James DeLong is having a hard time understanding Linus Torvalds' point about the GPL and reciprocity:

I don't write code. So, in Torwalds' view, we cannot trade. But a broader view is that we can trade because reciprocity includes me giving him money, which he can then use to buy something that he cannot make himself.
The market is a giant reciprocity machine. Why is this a hard point?

As Time Lee has pointed out in posts on Tech Liberation, DeLong has a hard time understanding a view of copyright that doesn't involve money or dichotomies between anarchy and total control. Torvalds is not thinking like a businessman here, but rather like a software engineer. He is not interested in simple financial contributions from a variety of sources, but rather technical contributions. I'm sure that he'd be pleased as punch for people to contribute money to a fund to pay core contributors who are not gainfully employed in a way that advances Linux development, but, again, that's not what he was referring to here. What he is interested in was the little matter of getting something for giving something.

I agree with Torvalds that if the GPL is going in the direction of restricting developers, it is not a good license. What DeLong clearly missed was that Torvalds chose the GPL v2 because it best fit his primary goal, which was to force people to give back code if they use his code. The matter of getting paid was less important than making sure than people could not take his code, modified or not, and make money off of it without giving back to his codebase. Clearly, reciprocity between code and money is not what he was referring to at the time. It was reciprocity between developers, not developers and businessmen.

To put it another way, would you want a developer to talk about reciprocating in code when your primary interest is money? For most people, in fact for pretty much everyone, the answer is clearly no.