Reposting my thoughts on bounties from a reddit thread a while back:
Well. Here is the problem, and why I think that bounties just don’t work as a concept.
Let’s pick an item from bountysource at random. I chose https://www.bountysource.com/issues/3933006-massive-memory-leak-in-unity-panel-service-and-hud-service-when-invoking-the-hud-on-firefox-profiles-with-large-amounts-of-bookmarks-lts-12-04-14-04. There’s obviously some sort of bug in the HUD when Firefox has lots of bookmarks. That’s going to involve a developer setting up a Firefox profile with a load of bookmarks, building a debug version of the HUD, running Firefox, seeing where the HUD is incorrectly using lots of memory, working out how to fix that, fixing it, submitting a change back to Ubuntu, getting that change reviewed, possibly fixing any review comments, and then seeing that the new code gets packaged and fixes the problem. I don’t know for sure, but that’s gotta be a minimum of, say, four hours work. It might take longer (it might take a lot longer) but let’s imagine it takes four hours.
Average salary for a UK software developer is £30,109, but let’s take the lower limit of what they say there, which is £20,639, or 20639/52/40 ~= £10/hour. So that’s £40, or sixty US dollars.
What’s the bounty? Ten dollars.
(note: it was at time of writing. It got jacked up as a result of this discussion! This is obviously nice but doesn’t work as a general principle!)
So it’s asking someone to give up a minimum of half a day and basically not get paid for it, or get paid a sixth of their normal rate for it. And that’s if it only takes half a day, and only if you’re currently at the absolute rock bottom of the pay scale.
The usual objection here is that bounties aren’t there to incentivise a developer to do a thing that they don’t want to do (which is what salary is for); it’s there so when you fix a bug that you were going to fix anyway, you get a nice little ten buck reward for doing it. Or possibly that if you decide you’re going to fix a bug today, that you’ll look at the list of bugs that you might fix and pick the one which has a bounty, because, hey, free ten dollars! That’s OK; it’s a reasonable argument. But that argument means that bounties themselves don’t incentivise anyone to fix a bug… and therefore there is little to no point in hassling people to add more bounty money, because it doesn’t actually get more bugs fixed.
There is another argument which is rarely explicitly stated, but goes something like: well, this is free software. It doesn’t cost anything. So it shouldn’t need to be paid for at market rates; paying someone twenty dollars for a week’s work is perfectly reasonable, because freedom. Screw that argument.
Anyway, that’s why I don’t like bounties. If you want something fixed and you’re willing to pay for it, hire someone to fix it; that’ll get it fixed, and you’ll pay market rate for their time. If you want something fixed and don’t want to pay for it, then I really don’t believe that waving the price of a coffee at someone will help, in the same way that demanding that it be fixed because you need it won’t help. Instead, give a developer an actual incentive to fix the bug you care about – maybe offer to fix a bug that they care about in return, or offer to contribute to their project by making a tutorial video or writing some new user documentation or knitting them a scarf, or ask them what you can do in return so that they make your thing a priority instead of what they planned to do.