CUPS relicensed to Apache v2

From the GPLv2: "The source code for a work means the preferred form of the work for making modifications to it.". If you wrote it in binary, you could redistribute it in binary. But that's not what's done, so I don't see the point of your example.

>Freedom is the freedom to do things that others may think are stupid and pointless, so long as they, don't involve hitting someone over the head or such. To me, that means Freedom absolutely involves (A).

Except things aren't so simple at all due to things like private property and physical force. Your freedom to wander onto someone else's private property is restricted despite wandering being non-violent; your freedom to wander is considered lesser to the private property owner's freedom to choose who gets to be on their land. The freedom to use physical force is granted to police when enforcing the law. My point is that different freedoms aren't all equally valuable -- some are valued more than others. I don't know a mechanism for evaluating the values of various freedoms, presumably I may learn by reading philosophers who discussed freedom such as John Locke and John Stuart Mill, but I have yet to do so.

>I think what you are saying could be (mis?)interpreted as being analgous to "Bob should not be free to become an unemployed rock musician, because if instead Bob had become a cardiologist, he could have helped me to live longer, therefore his freedom to make music I consider to be causing me medical harm".

Perhaps in the rough sense that you are *compelled* to do something, but being compelled to *release source code* and being compelled *into an entire career* are two very different levels of compulsion.

>Now that I've been through more discussion (which reminded me of highly nuanced considerations I had half-forgotten), I can posit that in this case, though I have no idea why Apple did choose to go GPLv2->Apache2 instead of perhaps GPLv2->GPLv3 (amongst many other possibilities), I could see it as a reasonable choice based solely on how the GPLv3 preamble seems to be dismissive of the value of (software) patents *in general*, wheras the anti-patent language in Apache-2.0 _could_ be interpreted as narrowly directed against the perceived abusive tactics leveraged in the KAMIND/JMRI case of suing for damages based on number of open source code downloads.

>I don't expect anyone to necessarily bother parsing that, in the same way I meant it, and agree with it. But it is how I'd prefer to view the issue :) Though I do realize it is a complex spectrum of people with widely varying levels of interest in and opinions about the issue. Clearly I see respect for, or dismissiveness of, the patent system in general as being at the heart of the issue. And after reading a significant fraction of one single bonafied patent from 1998, I must admit it's a hard thing to love. In this case, my point is that some people might find the importance of respecting the patent system in general, to be more critical than the (quite significant) differences between Apache-2.0 and GPLv3. Certainly it's plausible to me that Apple might be somewhat along those lines - God only knows how many software and other patents they own. Certainly CUPS contributors could not have been under any misimpressions about their copyright holder's general stance towards the patent system.

My concern was a move away from copyleft moreso than the licenses' stance on patents, but perhaps patents were the primary consideration.

Sure; there also wouldn't be DMCAs, NDAs, ToCs, &c. But there are, and there is copyleft.

Nope, but I've been there myself in the past.

>And actually, for the project I want to write, I'm hard pressed to tell the difference between the two! In other words, if I licence my code GPL, and write it the way I intend to write it, the recipients will be able to treat it pretty much as if it's BSD!

Sounds like a service-based project, perhaps Affero :P?

The JMRI folks are a harsh reminder of what can happen. In short -- a company took the JRMI projects's open-source software, patented it, and then tried to shake down the original authors for patent infringement. The original authors eventually won, but at the cost upwards of half a million USD, and at least four years worth of grey hair.

But for me personally, I write Linux drivers for printers, usually via reverse engineering. Some of what I've reverse-engineered is surely covered by patents (not that I have any intention of searching to find out!) These reverse-engineered drivers, in many cases, produce better output and are more featureful than the manufacturer's, to the point where the manufacturers themselves are utilizing them -- while not publicly acknowledging or supporting my efforts. That latter bit in of itself is fine (as long as they respect the license terms!) but I have been told by lower-level minions in more than one company that there have been ongoing internal debates about whether or not to sue me over the very stuff they are either pointing their customers towards or actually using for other internal products. (Because I obviously must have stolen something; there's no way anyone could have possibly figured out their HIGHLY PROPRIETARY algorithms without being naughty!)

So. I release my stuff under the GPLv3 in part because its patent clauses remove one of the weapons that could be used against me -- or at least substantially raise the cost of using it. They don't actually _need_ my code; they are the one with access to documentation, hardware, and the actual engineers who designed/built the gear, and could have written something for Linux themselves if they'd cared to invest the time and money.

(as an aside, last weekend was the 10th anniversary of my first printer hackery. How time flies when you're staring at multiple hex dumps)