm saying that the economic model of the cellphone community-- And the economic model of office productivity. Why on earth would you buy Microsoft Office if you can get OpenOffice for free?

Nat: Can I say ... in my experience, OpenOffice has been slow and buggy. So they'd pay for something that's slow and buggy in different ways. [laughter]

Jonathan: Have you used StarOffice 8? I would invite you to return to using StarOffice 8. Obviously, we're working on the slow and buggy part along with the community. It's good news, there's competition there.

I think the economic models are distinct from the network models and the openness. Look, Yahoo!'s not open source. It's not. If you want to get access to source code at Yahoo!, there's some that they'll provide but they're not going to give you it all. But Yahoo!'s free as a service and they've found other ways to go build value around it. I think the economic models of free products are certainly changing because the network is so simple now to connect with a consumer opportunity, but I think the technical models and the intellectual property models--there's going to be a lot of change there in everything from patents to closed networks to strong authentication.

Nat: So for you, the "free" in "free software" is all "free" as in "beer".

Jonathan: No, I think it's a see-saw and see-saws don't work unless you have somebody on both ends of it. Part of it is free as in the economic definition of free, and part of it is "freedom", which is the freedom to make choices. Absolutely I believe in both.

Nat: Fantastic, thank you for joining us today.

Jonathan: Likewise, it's been a pleasure.

Nat: Don't forget your cellphone!

tags:   | comments: 1   | Sphere It

Previous  |  Next

0 TrackBacks

TrackBack URL for this entry:

Comments: 1

  Daniel [11.03.05 03:39 AM]

We've also just posted a conversation Tim had with Jonathan Schwartz and Mitchell Baker at Web 2.0 this week on

Post A Comment:

 (please be patient, comments may take awhile to post)

Type the characters you see in the picture above.