Build it or License it

Billy Newport complains that he can find code to do what he wants freely and easily on the Internet, but that he doesn’t want to have to go through the hassle of actually getting permission to incorporate it into an IBM product. He goes on to say that open source software doesn’t increase programmer productivity, because he can’t find code that he can use for his employer without any strings attached.

Open source code does increase programmer productivity within the open source community. If you go ahead and write the data structure code for IBM, then IBM owns it, and nobody but IBM reaps the rewards. Viral open source licenses give the open source community the same advantages. When I distribute code licensed under the GPL, the open source community reaps the rewards. It’s a feature, not a bug, that IBM can’t use this code without playing by the same rules. I, for one, would love to see IBM[*] gaining the rewards of participating in the open source community. This doesn’t mean taking what you want from the community without giving anything back (nor does it mean whining about what the community won’t give you for free).

If IBM wants the software, they can either build it or license it. Whining about it doesn’t change those options. The fact that the Internet provides an accessible software marketplace with plenty of code under a variety of licenses doesn’t change those options. You can either build it or license it. The references to Apache and GNU licenses in the original article are really just passive-aggressive FUD. If you find code you want, but you can’t agree to license terms with that code’s owners, then you either keep shopping for code you want to license or decide to build it yourself.

[*] I am aware that IBM does participate in the community. I’m responding specifically to Billy’s post, in which he assumes the role of a “commercial” entity himself, and says the GPL and LGPL are useless.