Where Openness Prevails
Here, in Hopkins's Center for Networking and Distributed Systems (CNDS), the spirit of openness born in the early days of computer science still prevails. Yet each new technological tool created here also offers a choice for faculty and student researchers. Do they:
A) seek a patent on the invention to better pave the way for a start-up company? or
B) give the technology away to the great, unwashed computer programming masses and hope it becomes a wildly popular standard that brings the inventors prestige and lucrative consulting contracts?
Often, that choice is B: open source.
As patents have become more common in computer science, open source has been proffered as a nonproprietary means of superior software development. The term applies to free software in which the source code, the "recipe" for a program, is distributed along with the program; an attached license says the software can be used, modified, and even commercialized by those who agree to keep the source code available.
To many, the advantages are clear: millions of people can use and find the bugs in a new product, such as the Web server program Apache, and programmers can better analyze the technology and build on it.
"Computer science depends to a large extent on the network effect," says Darren Lacey, Hopkins information technology licensing associate. "There are lots of ways to become the standard. You can exercise monopoly control, like Microsoft. But if you don't have a monopoly, you need to have the widest implementation possible. That runs counter to patents."
In recent years, open source has taken on the mantle of a philosophy, especially among lone programmers or software start-ups working outside the halls of Microsoft and other mega firms. The increasingly popular operating system Linux, which can be downloaded for free off the Internet, has become the movement's poster child. But many programmers are still figuring out which way to go.
"The whole question of intellectual property rights is becoming hopelessly complicated," says Jonathan Shapiro, Hopkins assistant professor of computer science, who has 15 years' experience in the industry, including as co-founder of HaL Computer Systems, a California software and hardware firm.
Software is often copyrighted, which requires royalties for use but still permits competitors to learn how a program operates, and to reproduce it, if they don't use the original source code. Patents, applied to software only in the past two decades, are more restrictive, providing greater protection.
Yet the software industry is grappling with the patent system's rule of exclusivity, which lasts 20 years from the time a patent application is filed. That's tantamount to forever in an industry in which products have a 12-to-18-month life cycle. Notes Shapiro: "This hampers commercial innovation, not just research." He believes a two-year patent period in software, and even some hardware, might be more logical. In addition, some patents have drawn controversy for being too broad, such as Amazon.com's patented one-click online shopping method.
Just how the open source ideal can be economically feasible remains to be seen. Giving away software for "free," advocates argue, can still create income streams from technical support contracts, as well as increased demand for related products, such as hardware, or kits that make the free software easier to use. (Take Red Hat, a company that packages and sells Linux, with a manual, for about $50.) There's also a variety of hybrid license deals, including those that allow widespread use by anyone, even companies, at no cost, yet require royalties for any resale of the product.
The open source approach can be fraught with legal questions over the terms of various licenses, and other business uncertainties. But in the recasting of rules for the new lab-to-market economy at universities, some researchers say such counterintuitive approaches are just what should be given a second look.
"The real question is: How are we going to be successful?" says Yair Amir, associate professor of computer science in Hopkins's CNDS lab, which has developed tools to help Internet Web servers and other networks communicate faster, or with fewer glitches (see the Web site: www.cnds.jhu.edu). Says Amir: "Open source, or not open source. Patent or no patent, to me these are all just vehicles." -- JCS
RETURN TO FEBRUARY 2001 TABLE OF CONTENTS.