Whether or not your Component extends or has GPLed classes injected into it or instantiates GPL classes before using the methods only adds to use of GPLed benefit you have gained. Yes, those are ways of invoking the viral effect of the GPL.
Outside of a US court, I suspect that argument will fail (not that I can't see the logic). As an example (albeit in a different arena). I write a proprietary FTP client for GNU/Linux, when listing the local directory it calls stat, which passes the data back having gone through the appropriate filesystem module.
Now, I've derived benefit from GPL code (actually, technically I think the stat call in Linux might originally have been BSD) - the kernel and the filesystem. So I've derived benefit from GPL licensed code.
It's a bad analogy for quite a number of reasons, but you see what I'm getting at. Assuming anyone was stupid enough to go to court rather than comply, it's the kind of response I'd expect to see to an argument of that nature.
I realize you are not (quite) pro LGPL, but you are sensible and I suspect you'll get there when you realize there is not a snow ball's chance in hell PHP developers will use the library under that license and there are TONS of other (arguably better at this point) options they can choose from.
My stance is more philosophical - a massive proportion of the code I've written won't ever see the light of day (in the OSS sense), and offsetting that is a good thing.
I do agree that LGPL would make the framework more attractive to developers, and it's quite possible we'd see some additional contributions coming back. I just don't know if it's worth dropping an ideal that we as a community have espoused for so long (although LGPL was debated quite a while back, 99% of users will only have ever heard the GPL message).
I do think, though, that communication is absolutely key
. There's already uncertainty in the air (which license will we used), added to by the question of whether any devs will walk (I agree with you btw, that's a personal decision for each dev). Whichever way the vote swings, things need to be well communicated.
I've seen a lot of posts on the benefits and drawbacks, but there hasn't been any real sign of forward planning/analysis (from a communication perspective). As an example:
If it goes LGPL, you and I both know that it's likely to take some time for developers to start using Framework, so it's going to take some time to see any contributions coming back. We've got a lot of catching up to do, in mindshare and (to some extent) in functionality, which will only extend the lead time.
Without decent communication, the end result of that is going to be rants about how the license has changed and there's been no benefit.
The biggest risk, at the moment, in my eyes is that of the sense of community being hurt, which is only going to lead to a divide if people start thinking 'told you so'.
The point I'm trying to make is - As someone who's almost constantly in trouble for opening his mouth, for something like this, communication should not
be left to developers. Traditionally, good communication (and that doesn't include talking to other devs) is not a strong point, usually leading to hurt sensibilities and a breakdown of communication.
Earlier in the thread, I called out Don Gilbert for the way he'd phrased/stated things. In fairness, it was a bit rich because I quite regularly come across in much the same way (though I'm getting much better
The devs will have their opinions, and I'm not suggesting for a second that they shouldn't be allowed to express them, but the issue of licensing isn't actually a development question (though obviously it does have an impact).
I know you've feel there's a strong case for LGPL, given that this thread is all-but unreadable to the average user, perhaps a blog post espousing the benefits (and analysing the drawbacks) would be a good way to go - it's a pity the debate never happened as it would have helped to the same thing.
Niv in my opinion this isn't about attracting the proprietary developers. It's about opening the door to the PHP community of developers who don't use GPL (but use MIT or similar) - these aren't the proprietary developers but people like us, who, for whatever reason have decided GPL isn't for them.
I agree that the (ideal) target audience is the non-proprietary but no-gpl developers, but the counter argument is that the LGPL route opens the door to both. Only time would/will tell which make the most use