Gallery

Has Open-Source sold out?

Posted on 3 Sep 2008 at 12:16

The model highlights the fact that open-source companies can be identical to proprietary software companies except in one respect: they are simply taking a bet that the efforts of their self-selected unpaid worker bees will produce better results in the long run than carefully selected well-paid worker bees. This is a matter of faith for open-source believers, but is it right? There are still questions that the open-source approach needs to answer.

First, can open source keep up with commercial software? Nobody doubts the quality of Linux, but it has taken the best part of two decades to get there. Ubuntu - backed by multimillionaire Mark Shuttleworth - has come a long way as a desktop Linux, but it still lacks the polish of Mac OS X, Windows Vista or even XP. Firefox raced ahead of Internet Explorer, but that was because Microsoft took five years off. OpenOffice still seems inferior to Office 2003, and is nowhere near Office 2007. GIMP isn't a patch on Adobe Photoshop for usability. The idea that open source always delivers better software doesn't stand up.

Second, can open source do the boring stuff that nobody wants to do? Commercial companies such as Microsoft spend a lot of time on features that provide accessibility for disadvantaged users, for example, because they're required for government contracts. How many open-source programmers are enthused about meeting Section 508 of the US Disabilities Act? Sun had to assign paid programmers to the task. The Gnome Foundation, with sponsorship from Google, Mozilla, Novell and others, has offered $50,000 to individuals via its "accessibility outreach program". There are ways round problems like this, but they shouldn't be problems in the first place.

Third, can open source deliver breakthrough innovations? So far, a lot of open-source projects have duplicated existing commercial software. Can they deliver the sort of innovative functionality that requires expensive usability research and testing? Few open-source projects have the knowledge, the staff or the funds to do the job. It's true that the two-guys-in-a-garage model sometimes works, but it's not clear that it's effective at an industrial scale.

And finally, are there enough open-source programmers to go around? Linux has thousands of contributors, but once you get past the big names, most open-source projects have few or none. In his paper, Cave or Community?, Sandeep Krishnamurthy studied the 100 most active mature projects of the thousands on SourceForge, and found the average number of developers per project was four. Only 19% had more than ten developers, and 22% had only one. If the software industry is going open source, where are the programmers going to come from? Unless you pay them, of course. That works.

This doesn't mean open source has failed, or will go away. It's here to stay, and will probably lead to more reliable - though maybe not more imaginative - software. The signs are that open source is being absorbed into the mainstream as just another way of doing business. If you were hoping for an apocalypse, it could be a long wait. Sorry.

Author: Jack Schofield