> Open-source forever
> Transparent code, permissive license, and a community-driven roadmap.
Which I was going to mention is contradictory, because the point of permissive licenses is that it does not have to be Free forever. But the license is actually GPLv3 instead. So still contradictory wording, but the "permissive" is the part that isn't correct :-)
No, the point of permissive licenses is that third-party derivatives, which have no impact on the licensing of the original, don't have to be free ever, while the point of copyleft licenses is that they do.
Neither has any effect whatsoever on what future first-party licensing can be; a commitment to "open source forever" by the copyright owner is mostly orthogonal to what kind of open source license the copyright owner offers. (Now, its true that if the owner accepted contributions under a copyright license rather than under a CLA, they would likely have no practical choice but copyleft now and forever, but that's an issue of the license they accept on what they can offer, not an effect of what they offer itself.)
(OTOH, using "permissive" for GPLv3, a copyleft license, is actually contradictory, as you correctly note.)
Only the code that is yet to be released is not.
From an end user perspective BSD/MIT and GPLvX licences offer the same guarantees. It only is different if you are a contributor or if you intend to distribute modified or unmodified code yourself.
> No, the point of permissive licenses is that third-party derivatives, which have no impact on the licensing of the original, don't have to be free ever, while the point of copyleft licenses is that they do.
This should read "Yes, [...]".
The point of permissive licenses is for people to sublicense it. You can use this to sublicense the software using a license that actually enforces it must remain Free (see Redict for an example) but it is almost always the case that someone uses this headlining feature of permissive licenses to lock the code up and extract rent.
Your next paragraph frames single contributor or CLA as the two primary development patterns, when that those are absolutely the exception, especially if we exclude repos for things like AoC our homework, which are mostly single contributor.
It's guaranteed to have at least the terms of the permissive license (usually requiring attribution), but no, it does not guarantee code released under a permissive license will remain available under permissive terms. That is literally the point of the permissive terms: so people can apply more terms under a sublicense.
> From an end user perspective BSD/MIT and GPLvX licences offer the same guarantees.
No they don't. I can decide to stop distributing a BSD/MIT licensed application in both source and binary form, in favor of only distributing it in binary form under a sublicense. As a user, this is not "open source forever". This is "open source until we use the distinguishing feature of the license to make it not open source".
GPL, assuming multiple contributors and no CLA (both of which are extremely common) ensures this cannot legally happen (unless they somehow get all contributors to agree against their exercised rights).