Peer review and public acessibility makes Open Source superior

Within a day of the September 11 attacks on the World Trade Center civil engineers around the world understood why the towers fell. The analysis was carried out - or at least begun - at Sydney University, on the other side of the world from New York, by engineers with copies of the WTC designs.

It exemplified the civil engineering profession's attitude to intellectual property: designs are copyright - but they're also published.

It's an example pertinent to the debate about Open Source software. Like the office towers we inhabit, software is becoming embedded in our society's infrastructure. Software engineering wishes to become a profession; software engineers as individuals wish, rightly, to make a living.

Tony Healy's error is to assume that publication and profit can never live under the same roof.

The overriding ethic of Open Source software is not its price but its publication. Software can be sold but it must be published: the source code, like the construction drawings of the WTC (or closer to home, the Anzac Bridge) is a public document, available for peer review.

Copyright is used to achieve this: the owner of the work first asserts copyright, and then specifies the conditions under which the copyright work may be used by others. In licenses such as the GPL, these conditions specify the need for users to inherit rights downstream - if I obtain software under the GPL, I must give other users the same rights as I have.

The explicit rights regime of the GPL keeps its outputs in plain view and associates authorship with output.

To many of those in today's software industry this is a threat. Healy's argument is that nobody will pay for something they can get for free and the best engineers will have their work stolen by the worst.

Who Will Buy?

In a sense, Healy is right. If software is "free", why pay for it?

This, however, is a view which trivialises software as an engineering discipline.

In the business software market the software itself is already a shrinking component of the value of a project. Expertise is worth a lot more: the expertise to assemble different pieces of software into something that works smoothly; the expertise to customise the software to particular purpose.

Revaluing the "licence" component of a large software project to zero doesn't get rid of the software engineering needed to build the project.