As chief technology officer of San Francisco start-up Laszlo Systems, Temkin is selling the idea that the long-predicted era has arrived in which software applications will live on Web sites rather than desktops.
Now Temkin is betting that a recent open-source gamble will propel his company's vision of Internet applications ahead of its formidable competition.
With the October release of Laszlo's core technology to an open-source development group it created, the company is trying to take on an established company, Macromedia, which is, in turn, going after the fiercest software adversary of them all--Microsoft.
Microsoft's XAML language is designed to build Windows applications tied closely to the Internet. While some companies arethat run with Microsoft's .Net client, the proprietary XML dialect is meant for use with Microsoft's ambitious Longhorn operating system. While Longhorn has suffered chronic delays and , few doubt its impact will be significant once released.
More immediately, Macromedia is busy promoting its Flex server software, which runs Internet applications written in Macromedia's MXML language. Macromedia recently announced aand .
Temkin recently spoke to CNET News.com about the company's view of the computing present and future, and why his sights are set on besting Macromedia.If Web applications are here already, what do we need Laszlo for?
There is still a serious gap in what Web applications provide. Today's Web applications are based on the page-by-page HTML model--which was never designed to support applications--and what the end user sees on the vast majority of Web sites is a series of flat pages, connected by links. This is an awkward way to drive an application.
LZX is Laszlo's client-independent XML description language for rich-client applications. The Laszlo Presentation Server compiles LZX into SWF (Macromedia's Flash format), which runs in any browser on any platform that has Flash 5 or later installed. In the future, LZX will be compiled into client formats other than SWF, such as Java and the .Net client.
MXML is Macromedia's Flash-based XML description language for rich-client applications--this is the core of Flex. The Flex server compiles MXML into SWF, which runs in any browser on any platform that has Flash 7 installed. And XAML is Microsoft's Longhorn-based description language for rich-client applications. Unlike LZX or MXML, these applications are Windows-based, rather than Web-based, and more deeply integrated with the underlying OS and hardware.What does putting your presentation server into open-source development do to help you compete? How did you distribute your software before?
We have open-sourced our whole platform. There's no poison pill--no "if you're commercial, you pay us"--anyone can use it.
Our software was licensed the way application servers are: on a per-server basis. The more you serve, the more concurrent users, the more you need to pay. That's the way Flex is licensed. But now that the source code is open and available, we're working on establishing alliances to get this platform, especially the LZX language, to be a standard. That benefits us and the industry because this is the open Web. It's open source. It runs everywhere and runs in any browser. XAML needs IE and .Net.
How have developers reacted to your open source offer?
There's been a huge, huge response. We did as many downloads and developer registrations the first three days (after last month's release) as we had in the previous six months, roughly 3,000 downloads of the Laszlo Presentation Server. We're also talking to industry heavyweights who have strategic interests in having an open-source, rich-client platform that isn't proprietary and controlled by a single vendor, a zero-install rich Web experience on an open-source platform. And we expect those companies to contribute.
Who are some of your partners?
There's Akamai. We're working closely with them. And we're working closely with others in the industry who are interested in seeing an open-source, rich-client standard. We can't say who right now, but there are other companies in the industry who have an interest in having a standard of this kind.
We'll announce who it is when we're ready to announce who it is. But we are talking to a lot of companies of that type: large Web application providers and some infrastructure providers.
How seriously do you think Macromedia is taking your open-source offering?
Macromedia has already responded to this. On the very day we announced, Macromedia said that Flex now has a free noncommercial license that will be available in November. Interpret it as you will. But they're responding to this on a marketing basis. If I'm an application developer and I have a choice between something that is closed and proprietary, that costs $12,000 for two CPUs, and is linked to the Flash 7 client, or I can use an open-source product that offers me wider deployment options because I can use Flash 5 and 6 and it's unencumbered, what am I going to choose?
(Macromedia responded that Flash 7 distribution has reached 76 percent of PCs. "The advantage of running on Flash Player 6 is fleeting at best," a company representative said.)Do you really think Macromedia made the free Flex announcement in response to your open-source giveaway?
When we announced this, Macromedia had their talking points out there and said this is the last act of a dying company. They didn't say it in public, but it was a very consistent message that we got from more than once source. Our funding announcement refutes that, of course. (Laszlo has received a $5 million Series B funding round led by Mitsui Venture Partners.)
The Laszlo Presentation Server includes a number of open-source packages. We use the Apache HTTP client and Tomcat (an open-source servlet container), and we bundle that with our product.
The obvious question is that with your flagship software now available open source and free of charge, how are you going to make money?
For those who have downloaded the open-source client, we offer support and training. Second, we're going to do custom application development. Customers will come to us and say, "I love those apps. I want your design talent applied to my problem." Third, we will offer commercial applications on top of this open-source platform. We will have a shopping cart, for example, or a Web e-mail application built on top of this. That's over the course of 2005.
We've been around longer than Flex. We're deployed on more prominent services, including EarthLink, SBC Yahoo's broadband service, and the La Quinta home page. Some of these sites have millions of users. This is a very stable and robust platform.
So look at the alternatives: Laszlo, a small start-up--and yes, people get uneasy about that--or Macromedia. With Macromedia, the question isn't if the company goes away, but what if it stops supporting Flex? Macromedia has a track record of canceling server products, a whole raft of them--Generator, Aria, LikeMinds, Spectra, SiteSpring. In the commercial world, the size of a company doesn't guarantee the stability of a platform. Macromedia's solvency isn't what determines what's supported in 10 years. Open-source software gets immortality, in a way, if it ever sees adoption. My belief is that you can't get a new platform adopted unless you're open source or unless you're Microsoft.
Take, as an example, JBoss, PHP, Python, Perl, Eclipse to a certain extent--these have been massively adopted and they're all open source. I can't think of a broadly adopted proprietary platform that's been widely adopted in the last five years. .Net is the exception that comes to mind--and that's Microsoft.
Apart from being comparatively insulated from the strategic vagaries of private companies, what else makes open source more attractive to developers?
Proprietary platforms are just not how developers think today. High-priced proprietary platforms stand in the way of casual development experimentation and interest and make it much harder to justify the purchase of one of these systems. With open source, you can--and do--have grassroots deployments within the enterprises. With proprietary platforms, you have enterprise sales driving top-down adoption. Laszlo has lived through this. We know what it's like trying to get a proprietary platform adopted--not without some success. But it's a tough battle.