3 ms·
The argument ignores that GPL is a giant stink eye for many people. It deters people from wanting to make enhancements in the first place... there's no point in
by unconed 14y ago
The argument ignores that GPL is a giant stink eye for many people. It deters people from wanting to make enhancements in the first place... there's no point in having rights to enhancements that are not being made. It doesn't matter if technically they could still use it without violating the license, they simply do not want to.
GPL works great until someone builds a superior non-GPL-licensed alternative. Then it becomes irrelevant.
- gillianseed 14y ago'And permissive licences is a giant stink eye for many people. It deters people wanting to make enhancements in the first place. There's no point with enhancements being made when they are kept under proprietary lock and key.' There, your arguments hold no water. There is no 'right' or 'wrong' licence, there is only the needs of the code author/owner. And if someone builds a 'superior' alternative then it doesn't much matter what licence it uses, the point is it being 'superior' code which in turn is what would make something 'irrelevant'.
- unconed 14y agoIs clang getting more popular because it is superior, or because its license encouraged commercial interests to invest into it? Also, you seem to equate a permissive license with 'lock and key'. My experience is the exact opposite, having released software under both GPL and BSD-style licenses. I've received tons of emails from people asking me to change the license on a piece of GPL code over the years. Nobody's ever asked me to make something GPL that wasn't. It's a pretty good indicator that the GPL prevents people from getting shit done, but that few people other than idealists of the Stallman type really care about it.
- gillianseed 14y agoThere are commercial investments in GCC and has been so for a very long time, we have IBM, Google, Red Hat (of course) and others employing full-time GCC programmers. >Nobody's ever asked me to make something GPL that wasn't. Why would they? If you've licenced it permissively they can include it in their GPL licenced if they so wish. It doesn't affect their decision in licencing their code as GPL. >Also, you seem to equate a permissive license with 'lock and key'. No, I was pointing out that enhancements done to permissive code which is not being open sourced is in no way better than your assumption that if something is GPL then chances are enhancements will never be made. On the other hand, if there are enhancements made (and distributed) then GPL makes sure those enhancements will be open sourced and available to all recipients. >It's a pretty good indicator that the GPL prevents people from getting shit done, but that few people other than idealists of the Stallman type really care about it. I'll counter your anecdotal evidence with the following anecdotal evidence: Linus Torvalds and lots of other Linux kernel developers have stated numerous times that they think GPL(v2) is a 'perfect' licence and they attest alot of Linux success to that very licence choice. Now I don't understand why some people (like you) are so adamant at attacking other licences other than your particular preference. GPL and BSD/MIT etc are great licences which are generally being preferred for different types of code. GPL is mostly used for 'self-contained' projects (typically applications) while permissive licences are mostly used for component/framework/library style code. People _choose_ to licence their code under these licences, it's not unusual that people licence some of their code under copyleft and some other code as permissive. Choice is good.