Interview: Tim Bray opens up about open source

Co-inventor of XML also talks about the future of XML, blogs, and document formats

Tim Bray is director of Web Technologies at Sun Microsystems, but is perhaps best known as a co-inventor of XML. He also has launched one of the first public Web search engines, Open Text Index, and founded Antarctica Systems, specializing in visualization-based business analytics. Additionally, Bray publishes a blog and co-chairs the IETF AtomPub (Atom Publishing Format and Publishing Protocol) Working Group, which is focused on technologies for editing Web resources such as blogs and wikis. InfoWorld Editor at Large Paul Krill spoke with Bray at the Open Source Business Conference in San Francisco last week about topics ranging from open source and blogs to document formats and, of course, XML.

InfoWorld: With the Open Source Business Conference happening here, how do you leverage open source and be profitable with it?

Bray: Well, the notion is that the different kinds of business models you have around software are not specifically a function of whether it’s open source or closed source. I think even if you wanted to do traditional capital cost software licensing, the kind of thing that Oracle still makes a living on, there’s no reason in principle you couldn’t do that and still have the product [be] open source and anybody could download it and compile themselves. [We did an announcement about] Project Red October. All the software products are freely downloadable and freely usable, but unless you license them, whether that be by capital cost or subscription, you get no indemnity, no support, no nothing. I don’t actually see any tension between open source and making money. Open source is a good characteristic for software in general to have. It seems like we’re all beginning to agree, and our plans for making money from software do include open source by and large.

InfoWorld: The gentleman from Microsoft this morning basically said he didn’t think the model of selling tech support calls is going to be sustainable. Can you make money just selling support without taking any money for the software itself?

Bray: Well, we think so. Yes. Absolutely.

InfoWorld: How long have you been doing that and how has that worked out so far?

Bray: Oh gosh, I’m not on the business side of the software organization. I just don’t have a good clean answer for that one. But I will say that the historical model of trying to make your money by software [rights] to use licensing costs, I really do think that’s going to become increasingly unsustainable. Trying to cause people to pay money for something that is duplicated and the cost is zero, is at some level profoundly unnatural. And another issue that has always bothered me as a person who’s been not just a technologist but a business executive, software licensing tends to produce some profoundly bad accounting. You know, trying to treat expenditure for software as a capital cost which you then depreciate gets you even further than one normally goes into their own accounting science fiction. In fact, software is a piece of business infrastructure that you have to have there to work on a day-to-day basis, and I would think should be expensed in a rational way. And I think that was also the main thrust of ["IT Doesn't Matter" author] Nick Carr’s speech today, about this notion of building your own capital cost IT infrastructure, as opposed to getting it like electricity on a subscription basis, is it probably doesn’t have legs for the long haul.

InfoWorld: I had somebody else at another company tell me the other day that open source brings down revenues for everybody. Would you agree with that?

Bray: Well, we don’t know yet. Right now, we’ve got this completely ad hoc hodgepodge of business models out there. Clearly, Microsoft makes the most money of anybody and they’re not open source at all. But then again, they’ve got a monopoly to help them. So the question of how you would maximize the revenue for the competitive marketplace is one I don’t think we know the answer to yet. We’re betting that closed source is sufficiently unnatural in the long haul, that open source is the way it’s going to end up.

InfoWorld: At what point do you say, Our revenues are so much less now that we’re not getting licensing, that we’re going to have to cut your salary in half. I just wonder where does this all end?

Bray: Well I just don’t buy that picture. If you look at the typical subscription arrangements, such as the ones that our friends over at Red Hat are being quite successful with, the amount of money you end up paying, it may not end up being that dramatically less. It’s just that instead of it being as a capital cost, it’s as a monthly subscription kind of basis.

InfoWorld: I don’t have the figures in front of me, but haven’t Red Hat’s revenues not been all that fabulous? They're nothing compared to Sun or Microsoft or anybody like that.

Bray: Well, yes, except we’re not a pure software company, and they are. And in terms of the number of deployed seats they have, they have orders of magnitude less than Microsoft. I think actually it’s going to be a hard apples-to-apples comparison to make.

InfoWorld: Borland last week, decided to sell off not only its Java tools but its Microsoft tools. What do you see happening there? Is that open source or commoditization of tools?

Bray: I think that the conclusion is there’s no money to be made selling IDEs, that’s all.

InfoWorld: And why isn’t there?

Bray: Well, because there are three excellent ones. If you’re in the Microsoft space, you’re locked into [Visual Studio .NET], which is actually an excellent product. On open source in the Java world, there’s NetBeans and Eclipse, both of which are excellent; both of which are free. And in the LAMP (Linux Apache MySQL Perl/Python/PHP) world -- you know, dynamic language, that world -- there isn’t much at the moment. But what comes along will doubtless be open source. That’s now the cultural expectation.

InfoWorld: Could you maybe explain your role in the founding of XML and has XML exceeded your expectations?

Bray: In 1996, it was obvious that the Web was winning, and it was also obvious that people wanted to do a lot more things over the Web than just look at Web pages. And it was also obvious that HTML probably wasn’t the right thing to serve as the basis for what we now call Web services. And there was this earlier technology called SGML (Standard Generalized Markup Language), which had been used in big, big publishing applications, like the legislation of the European Union and the technical manuals for Boeing, which was good, but it was too big and too complicated. And so the population of people in the world who actually knew SGML and understood the Web was maybe 12 people. So that was the 12 people that ended up on the working group, and I co-edited the specs. I’m a co-editor of XML 1.0. And we basically took SGML, threw away the 90 percent that nobody ever used, made it a little bit more Web-friendly, addressed things with URLs, and that was XML. So I was in the middle of that group. Has it exceeded our expectations? Oh, my goodness, yes. We thought we were building something that would enable somewhat more efficient publishing of Web pages to multiple devices and so on, and the explosion of creativity and energy around XML has wildly exceeded anything we could have possibly dreamed.

InfoWorld: Where do you see XML going and where do you see Web services going?

Bray: Well, those are two wildly different subjects. I mean XML is basically being used everywhere that you have heterogenous computers, that’s to say heterogeneous technology where you’ve got different operating systems or different databases or different anythings, and you need to interchange data, well, XML now seems to be pretty much the de facto way to do that. So XML is essentially everywhere, but at the same time, it’s more or less invisible because [of] all that’s happening behind the scenes where you don’t see it. I think that the recent interest in XML as an office document format, both in the standardized version of ODF (Open Document Format) and the Microsoft stuff is a harbinger. I think there’s a huge upside for everybody in that when the document formats become open, stable, and sort of generally purposed reusable for everything. The world’s most successful XML application right at the moment is RSS. In terms of the volume of data, the number of feeds, and that is a huge source of change, not just on the technology front, but also culturally. The notion that we have a communication commons where anybody can write and there are no gatekeepers. It certainly doesn’t mean we’re going to do away with analysts or journalists or conventional letters or anything, but it does mean that the shape of communication becomes different. So in terms of impact on the larger human culture, that’s big. Switching gears to Web services, Web services is really confusing. The core notion of Web Services is that instead of trying to define enterprise APIs that extend across the network, as in CORBA or DCOM, you say -- well, we’re not going to define the API, we’re just going to define the messages. You know, I’m going to send you an XML message that looks like this and you’re going to send me back an XML message that looks like that. And the advent of XML was a [gating] factor because it gave us a lingua franca that we could agree to use in these messages. I think you’d be hard put to find somebody in my profession that doesn’t basically buy into that concept. The notion is that since everybody’s IT infrastructure is heterogeneous and is going to stay that way, and since everybody’s IT infrastructure is networked and is going to stay that way, messaging passing seems like the only sane way forward. Having said that, there’s this huge pile of technology, it’s generally referred to as WS-* (spoken as WS star), [where you have] 1,000 or more pages of specifications and various standards organizations, many of them still in IBM or Microsoft back rooms, and yet to emerge or be solidified. And the party line of some of the big players in our industry and some of the analysts, the Gartners, and so on, of this world has been, "Well, this is the future. WS-* is the future, and if you don’t do that, you’re in the wrong quadrant." Gee, it seems to be a long time coming, doesn’t it? I mean there’s been a lot of effort going on and the deployment of applications isn’t that big. So speaking officially for Sun, there is no doubt whatsoever that we will pursue that stuff. We will do what it takes so that our customers can interoperate with the WS-* [stack] to the extent that it actually stabilizes and ships.

InfoWorld: I was at a conference a few weeks ago where, basically, the verdict on Web services was that it was just so mired in confusion with these standards and specifications that people don’t know where to start.

Bray: I have a lot of sympathy with that viewpoint. We’re not promising to interoperate with whatever the WS-* castle in the sky is, we’re promising to interoperate with what Microsoft shipped. And we will do that. And I think IBM will too. So the core big picture vendors will actually interoperate with each other. So the interesting question is, what you raised is, does that change the world? Does everybody sign up for that? Well, I don’t know. I tend to be a skeptic. I observe that is doing tens of millions of Web services transactions a day and making money, and generally speaking, ignoring all that WS-* stuff. They are just doing straightforward XML over HTTP, without benefit of WS-*. And I think that that simple model, what we call the REST (Representational State Transfer] model, had a lot of legs and is going to deliver I think a lot of value, both in the Internet space and in the intranet enterprise space.

InfoWorld: So does REST supersede SOAP at some point?

Bray: No, I don’t think so, I think probably there’s value in SOAP. I could get incredibly technical here, but the notion of REST is that there are a large number of resources out there on the Web and a small number of verbs which interact with them. Basically create, replace, destroy, is a good model for building applications. Now it may be the case that to do that you would exchange messages that are in SOAP envelopes. But this notion of abstract multi-step Web services with intermediaries, that are [declared] with WSDL and published with metadata exchange and governed by WS-Policy, and so on and so forth, and where your new things are [passed] around with WS-Eventing and so on, I really have a hard time thinking that that will be as big a deal as the prognosticators say.

InfoWorld: You mentioned the documents issue. Do you think Open Document format will win out over Microsoft Office Open XML?

1 2 Page 1
Page 1 of 2