The Future of Technology


20 Feb 2004

University Foundation in Brussels

About 20 invited guests, including IT industrials, ICT authorities and VUB professors sounded out, together with invited speaker Tim O’Reilly, who elaborated on the future of computing and Open Source Software.


Rector Benjamin Van Camp welcomed the invited guests and the invited speaker Tim O’Reilly. The Rector hinted at the fact that maybe VUB Doctor Honoris Causa Richard Stallman would drop by and according to CROSSTALKS operational director Marleen Wynants, the chances were pretty open...

VUB Professor Viviane Jonckers introduced Tim O’Reilly. O’Reilly is the founder
and CEO of O’Reilly & Associates, often cited as the best computer book publisher in the world. But O’Reilly is in the first place a most innovative entrepreneur and pioneer in the netculture. Furthermore he aims at catching and communicating the knowledge of real innovators through books, online media and conferences.

Tim O’Reilly: I just want to start with the traditional wisdom on open source – the basic idea some people have is that open source is about software licensing terms and in a large absolute sense that is true. But people also think that open source is anti-business, and there is various elements in the free and open source community that do say that anything that is not free is bad. The reality though is that open source developers are fairly comfortable with a mix of free and proprietary since it’s just not about licensing. The boundaries between the two are very blur in many open source projects.
The tradition I come from and which I represent is the Berkeley Software Design or BSD tradition. It’s one of the main open source traditions together with the free software tradition of Richard Stallman. The model of BSD being: “We’re doing research and we want to share that research. And what you do with the research once we’ve given it to you, is up to you.” At Berkeley, we believed in “copy central” until Richard Stallman came along and said copyright is bad and started this new thing called copyleft. We figured that when we gave out the code, people could take it down and copy it. It was a kind of university research tradition to increase the sum of human knowledge and to disseminate it. And the Berkeley tradition has always assumed that somebody will take that source code and will build proprietary software on top of it.

But regardless of whether you are at the Berkeley side of things or in the Free Software tradition of Richard Stallman, I believe that the licensing debate is largely a red herring since it doesn’t represent the real significant thing of open source. That’s why I’m gonna talk about the paradigm shift. Everybody of you has probably at some point in your career read or heard about Thomas Kuhn’s book from 1963, The Scientific Revolution. I refer to it since from this book stems the idea that paradigm shifts apply very powerfully in technology and computer science industry. Let’s go back 23 years ago. IBM made a decision - we don’t know really why they did it and the guy who did it was killed in a plane crash not long after that, so we cannot go back and ask him what he was thinking, - but for some reason IBM changed the world through the decision that they were going to produce a personal computer. And they were so open about it that they gave away the specifications so that anybody who would want to make a similar personal computers, could do so. And as we know, that PC hardware model took over the world. There were a lot of competing companies, Apple being the only survivor. Apple is the only company who is still in the same old business as IBM was: you build an integrated system of hardware and software. When IBM released these PC specs, they changed the industry in a very profound way but they didn’t fully understand the implications of what they did. And so for the first decade of the PC revolution – everybody was doing what they were used to doing: trying to build a better pc. Some, like Dell and Compaq were looking for a new proprietary bust that would sort of enhance the PC standard. Until Dell realized that they were actually failing at building a proprietary bust and they started to embrace the commodity model. So they became better at logistics and sales, at marketing and distribution and now Dell is still the largest pc vendor.
But there was a second and more profound element to the paradigm shift: and that was that once you had commodity hardware, software became separated from it. If you had a commodity hardware model, you could have a commodity software model! IBM made a massive miscalculation by saying, we control 90% of the market, so of course you little guys from Microsoft, you can go and build software for the other 10%. What they hadn’t realized was that the 10% would soon be the former 90%. Not soon, it took 20 years, but my premise is that we are at the same paradigm shift happening with software.

I’ll ask you a question which will probably illustrate where we are caught in old thinking and need to move to new thinking:
How many of you here use Linux? (about 4 people of the 18 dinner participants raise their hands)
How many of you here use Google?(about 14 people present raise their hands)
You know that Google is the largest Linux installation in the world? (laughter) Right. But what you’re telling me here is that you are thinking in personal computer terms. And I think that the most important what is happening with open source is this “getting out of the personal computer”.

You also probably heard already this discussion between open source advocates and open source opponents who say: ”There are no user friendly applications on Linux!” and the Linux advocate says: “O yes! Open Office is getting really good and Gnome is actually catching up to Windows.” But there are already killer applications like Amazon and Google, that are used by tens to hundreds of thousands of people and they are all running on Linux!
So people are just missing that the rules have changed again and I think we have to start thinking on what the nature of this paradigm shift is that we’re in the middle of. The paradigm failure is deepest in fact in the open source community. People operating in Linux companies are fixated on catching up with Microsoft when in fact we’ve already gone far beyond Windows. The greatest success of open source was in fact the internet. The internet came out of funded work as part of the Berkeley project – released in the public domain effectively and Microsoft took it up in their products. All kinds of people from commercial business built their thing on top of it but with the internet, we did get a new infrastructure which is an open source infrastructure. On top of that we built these new application layers like Google, Amazon, online mapping, and we’re just at the beginning of that stage of course. And if we’re looking at some of these applications, we see that the source code is not free. Also the old paradigm, you have to give someone software to run on a local machine, that doesn’t happen any more! Google doesn’t give out any software, you use Google simply by going to their website. It’s a new set of rules. So the open source licenses are somewhat irrelevant at the layer of software. There is also an issue that the value of these applications isn’t entirely in their proprietary software. But there are fairly substantial analogies between what has and what Barnes& has or between what Google has and Alta Vista had. So what are long term trends that are at the root of the open source paradigm shift?
I think there are three. The first being that software is becoming a commodity. Not all software, if you look at the hardware paradigm change, not all hardware became a commodity either. In fact we built a new monopoly business in Intel so that all those commodity pc’s got labeled Intel Inside. So there are opportunities for pieces of proprietary IT embedded in a large commodity system. Look at Amazon and Google, they could not have existed without commodity hardware AND software. They build their thing out of really cheap commodity parts, they didn’t have to buy anything proprietary. Amazon said that when they switched from digital Unix to Linux, THAT was the critical step for Amazon to become profitable. But it’s also that these systems built on top of the open source platform are also highly user customizable. And not just at the level of software.

The second trend is that all of these next generation applications have a large user engagement component. So think about E-bay, that’s very obvious: the customers ARE the application. But it also counts for Amazon, Amazon has figured out a way to have its customers contribute, not to the software, but at the data-layer. More than 10 million customers supplied reviews to and that’s actually the critical difference between and say Barnes& All of the publishers content is the same, but there’s a greater flow of customer attention and more user content that gives the site much more value. That’s a principle we could take from open source software and this principle of collaborative engagement, of an architecture that encourages participation should be brought up to stack the data layer. You can argue the same point, less forcefully with Google. The critical advance that Google brought to the search engine game and blew away competition like Alta Vista, was the page ranking algorithm: What do the users of the internet think about these pages? Every time anybody lights a link on the web, one contributes to Google. Again, that technology is now copied by other people but it was a big shift in thinking about search engines. You see that often when technologies are becoming commoditized. Also in the car industry, everything become more driven by design and branding and the like. Architecture critic Dave Hickey said: “After World War II, General Motors stopped marketing cars for what they do, and started marketing them for what they mean.” And the computer story starts to go in the direction of what it means to use an Apple... Google beat Alta Vista because they figured out how to do cool branding. Which is another byproduct of software commoditisation.
There is a third trend – the one about network enabled collaboration. Open source really grew up alongside the Internet. It’s what developers do! And a lot of the benefits of open source are about network effects. Before the internet, how would you be able to have Linus Thorvalds from Finland and Alan Cox from Wales collaborate on a project? This massive move shifted the power from businesses to individuals because the network allows individuals to associate around a project. Most of the key open source projects are international, with key developers who never met in the flesh. At least they do now since open source is big business, but at my first Perl conference in 1997 - before I organized the meeting in which the term open source was adopted - most of these people had never met! And even in 1998 at the so-called Open Source Summit in Palo Alto, Linus Thorvalds still had never met Larry Wall who created Perl – and I said “Oh, we gotta do some marketing here and bring this together as a movement”. There were all these different communities who were basically reaching out across the net. I’m gonna quit here and prefer to enter into discussions with you.

There is one essential thing though to reflect upon, especially in a group like CROSSTALKS. When we are thinking about the future of computing, I believe that the fundamental future that we are facing, is that we are moving to a network platform on which we will look back and say, “Yes, we were building an internet operating system”. The question that we all face is: What kind of operating system is it going to be? Is it going to be an operating system that is controlled by a single entity? Microsoft believes firmly that that would be a good thing to do. They could built the next layer and are going to own it just like the previous one. But the market rejected it because it was Microsoft but I believe that a lot of people to a more of lesser degree have the same ambition! And it is very easy for business models when you don’t understand the paradigm, to fall into the place of somebody who doesn’t expect it. Obviously IBM didn’t know what they handed over to Microsoft about 20 years ago. And on a smaller scale we’ve seen similar things happening on the internet: the most critical piece of software on the internet is BIND – the Berkeley Internet Name Daemon - software that runs the domain name system, maintained by this long haired programmer in Redwood city California. It was worth hundred thousands of millions of dollars a year, a monopoly business of which nobody knew it was going to be valued... until it was handed over to monopoly! And I think we’re going to see that happen again and again : that somebody is going to fall onto something that turns out to be quite critical and make it big business. Because we haven’t thought long enough about the long term trends yet! I think that open source has really triggered a set of cascading changes that are well beyond of what we normally think of – the software licensing debate – and is really part of a much bigger movement. We have to think deeply about what kind of next generation we are going to build. I think that it’s going to be one that has a lot that coming from the LINUX and internet heritage: a large architecture with lots of pieces that are loosely joined... Basically it will be systems that are loosely coupled by well-defined open standard communication protocols rather than a tightly integrated system that is controlled by a single player. And that’s the challenge that we as an industry have to make sure happens!

So far the talk given by invited speaker Tim O’Reilly. The discussion started right from here on and continued even more informally while dinner was being served. Somewhere in the middle of the hors d’oeuvre, the doors of the dinner room were opened and VUB Doctor Honoris Causa Richard Stallman stepped into the room. For the rest of the story and the animated tabletalks that took place, you’ll have to talk to one of the participants of the dinner or wait until the first CROSSTALKS publication...

PARTICIPANTS: Benjamin Van Camp (VUB Rector) Viviane Jonckers (Vice Dean of the Faculty Applied Sciences -VUB) Marc Nyssen (Prof Medical Faculty-VUB) Dirk Van Berlaer (Prof Faculty of Economic, Social and Political Sciences VUB – Board of Directors Ethias) Jacques Tiberghien (Prof Faculty Applied Sciences VUB) Onno Timmerman (Sp.a Research Cell) Karel Uyttendaele (Head Belgian ICT Policies with State Secretary Peter Vanvelthoven) Bart Mariman (Business Development Director Siemens) Walter Van de Velde (Scientific Director DISC) Bruno De Vuyst (VUB consultant Industrial Policy/Manager BI3 Fund) Steve Dierckens (Client Relationship Manager Public Sector IBM) Dirk Van der Sanden (Market Planner- Univation Technologies) Sonja Haesen (Head Interface Cell VUB) Marleen Wynants (Operational Director CROSSTALKS)