« March 16, 2004 | Main | March 18, 2004 »
March 17, 2004
OSBC2004 Wrap-Up
With OSBC2004 drawing to a close, I've got a few thoughts:
- First, Matt Asay and crew did a great job, especially considering this is the first of its kind. There were a few minor things that could be improved in overall conference organization, but the conference was great and so was the location.
- I found most of the presentations by company reps to be just that. They were, for the most part unimaginative descriptions of why that company was so great. The exceptions were Steve Korn from EDS (since he wasn't there to talk about EDS, but rather their clients) and Martin Fink from HP. Maybe companies trying to figure out a business model found them more useful that I did.
- On the other hand, there were some excellent speakers including Clayton Christensen, Tim O'Reilly, Ray Lane, and Larry Lessig. They were all outstanding and offered real value.
- While I didn't attend any of the legal tracks, there are about 200 lawyers at the conference. Open source seems to be getting quite a bit of attention from the legal community. Maybe that's a side effect of the SCO lawsuits. There are a number of cynical comments about lawyers one could make here, I'll refrain. :-)
- The attendee list included quite a few people I know and I enjoyed getting together with them and having a chance to catch up.
I wonder, as I imagine this conference next year, how it will evolve. Will there be significant changes in the open source world that will change the primary message of the conference or will it be the same message, more or less, presented to a new, and perhaps larger, set of attendees?
I guess this question comes from a feeling I've had at this conference that it lacks the energy of the O'Reilly Open Source conference (OSCON). That's a not a dig at OSBC, its just an observation that people who are creating new things tend to be excited and every year there's lots of new developments to report at OSCON. You can feel it. OSBC has lots of suits in attendance and they tend to be a lot less exciting. What will Redhat report here next year? Another profitable quarter? That's important, but its not exciting. I look forward to seeing the agenda for next year.
05:03 PM | Recommend This | Print This
OSBC2004: Tim O'Reilly on Rethinking the Boundaries of Open Source
Tim O'Reilly is speaking on "Rethinking the Boundaries of Open Source." Tim starts out talking about the "PC paradigm shift" when IBM created commodity computing and did away with proprietary advantage. Perhaps the most important result of this shift was that power was transfered from hardware to software. This birthed Microsoft.
We're in a similar paradigm shift today. Tim illustrates this with his now classic question "how many of you use Linux/how many of you use Google? The point being the everyone who uses Google uses Linux. The desktop is no longer the center of the universe. LAMP is a generic back-end platform that is often used to present a platform-agnostic front-end. These LAMP applications are being created by open source developers and run on open source platforms, but the source code is not distribute (and wouldn't be useful to many developers if it were). Licenses are triggered by distribution. The value in these applications is in their data and their customer interactions, not their software.
Free and open source licensing was invented to keep doing something old, not to do something new. Source code was traditionally open and close-source was new. Stallman invented the GPL as a defensive mechanism to keep the party going. GPL can be problematic because its too restrictive. Less restrictive licenses like Apache and BSD licenses work.
Here are some trends that Tim sees:
- Commodity software with and open architecture
- Information applications are decoupled from both hardware and software
- Competitive advantage and revenue opportunities move "up the stack" to services above the level of the single device.
- Value is based on data and customer relationships, not proprietary software.
- Intel is still inside, but so is Cisco and eventually others--there's plenty of room at the bottom as well as the top.
Collaboration, not licensing is the real mother of open source. One aspect is "the adhocracy," like minded developers can find each other and self-organize in ever-shifting groups that free associate to achieve a goal. A second aspect is that software development teams can be distributed worldwide. Third, power shifts from companies to individuals because of the increased visibility of the network---everyone's a free agent. Fourth, users help build the application.
The first open source revolution happened in the 90's. The commercial Internet is a direct result of open source and even public domain software including UUCP, Usenet News, SLIP, DNS, Bind, Sendmail, Apache, and the WWW. The whole ISP business is essentially paid access to what used to be free. Right now, we have basic cable---the commodity components. Where are the premium channels?
Tim talks about the architecture and participation and Clay Shirky's three methods of building large collected works: (1) pay for it (Yahoo!), (2) use volunteers (Wikipedia), or (3) let users build it as a by-product of their own self-interested actions (Napster). This last method is used by eBay, Google, and Amazon to add value to their services.
Engineering reliable systems from independently developed components may be THE key open source business competency. Websphere is an example of a mix of proprietary and open source software. This is analogous to Compaq in the early days of the PC revolution. Tim asks "who's the Dell of the commodity software market?
03:46 PM | Recommend This | Print This
OSBC2004: Jason Matusow on Shared Source at Microsoft
Jason Matusow, Microsoft's manager of its shared source initiative, is speaking on shared software (his term) and Microsoft. He breaks the topic into "community," "competition," "functionality," and "intellectual property." There's commercial and non-commercial software and open source software fits into both camps
Companies have business goals and when those goals can be met with open source projects, open source projects will be applied. He cites Eclipse as an example. IBM gave it away to reduce expenses and gain mind share for a Websphere development environment. He notes that and eclipse is a total blocking out of the Sun.
How can Microsoft learn from OSS? The strategic bet Microsoft has made has been around innovation in integrated solutions. This has served Microsoft well. There is a cost to modularity. Modularity leads to flexibility, but that drives integration cost to some other part of the equation. He disclaims the nefarious intent of commercial software companies to get "customer lock-in."
The benefit of source code transparency to a business is not necessarily technical for most businesses. Fewer than 5% of them ever look at the code. The benefit is one of trust.
He defines a spectrum of Microsoft's openness. Windows will be closed with only specific grants to certain organizations to see the code. ASP.NET, Windows CE, Visual Studio.net, .NET and others will have derivative rights. Microsoft research will provide works to the public domain. Eleven of fourteen Microsoft programs provide derivative rights. He gives several examples of Microsoft sharing source code with companies, individuals, and universities.
In the end, the talk is mostly what we've heard from Microsoft before: We're not evil. We're playing ball with people. We think open source has some value. But we're not going to eviscerate our business. One can hardly blame them for that.
10:40 AM | Recommend This | Print This
OSBC2004: Ray Lane on the Macro-Economic Impact of Open Source Software
Ray Lane, a general partner at Kleiner-Perkins is discussing the macro-economic impact of open source software. He talks about the parade of people who come into his office everyday who talk about proprietary software because its (their words) higher quality and better able to differentiate. Ray believes that we need to change how we do things and become more modular and open. He thinks that recovery in the software market requires a sea change in development methodologies.
Software is still a buyer's marketplace. Customers all line up at the end of the quarter and beat the vendors up on price. This doesn't feel like recovery.
What's the next big thing? That's the number one question Ray is asked. Its idiomatic of our industry that we expect new things to come along every so often and shock us out of "the old thing." Often we don't recognize the "big thing" like client-server or PCs until later.
Ray does not believe Open Source Software and Web Services are tectonic events on the scale of the transistor or operating systems. Rather they are tactical events that are still important. VC's looks to invest in tectonic events. Ray contends that we need to learn to invest in more tactical changes as well.
IT along doesn't change the equation. Saber (the airline reservation system) didn't save TWA. IT enables management to execute their business processes. He references Carr's "IT Doesn't Matter" article and concludes that IT matters if it helps management.
The nirvana of the 90's has melted away, but even with that, the forecasts of the 90's around the Internet bubble are coming true. Witness Amazon, eBay, Google, and others.
Global peace was shattered by 9/11. Market economies have been beat-up by scandal. Global growth has stagnated. Technology has more capability than anyone's ability to use it. "It's like the Concord." Companies spent countless hours trying to pick "the right ERP system" for example, even though it doesn't really matter. Throw a dart and any one of the systems will be good enough. The problem is that we then "regress" the new system to match what we've done before and make it inflexible and hard to use.
Ray seems some constraints that the software industry needs to pay attention to.
- Volatility. We cannot resolve the assymetric threats that terrorism brings to our world.
- Our businesses are transparent. Why would anyone want to be CEO of a public company? There's Sarbanes-Oxley, but also the transparency imposed by new business models whereby customers and partners have to be part of the equation.
- Low growth. Focus on profitability, not growth.
There are also significant enablers:
- Knowledge worker globalization. No matter what your politics are, you have to decide where you stand on free markets and globalization. India and China have focused on Math and Science over the last 25 years and we can either use them of someone else will. If we don't, we'll be France.
- Network based service providers. ASPs were a terrific business model, but VCs weren't willing to invest capital in it. In the service business, you can't have your customers beta test it (unlike software). Customers expect service providers to be reliable from day one. He references salesforce.com as an example of a model that will become very popular.
- Rationalize IT infrastructure. Many CEOs feel like they didn't get a return on their investment in IT in the 90s. Opportunities exist for software companies to provide ways for companies to exploit the IT infrastructure they now have.
- Portable computing. Laptops, PDAs, wireless change how people work.
The new enterprise needs to be able to grow or shrink to meet demand. They must recognize that investments in fixed assets and labor is bad. The technological progress of the software industry is its vulnerability. The software industry is too big and too expensive relative to the value it generates. Its too labor intensive.
In the 90's engineers were scarce. Now managers are scarce. In the housing industry, there are two industries: building and renovation. We must learn to concentrate on "renovation" instead of "invention." Companies feel like they've spent enough on IT. They might be convinced to spend more if you can make what they already have work better. Think "op ex" instead of "cap ex." Think "service centric" instead of "component centric."
Ray says the service-oriented architectures and "federated data" is real. The ERP era created too many databases and they have to be rationalized.
Ray went out and talked to 30-40 CIOs of Fortune 100 companies. They liked open source software because of lower TCO, equivalent functionality, and quick bug fixes (I missed a few). They didn't like informal support, the velocity of change, no roadmap, functional gaps, licensing FUD, and lack of ISV endorsements.
We're moving into the "golden age of computing" where everyone in the world will participate in the computing industry as a user or developer. Every industry goes through this stage. Open source software is an important part of that. Software is a service and a service company has different DNA than a traditional software company. There's nothing so pure as a new software company. Yet, there's nothing as scary as a mature software company. A handful of people make a software company work and the rest are providing service. We need to recognize that.
09:08 AM | Recommend This | Print This
OSBC2004: Chris Stone
I missed Chris Stone's keynote yesterday, but c|net News has a report of what he said.