Major legal victory for open source in US (Java Model Railroad Interface case)
h-online.com
h-online.com
Of course the important precedent isn't the final settlement. At this point that was a foregone conclusion. It is the US Court of Appeals for the Federal Circuit a year and a half ago. See http://www.cafc.uscourts.gov/opinions/08-1001.pdf for the details of that settlement.
This precedent was doubly important given how horrible the first ruling in this case was.
The GPL says that derived works can only be distributed under the GPL. The restriction is on distribution, not license. If you accidentally (or whatever) write something like a gdb port to a new language and try to sell it under a proprietary license, the FSF can make you stop selling it under that license.
But that's it. Your new code doesn't "suddenly become GPL and public". It's your code, and you get to decide the license, even if you can't legally give it to anyone else. I guess if you want to pick a metaphor, it's been quarantined. It hasn't been "infected".
No court is every going to issue a decision that says "The GPL is OK by me!" any more than it is one about the Windows 7 EULA being cool beans. License text is a private affair, courts only step in when they are violated, and only void them when they're actually illegal. This court didn't. That's as much as the FSF could ask for.
So while this case doesn't specifically say "The GPL is valid", it does say that violating a copyleft license can put you on the hook for serious monetary penalties. This is crucial, because those give GPL compliance enforcers another major weapon in their arsenal (in addition to injuctions) to bring companies into line.
Yet another advanced peripheral system, Train CNC!
Simile: "It's like CNC for trains!"
Literal truth: "It's called DCC actually"
If your point was to add information, then there was a better way to do it.