Curl: Get With the Program

Baldwin Louie — Mon, 06 Aug 2001 15:00:00 -0700

The Software Development Forum audience in the Palo Alto, Calif., Cubberly community center was one tough crowd. Before attempting to sell the hard-core programmers on his groundbreaking Curl software, Brent Young made sure his company's pedigree was high in his PowerPoint display. There on the second slide, listed as a founder, was "Tim Berners-Lee, inventor of World Wide Web."

"HTML is good for what it does. Java is good for what it does," Young declared. "What the Web needs is a merger of the two." Curl, he claimed, will deliver everything these ubiquitous Internet languages offer and more.

The audience, about 60 strong, was skeptical, Berners-Lee or no. "Another bid by MIT's failed linguists to produce yet another language," one programmer, Brian Topping, scoffed.

By the time Young had finished, though, the techies had come around. His demo, a Web application recently deployed by electronics company Siemens, was a graphically rich and interactive financial analysis tool with the fast, smooth performance of a desktop app. At the end of the session, Topping was among the crowd of programmers that mobbed Young, eager to learn more.

Curl's ubergeeks have created a programming language they claim encompasses everything HTML and Java can do, along with a browser plug-in to deliver Web content, à la Macromedia's Flash. Aiming to re-engineer the Web, they face an array of entrenched technologies. But investors have bet $52 million on its potential.

The Curl technology emerged from an MIT research project funded by the Defense Advanced Research Projects Agency, or Darpa, the government agency that gave birth to the Internet. In 1998, with the project completed, several team members launched Curl in Cambridge, Mass., MIT's hometown. Brent's father, Bob Young, whose background is corporate financing, came in as CEO. Twelve MIT-connected founders, including Berners-Lee, put up the initial $1 million.

Curl addresses a core problem of the Net: It's a programming muddle. Basic Web pages are coded in HTML. Interactivity on a Web page - such as forms that visitors fill in or a shopping cart - demands a quantum leap in programming ability, to JavaScript or the equivalent. To create richer content, such as animation, requires Flash or a truly complex language such as Java or C++. Maintaining a Web site that mixes these technologies can be a nightmare.

The wireless Web may be the next frontier to companies, but to consumers it's still a pipe dream. A recent survey from Cahners In-Stat found that a mere 3 percent of respondents had used wireless Internet services.

Aside from technical limitations, a common complaint among consumers and industry analysts alike is that there's not a whole lot to access.

Intava, a startup company whose software allows the "little guys" to build wireless Web sites, hopes to change that. While there are at least a dozen so-called wireless development platforms already, they cost thousands (and sometimes hundreds of thousands) of dollars. Intava's product, dubbed Gravity, will sell for less than $600. "It's been very hard and expensive to build for mobile, so there's not a lot of content out there," says CEO Troy Carroll.

Generating lots of content has worked before. NTT DoCoMo's I-mode service in Japan has over 25 million subscribers accessing the Net from their phones, in part by allowing anyone who knows basic HTML to build an I-mode-compatible wireless Web site. According to NTT DoCoMo, there are 1,480 officially sanctioned sites and over 40,000 independent ones.

So far, thousands of developers have taken Intava's bait, downloading free beta versions of Gravity. The ultimate goal is to achieve the same scale as Macromedia's popular Dreamweaver Web-authoring tool, which sold 1.5 million copies. Now that could lead to some serious surfing.

Curl's language tries to cover the spectrum of needs, from basic Web page content to sophisticated programmable features. Also, the code that powers the processing for a Curl application plugs in to a user's Internet browser. This shifts the computing burden away from the site's servers onto the visitor's PC, saving network bandwidth and costs.

Curl faces the familiar chicken-and-egg problem of new technologies: No user can see a Web page written in Curl without the browser plug-in. But how do you get people to download that plug-in when no Web pages need it? And how do you get developers to work with it when there is no user base?

"It won't fly," says Dave Winer, well-known Internet software guru and CEO of Userland Software. "It's typical an idea like that would come from academics. They rarely have respect for installed bases."

Bob Young is intensely aware of such skepticism. That's why he's initially going after enterprise customers like Siemens. When a Curl application is deployed on a corporate intranet, the IT department will install the plug-in on the requisite desktops and the chicken-and-egg problem disappears. "This way, getting the company off the ground doesn't depend on critical mass," insists Young.

But that doesn't mean Curl dreams small. "Ultimately, we think it will be ubiquitous," says Young. "This could be a billion-dollar business."

Berners-Lee wouldn't agree to an interview concerning Curl. His office cited the neutrality of his position as director of the World Wide Web Consortium, the Internet's "governing body," but did confirm he has an ownership stake and acts as an adviser.

Another respected Internet programmer, Tim Bray, co-author of the original XML specification and CEO of software firm Antarcti.ca, agrees that Curl faces some difficulties. But he points out that Macromedia once jumped the same barrier with Flash. That software now comes bundled with major Web browsers and claims 97 percent Internet penetration.

"If you meet a need," says Bray, "you can get out there awfully fast."

The Industry Standard - Curl: Get With the Program - Comments

Curl: Get With the Program