It's pretty disparaging that there are people on HN who feel confident enough to enter into discussions about software licencing —albeit on a throwaway account— but don't understand that software can have more than one license.
I'd like to think I do understand all sides of this, but I will admit AGPL does give me pause for thought. It makes sense for standalone projects which don't have to be integrated. But at component and library level, when there's so much MIT/BSD/etc on npm/pypi/etc, I don't have to think about it.
Does make me wonder why there isn't an ALGPL: "Use my network-expressed component as you like, but if you alter its source, you also need to redistribute the changes under the same license"…
> Does make me wonder why there isn't an ALGPL: "Use my network-expressed component as you like, but if you alter its source, you also need to redistribute the changes under the same license"…
I'm fairly sure there's a GPL like that, either the GPLv3 or a version of the AGPL. It definitely doesn't let you alter the source for network servers, though, exactly like you say.
"It has one added requirement: if you run a modified program on a server and let other users communicate with it there, your server must also allow them to download the source code corresponding to the modified version running there."
Notably, the Canvas learning management system is licensed this way, and all of their JavaScript is unminified so you can download it from the server.
I'd like to think I do understand all sides of this, but I will admit AGPL does give me pause for thought. It makes sense for standalone projects which don't have to be integrated. But at component and library level, when there's so much MIT/BSD/etc on npm/pypi/etc, I don't have to think about it.
Does make me wonder why there isn't an ALGPL: "Use my network-expressed component as you like, but if you alter its source, you also need to redistribute the changes under the same license"…