Two different "risks" here

Two different "risks" here

But guess what? The team of people who were _threatening_ to sue Cisco over the GPL back in 2003? They didn't actually file suit:

http://www.forbes.com/2003/10/14/cz_dl_1014linksys.html

This is factually wrong. They actually did that.

All the work was done _out_ of court, and Cisco cooperated to _keep_ themselves out of court.

As you describe in other place this "cooperation" was in form "let's sweep it under the rug, continue with sweet talk and hope FSF will eventually go away". It didn't work: FSF is extremely patient, but this is more then compensated by it's perseverance.

The threat of a GPL lawsuit was just saber rattling to get Cisco's attention, they negotiated the release of source code _without_ever_filing_suit_.

Yes. Once. And it was incomplete. And then Cisco issued new versions of both software and hardware - again with GPL violation. And as your admit when they started new "proper" effort they had no plans to do anything with old violations.

All the open drivers for chips from nvidia and ati and broadcom and such we've gotten since then happened either because a company _chose_ to be nice (like Intel, hiring the xfree86 guys to push back against microsoft a bit), or because we _reverse_engineered_them_.

Yes, but can you explain just why companies suddenly decided to be nice when they dragged their feet for years? And why are you so sure lawsuits have nothing to do with that?

All this talk about lawsuits being necessary for hardware support does not match reality, and certainly was not the goal of the guy who started the suits (I.E. me).

You may have started it, but your certainly don't have power to stop it. Simply because you don't own rights for a lot of GPLed code.

I have first person experience here: you do not.

As Kleiner said: It's difficult to see the picture when you're inside the frame. You can only talk about direct consequences of all that activity, but how can you claim that Broadcom's change of heart is not related to SFLC and SFC efforts?

You are correct when you say Never think you're _irreplacable_. Nobody is. That's certainly true. And if your goal is to make software used by millions then these lawsuits may hurt. But Stallman's initial goal was completely different: the goal was to be able to hack on the software you "own", not to spread your own software as an act of goodwill. GPL was not a goal in itself, but means to this end: if someone uses your code then GPL guarantees that you get your own code back with appropriate changes! Clever hack - but it was broken by tivoization and by blatant ignorance of GPL conditions. Solution to tivoization is obvious: GPLv3. It makes sure that manufacturer can not forbid you to modify your own code. Solution to GPL violations are lawsuits. If manufacturer does not give you enough data to start hacking on your code then what's the point of the whole exercise?

From this POV lawsuits are quite successful: a lot of companies published enough sources to make the whole classes of devices hackable. The fact that GPL stopped being the most popular license in the world is unfortunate side-effect, but alternative (GPL continues to be most popular license while manufacturers effectively treat it as BSD license) was clearly worse from FSF's POV.

P.S. It's funny that you argue are so adamantly against SFC's efforts where the only "proven disaster" happened because of FSF. FSF certainly started it's efforts way, way before your involvement. In fact G++ story happened before busybox project was even started!