I write code somebody else wants and get paid for it as my day job. It happens to be open source. Some people write the code they want to write, but keep it closed-source. So I don't think your contrast quite works.
I think some of the "no money in open source software" unease isn't because people would like to get paid to write whatever code they feel like, but a desire to retain the benefits of having a massive amount of open source code out there (less reinvention of the wheel by multiple companies, low-cost low-friction way to bootstrap whatever actually interesting/novel software your company is doing, etc) but put it on a more sustainable footing where money is directed reliably enough at the people keeping it together that we can avoid the xkcd "one person in Nebraska" failure mode.
IMHO the underlying problem is value based pricing. Roughly that means you take how much money your software generates for your clients and try to capture as much of that as you can. That leads to huge incentive for companies to not depend on commercial software since as soon as that happens the vendor will take them to pound town in contract negotiations.
That fear makes it nearly impossible for something like Log4J to charge anything. Even if it's a penny per year per server you don't want to build on it because they can come back next year and make it $10 a year. And what are you going to do about it?
FOSS removes that threat but it also makes the path of least resistance to not pay anything. The ideal solution is something like "You have to pay a little bit but it's guaranteed that it will never be more than a little bit". But I don't see how to do something like that.
> That fear makes it nearly impossible for something like Log4J to charge anything. Even if it's a penny per year per server you don't want to build on it because they can come back next year and make it $10 a year.
I see it more as a function of scarcity. If it was really difficult to write a logging framework, and no one wanted to do it without getting paid for use, then anyone writing a logging framework would release it under a license that requires they get paid for use. But if there is just one logging framework that exists that meets people's needs and is free (as in beer), then you end up with the situation you describe. Then all the other logging frameworks either need to find some sort of big differentiator that is hard to duplicate and that people will pay for, or they just stop charging.
And since we're talking about a logging framework, something that isn't very hard to build yourself if you confine yourself to the likely very small number of features you need... sure, no, of course the idea of paying for one is just silly.
FOSS eliminates more uncertainty than price “supply” uncertainty. Since it’s free, you don’t have to know the “demand” value it provides to make a good purchase. Furthermore, you can always add features you need.
All of this uncertainty is easy to deal with if you have a nice API. Swapping out databases, for example, is (in theory) near-zero cost. If the database vendor tells you they are charging more, it’s somewhat simple to switch (unlike, say, ad-hoc logging).
It is, isn’t it. The article talks about “open source is communism” but not authoritarianism, real communism. Which made me daydream about if the various licenses for FOSS required profit making companies to pay 100$ per year for all you can eat FOSS. And then it got distributed on some usage based basis. Would things be better? Not practical though.
I think some of the "no money in open source software" unease isn't because people would like to get paid to write whatever code they feel like, but a desire to retain the benefits of having a massive amount of open source code out there (less reinvention of the wheel by multiple companies, low-cost low-friction way to bootstrap whatever actually interesting/novel software your company is doing, etc) but put it on a more sustainable footing where money is directed reliably enough at the people keeping it together that we can avoid the xkcd "one person in Nebraska" failure mode.