In the comments section following Jim DeLong’s response to my open source article, programmer Virginia Warren points out that some of DeLong’s characterization of the “General Public License” under which open source software is released is misleading. She writes:

[The GPL] does *not* say, as Mr. DeLong claims, that software licensed under the GPL cannot be sold. What the GPL in fact requires is that the source code of any software licensed under the GPL must be freely distributed.

It also requires that software which incorporates GPLed code also be licensed under the GPL, i.e. if Alice wrote a software package, licensed it under the GPL, and Bob then took the source code from Alice’s GPLed software, made some changes and additions, in effect creating a new software package, Bob’s package would have to be GPLed as well, since he incorporated Alice’s GPLed source code in his software package. However, it’s better than that. If Bob disentangles his new code from Alice’s such that it is distributed completely separate from Alice’s code, he does not have to GPL it, even if it requires Alice’s code to run. So, we see that the GPL does limit software makers in one way–it prevents them from taking someone else’s work and passing it off as their own!

This is the alleged ‘viral’ nature of the GPL, and on this feeble basis, many writers have made the mendacious claim that any software written to run on Linux must be licensed under the GPL. The suggestion is that a software developer can somehow ‘catch’ the GPL, just by using or writing software for Linux, when in fact one must actively and intentionally incorporate pre-existing GPLed code in one’s software in order to ‘catch’ it.