> Sorry for suggesting a good practice to avoid legal liability.
Why is this good practice? Has anyone ever been exposed to a legal liability that could be creditably attributed to their merely looking at (as opposed to using) such code? (Although my suspicion is that the answer is 'no', this is an honest question.)
I think the idea is that it's going to be difficult, having looked at the code, to implement a version that is not similar enough to the original to trigger a copyright "derivative work".
For example, I believe it used to be the case (just a few years ago) that Microsoft employees were contractually prohibited from looking at GPL'd code.
But if alayne is planning on bringing that up every time someone shares a post that might contain GPL code, alayne is quickly going to become the most tedious person at the party, and that's why the downvotes are happening.
If your situation prevents you from studying and learning from GPL'd code, then just don't study it, and don't brag about not studying it.
To be clear, I didn't mean "why might there be fear about this?" (which I can understand, though it sounds a bit hysterical to me), but rather "has this ever actually happened (outside of a lawyer's dreams / nightmares)?", as a weaker version of the question "is avoiding this really good practice?".
P.S. Perhaps a good defence against a claim of copyright infringement would be "you can tell that my version isn't a copy, because it correctly finds 'nana' as a substring of 'banana'"? ;-)
I don't think it's good practise, and agree that it's hysterical, and think it stems from unthinkingly believing Microsoft's anti-OSS claims during their decade-long freakout about open source and the GPL.
I think that attempting to convince people not to participate in sharing and studying code with the rest of the world is an evil thing that they did, especially since the GPL was the most popular OSS license at the time.
I'm absolutely not anti-OSS. I personally think it is against the spirit of the GPL to copy code that is GPLed, though.
I'm just saying be careful about how you do your engineering. This seemed like an inexperienced developer and he mentioned glibc vocally along with the comment "I'll admit that originally I copied this logic from glibc's implementation without fully understanding it, but I've now taken a closer look at the Two-Way paper and understand what it does." It's certainly not good software engineering to copy implementation like that. It is good that he read the paper.
Now I regret even bringing it up. I don't feel that I deserve your ad hominem comments or characterizations of my motivations.
> I personally think it is against the spirit of the GPL to copy code that is GPLed, though.
Why? I assume that you don't mean literatim copying, which is obviously the whole point of OSS, but rather copying with modification; but, even then, a quick read suggests that http://www.gnu.org/copyleft/gpl.html#section5 explicitly permits this activity, as long as you meet certain conditions. Maybe I should have understood an implicit insertion "to copy, without acknowledgement, code that is GPLed"?
> Yes, it can come up. See the recent Google / Oracle lawsuit, where it was specifically presented that Dalvik was a 'clean room' implementation.
I have no doubt that it can come up—all sorts of ludicrousness can; but my question was rather whether there have actually been any legal consequences, not just legal challenges.
If this were a simple matter of some code being open and some not, this wouldn't be a multi-billion-dollar lawsuit. But the precise details of how Sun chose to write its licensing terms are extremely important.
Why is this good practice? Has anyone ever been exposed to a legal liability that could be creditably attributed to their merely looking at (as opposed to using) such code? (Although my suspicion is that the answer is 'no', this is an honest question.)