My personal journey from MIT to GPL (2019)
drewdevault.com
drewdevault.com
Companies have two choices : avoid GPL like pest or embrace it.
We (I, mostly) contribute back to open source at work. The only problem that is not solved yet is should we use GPL-3 code, because we sign software and there is the Tivoïzation clause.
I still continue to release my (Rust) libraries with Apache-2.0/MIT. Because of static linking almost imposed by the current Rust development workflow. LGPL would contaminate the whole software if statically linked. Or make people link to it dynamically. Maybe this is not my problem but the companies'. Should we push for GPL libraries?
MIT style licenses are about freedom for the programmer to do as they want, not code freedom. GPL is all about the freedom of the code, at the cost of freedom for the programmer to do as they want.
I've gone the same route myself, from BSD style on earlier things to GPL on most things I do now. I'm still looking for a good license to apply to hardware projects though.
This is the fundamental idea behind copyleft that so many people seem to misunderstand whenever they say that MIT/ISC licenses are "more free" because they come with fewer obligations: copyleft licenses like the AGPL force a vendor to give up some power to improve freedoms for the wider community. Compliance is only hard if you want to control others' computing; the whole point of copyleft licenses is to oppose that.