The majority of American computer programmers--that is, those whose jobs have not been shipped off to India or Romania--don't work at famous technology companies like Microsoft or Google. They toil away inside the big corporate data centers that are run by banks, insurance companies and other mainstays of the economy. These programmers spend their days writing and then tending the software that their employers use to keep their businesses up and running: tracking payments, counting inventories and all the rest.
But since one bank accounting program is much like another, that means that at any given time computer programmers at one bank are duplicating the work of programmers at another. This is true even though much bank software is purchased off the shelf from vendors like
The Eclipse software consortium (www.eclipse.org) is an attempt to do just this. So far the experiment appears to be a success. Eclipse started life at IBM in 2001 as a development effort involving the Java programming language. At the time IBM was walking a fine line with Java. The language was being pushed by
Sun is still closely associated with Java; the company switched its Nasdaq ticker symbol to "java" in 2007 in a so far failed attempt to rekindle the magic of its dot-com heyday. While the language never came close to becoming the one-size-fits-all blockbuster that Sun predicted, it nonetheless has carved out a role as a reliable element of large corporate software installations.
The initial Eclipse work at IBM on Java involved housekeeping programs used to help write real software. IBM decided the best way to keep the project going was to spin it off as a nonprofit foundation, which it did in 2004. There are now more than 185 corporate members of the Eclipse Foundation, whose budget this year is $5 million. Many members are technology and software companies; others include banks, insurance and telecommunications companies.
Eclipse works the way many consortia do; members collaborate by donating their programmers' time to work on shared projects, with the resulting programs available for anyone to use. Typical output: an all-encompassing program that simplifies the job of writing software for various kinds of mobile phones. While the foundation's software library is free for even nonmembers to take advantage of, companies who pay dues of up to $125,000 a year get to help pick the projects on which Eclipse programmers will work.
Michael Milinkovich, executive director of the foundation, says that when joining the group, members have to rethink many of the ways they see technology. The first assumption that needs to be jettisoned: that companies need their own versions of important programs as a way of having an edge on their competitors. "Every bank has a program they use to monitor consumer risk," he says. "But companies need to stop thinking about every line of code as a corporate asset. A great deal of code is exactly the same."
The alternative, says Milinkovich, is for companies to expend their energy accumulating other kinds of competitive advantage, for instance, the real-world knowledge about customers and marketplaces that gets plugged into a program. The best analogy is with spreadsheet programs: No company insists on writing its own version of Excel; the value is in the data inside a spreadsheet.
When thinking about collaborating on any project, especially software, companies want assurances they won't be sued for copyright or other offenses in connection with whatever programs they end up using. At Eclipse 3 of the 17 full-time staff members look through lines of code to make sure none are someone's private property. So far the consortium has been free of legal hassles.
Eclipse's membership has extended beyond the software companies that were its initial audience. (Microsoft is still not a member, though relations are good, as many Eclipse projects rely on Windows.) Now, says Milinkovich, Eclipse is looking to extend its reach further afield, working with groups for telecom and automotive companies that aim to develop industry-specific programs.
The fact that many of its potential members are struggling in the current economic environment is something of a calling card for Eclipse, says Milinkovich, since the entire point of the organization is to let member companies shrink their programming budgets. "People these days are interested in doing more with less," he says. "A crisis like this is too good an opportunity to waste."
Senior editor Lee Gomes covers technology from our Silicon Valley bureau. Visit him at www.forbes.com/gomes/.
