Jul 19

Tim O'Reilly

Tim O'Reilly

Four Big Ideas About Open Source

In my O'Reilly Radar Executive Briefing next week at OSCON, I'm focusing on four big ideas about open source:

  1. The architecture of participation beyond software. Software development was the canary in the coalmine, one of the first areas to show the power of self-organizing systems leveraging the power of the internet to transform markets. But it didn't stop there. What we're now calling Web 2.0 is a direct outgrowth of the core principles that made open source software successful, but in my opinion, many of the projects and companies that make up the Web 2.0 movement have gone far beyond open source in their understanding of how to build systems that leverage what I call the architecture of participation.
  2. Asymmetric Competition. One of the most powerful things about open source is its potential to reset the rules of the game, to compete in a way that undercuts all of the advantages of incumbent players. Yet what we see in open source is that the leading companies have in many ways abandoned this advantage, becoming increasingly like the companies with which they compete. I have no concerns about the ultimate health of the open source development model or the vibrant creativity of the open source community, but I do question whether open source companies really grasp the implications of the new model. I think that if they did, they'd be Web 2.0 companies.
  3. How Software As a Service Changes The Points of Business Leverage. Operations and scalability lead to powerful cost advantages; increasing returns from network effects lead to new kinds of lock-in. The net effect is that even when running open source software, vendors will have lock-in opportunities just as powerful as those from the previous generation of proprietary software.
  4. Open Data. One day soon, tomorrow's Richard Stallman will wake up and realize that all the software distributed in the world is free and open source, but that he still has no control to improve or change the computer tools that he relies on every day. They are services backed by collective databases too large (and controlled by their service providers) to be easily modified. Even data portability initiatives such as those starting today merely scratch the surface, because taking your own data out of the pool may let you move it somewhere else, but much of its value depends on its original context, now lost.
In short, you can see that I believe that there are serious challenges to the open source model. For all its success (and that success has been world-changing), it's important not to get complacent. The world is changing under our feet! The pendulum always swings between open and proprietary, and despite the apparent progress of open source and open standards, right now the pendulum is swinging the other way.

 I have always felt most passionate in preaching to the open source community. It is one that I love and esteem most highly. But it is also one that is in great danger of increasing irrelevance, because some of the premises on which it has based its thinking are wrong. I've always had a perspective a bit at variance from those of other open source advocates, and as the future unfolds, I continue to feel that that perspective is essential for open source strategists to understand and embrace.

(For a succinct recap of the evolution of that perspective, see my 1998 essay, Hardware, Software and Infoware, my 2003 essay, The Open Source Paradigm Shift, and my 2004 essay, What is Web 2.0? Also, see my previous entry about the Radar Open Source briefing session.

tags: open source, web 2.0  | comments: 17   | Sphere It

Previous  |  Next

2 TrackBacks

TrackBack URL for this entry:

I'm not there, but everyone else seems to be.

Search Engine Watch has a nice rundown of some of the ...

Read More

I'm not there, but everyone else seems to be.

Search Engine Watch has a nice rundown of some of the ...

Read More

Comments: 17

  Dan Zambonini [07.19.06 06:02 AM]

Sounds like it will be excellent discussion; I really wish I wasn't on the other side of the pond, and could make it.

I'd particularly be interested in hearing if you have any suggestions for the 'open data' problem, and whether you think Semantic Web technologies will play a role (or will perhaps be a distraction) in making our data more accessible? And, how can we de-centralise our data, and yet increase our ownership of it at the same time?

These are interesting times.

  A friendly neighbour [07.19.06 06:51 AM]

Now, writing short and distinctly has its advantages. But when it's too short it feels like buzzwords and hot air, without anything to grasp. For example:

> but I do question whether open source companies
> really grasp the implications of the new model. I
> think that if they did, they'd be Web 2.0
> companies.

I am sure there is valid and hard thinking that went into creating this conclusion of ours. But given that you present us only an assertion, without any details on how you come to that conclusion (or what you specifically mean by it), I didn't find your post all that enlightening.

I am sure most of the follow-up discussion will be centered around clearing up confusions about what was actually said or meant.

  Anonymous [07.19.06 07:50 AM]

Is the "network effect" the same thing as the "net effect"?

  Tim O'Reilly [07.19.06 08:03 AM]

A friendly neighbor - That's why I linked to the three essays at the end, which lay out my thinking in much greater detail. I didn't repeat it, because I've already written it. See The Open Source Paradigm Shift in particular.

But specifically as to why I question whether open source companies completely get the new model, here are a couple of things to chew on:

1. Given that web 2.0 is deeply about software as a service, in which operations is a core competency, the fact that open source companies keep their various deployment and operations tool as part of their proprietary value add is probably wise from a business point of view, but at the very least shows that their point of leverage is not necessarily what they claim that it is.

2. A big part of the new model is lightweight, rapid development, and services that are always changing, it concerns me that companies embracing the enterprise market are working to manage user expectations by slowing down the pace of development and creating the enterprise fork (cf RHEL vs. Fedora), rather than by building next generation tools to keep up with what is ultimately an internet-era software supply chain.

In short, I still see a lot of the open source companies acting like early PC companies (pre-Dell). That is, they were born in this new open world, but as they grew up, they emulated the practices of the world that went before. Eventually a company (in the PC case, Dell) came along, and more fully embraced the commodity world by building smarter, more agile processes, and got its advantage not from trying to add back in what was previously proprietary, but in discovering a proprietary advantage that was really new.

Compare Red Hat and Google, for example. Both are great companies. I believe both were born from the same vortex of possibilities. But as Red Hat has matured as a software company, it has taken on the business models of previous generations of software companies, while Google and other Web 2.0 companies took a different fork in the road.

It's tough to really embrace new models -- heck, we really struggle with that at O'Reilly as well, and don't have easy answers. But I can still see what I can't necessarily achieve. As Robert Browning said, "A man's reach must exceed his grasp, or what's a heaven for?"

  Roman Bischoff [07.19.06 09:13 AM]

Can you describe what it would mean for open-source companies to get Web 2.0 companies?

Run some scalable servers, provide Web-UI, some APIs and make use of some other existing Web 2.0 APIs?

For me, in the extreme exmple case of Sleepycat Software, Inc. it's questionable whether they could and should provide their software as service.

Is the number of financially successful Web 2.0 companies bigger than the number of successful open-source software companies? (I don't know)

  Sid Steward [07.19.06 09:22 AM]

I have doubts about the relevance of open data to web 2.0 users. Does a web 2.0 user really have anything in common with a free software (since you cited rms) advocate?

  Tim O'Reilly [07.19.06 09:48 AM]

Sid, you completely miss my point. RMS got all fired up about free software because as software became proprietary, he discovered that he could no longer fix problems with the software he was using. My prediction is that the proprietary, closed data problem will eventually hit users the same way, and an "RMS of the future" will start an equivalent movement about the rights of users to their data. (In some ways, these things are already bubbling up. See for example

  Sid Steward [07.19.06 10:42 AM]


Thanks for the reply, and thanks for your rich article and follow-ups.

I think you will have folks who are fired about about closed data, but that an open data movement won't go mainstream. Here's why.

Free software is a fringe culture (of which I'm a member), while web 2.0 is gaining momentum because it's moving into the popular culture. If mainstream folks were bothered by lock-in, then why don't they rebel against Microsoft Office's proprietary formats? Why don't they use Firefox or Linux? Because they don't mind the lock-in — it's part of the deal.

As long as web service providers treat their users tolerably well, their users won't mind the lock-in. GYM (goog, yhoo, msft) knows this.

Branding also factors into it. "Yahoo! or Google?" will be like: "Coke or Pepsi?" What do your friends use? Which is cooler? Is open data cool?

You will certainly have a fringe who gets fired up about closed data. I'm one of them — that's why I started kicking around the web cooperative idea. But I don't think the movement will go mainstream.

On the flip side, I also wonder how much free software folks care about cutting edge web 2.0 services. Talking to one free software adept about my web co-op idea, he replied: 'I've read that web 2.0 is about tagging ... tell me: what is tagging?'

Thanks for the link.

  Tim O'Reilly [07.19.06 10:50 AM]

Sid -- you wrote: "I also wonder how much free software folks care about cutting edge web 2.0 services. Talking to one free software adept about my web co-op idea, he replied: 'I've read that web 2.0 is about tagging ... tell me: what is tagging?'" My point exactly. I'm trying to get Web 2.0 on the radar of people who *ought* to be right in the middle of it, but who really aren't.

Regarding the consumer vs. geek edge comments -- important movements often start on the fringe, and hit the mainstream only indirectly. But that doesn't mean they are unimportant.

  Sid Steward [07.19.06 10:59 AM]

Tim- True and true. Thanks for taking the time to reply.

  Tim O'Reilly [07.19.06 11:26 AM]

Roman --

I think the scale of success of Web 2.0 companies definitely shows that delivering services rather than selling services to support software is a better business model. And I'm not just talking about companies like Google, who've delivered a higher level set of software as a service. But consider even the open source infrastructure that fuels the web hosting market. The real PHP play wasn't Zend, it was rackspace and their peers -- any hosting provider who effectively sells access to PHP as part of their offering.

For years, I used to point to Uunet as the first great open source success story. Rick Adams, the founder, was the author of B News, the dominant usenet news package at the time, and SLIP, the predecessor to PPP, as well as the hostmaster of the best-connected usenet and mail hub of the day (1988-1989). Uunet was one of the first ISPs -- originally a non-profit funded by Usenix -- but soon one of the main instigators of the commercial internet. And what were most ISPs and hosting providers selling? Subscription access to free software stacks, as a monthly service. (Microsoft later entered that market.) As for Rick, he retired a billionaire.

The SaaS story has been central to open source from the beginning. That model took hold for some programs, but not for others. And it became the dominant model for people who went further up the stack.

I hope this doesn't come across as ragging on Red Hat. They are a great company. But I still think there's a lot more to think about in how the internet can change the economic models for how we do business with free software.

  Stormy Peters [07.19.06 03:22 PM]

I read your post and your web 2.0 essays and went "wow, that's right on, that's how the world is changing." Then I read the comments and got confused about how people were interpreting the original post.

I think the point is that open source software changed the software industry (and I agree many of those companies have lost their "open-sourceness" and are now competing in traditional software company models), and just like open source changed the software industry, "open source" is changing other industries. Because I share my book reviews and my book wish list, how I look for new books has changed. Because I share my bookmarks on, how I find interesting new websites has changed. The world is changing because of web 2.0 and it is very related to what made open source successful and it is very much the same community that is jazzed about it, whether they call it web 2.0 or not.

For an example of another company that didn't get the web 2.0 model - Netscape posted today offering to pay Digg posters to create bookmarks on Netscape's social bookmarking site. I blogged about it here.

  Sid Steward [07.19.06 03:36 PM]


The open data discussion needs more teeth &mdash it's too theoretical. Here's a non-threatening open data application that all web service providers could rally around: a standard api or feed for accessing account logs.

All services generate logs, and it would be nice for users if they could monitor that data from a central location.

It might be unlikely, but I think it's a good, concrete example that could be used to fuel conversation.

  Sid Steward [07.19.06 03:52 PM]

Edit – replace: "standard api or feed" with: "documented api or feed." That puts it in the realm of possibility.

  Tim O'Reilly [07.19.06 04:01 PM]

Stormy, you are very much on my wavelength with your comments. It's tough to figure out how the open source/web 2.0 model applies in every case, but it's often easier to see where it's NOT being followed. (That was the point of my levels of the game post a few days ago.) Open source is more rooted in the internet than many people in the open source community realize, and so their companies move away from or under-emphasize things that should be central to their strategy.

  Swashbuckler [07.20.06 10:03 AM]

You really like the "canary in the coal mine" metaphor, don't you?

  C. Enrique Ortiz [09.04.06 07:07 PM]

As I read these "Four Big Ideas About Open Source", I realize (that with minor changes and reordering) they also are 4 principles that describe the current philosophy of how services on the web are produced and consumed...


Post A Comment:

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

Type the characters you see in the picture above.