3 ms·

> (I don't approve of this system, I am merely describing what I see as the status quo.) As has been said, being a FOSS "advocate" may put you on many sides of

> (I don't approve of this system, I am merely describing what I see as the status quo.) As has been said, being a FOSS "advocate" may put you on many sides of any copyright issue. > RIAA isn't going to waste time finding out if the recipients were planning to buy CDs otherwise. If you receive software under a license that is too restrictive for you, then there is always the option to pay the rightholder to grant you a more permissive license in exchange for money. Your comparison fails to understand how different these claims are. First, those CDs have a price on the tin. That is -- this claim would have "actual damages" which can be readily determined. Second, you're speculating a new license for either Linux or OpenZFS is possible, but I don't think anyone imagines either is readily available given the number of rights holders. If Oracle was offering to relicense (BTW only its portion of the ZFS source) for a known and reasonable amount $X, like your CDs, then your comparison might make more sense (but read on, the offer of another license does not preclude a finding of non-infringement). Thankfully, of course, copyright law provides for statutory damages where actual damages cannot readily be proven, as here. But those nominal/statutory damages divided up among Linux rights holders would be almost insignificant, and therefore this poster above is likely correct -- this may not be a legally cognizable injury. While I may agree re: this poster's damages analysis, I'm not even sure that's the best form of the ZFS argument. IMHO where anyone creates a functional software interface, whether it be via game console's cart or a kernel module, fair use allows one to create to software and use that interface. Period. Your license can say whatever it wants. Your license can say "all your software now belongs to use" and IMHO it is completely unenforceable. IMHO the courts have favored software interoperability and fair use for years. For example: In Sega Enterprises, Ltd. v. Accolade, Inc., the court denied copyright protection for “functional requirements for compatibility with the Genesis console . . . .” Under this approach, the right of the kernel module author to create a compatible module overrides any nominal copyright infringement created when that author creates static or dynamic links to kernel code. Sega’s Genesis console had no public API whatsoever, stable or otherwise, yet the court still denied protection to these functional elements. Regardless of the status of the API or system interface required for compatibility, any parts of a program that a developer must copy — such as kernel headers, definition files, variables, or mandatory Linux kernel function calls — in order to create a Linux-compatible kernel module would not receive copyright protection.[0] The case is much stronger re: ZFS because the court explicitly protected reverse engineering in Sega, whereas here the Linux kernel devs have published the code to their interface. Note, Sega even offered Accolade a license (just as we discussed the potential for an alternative license above) that would have allowed Accolade to create Sega-compatible games. Notwithstanding this offer, the court declined to find copyright infringement in Accolade’s creation of Sega-compatible games without a license. [0]: https://www.networkworld.com/article/836041/smb-encouraging-closed-source-modules-part-2-law-and-the-module-interface.html https://www.networkworld.com/article/836041/smb-encouraging-...

Y_Y 2y ago
> As has been said, being a FOSS "advocate" may put you on many sides of any copyright issue. Who's an advocate here? > Your comparison fails to understand how different these claims are. Comparisons don't understand anything. I mean to draw attention to aspects shared by two different things. > I don't think anyone imagines either is readily available given the number of rights holders Me neither, but I don't think that's pertinent. Why do you? Finally, you make an argument about interfaces, but I think the issue here is the use of CDDL licensed code, rather than the (re-)implementation of an interface. To me it seems like the finding in Sega v Accolade aren't relevant here.
- mustache_kimono 2y ago
  > I mean to draw attention to aspects shared by two different things. Yes, I understand you meant to reason by analogy, and I told you that I think your analogy is flawed. When you reason by analogy, perhaps you should be prepared for someone to say, "That analogy doesn't make sense in this context." Mostly because -- your analogy doesn't make sense in this context. One, is an actual damages case, and the other is statutory damages case. When we are strictly discussing damages, your comparison is an apples to oranges comparison. You said: >>> If you receive software under a license that is too restrictive for you, then there is always the option to pay the rightholder (sic) to grant you a more permissive license You suggest "there is always the option to pay the rightholder (sic) to grant you a more permissive license" and I'm saying that it's impossible in this circumstance? There is no one stop shop for the licensee. There is no 50 stop shop. In your scenario, where we are to imagine an incompatibility exists, you also imagine a theoretical alternative license would be relevant to a court. And I'm saying, just as relevant, is the actual impossibility of obtaining the consent of either all ZFS licensors or all Linux licensors. We've created effectively perpetually licensed software via our shared use of FOSS. And now one party we imagine saying -- actually these two works will be incompatible forever. You think a court won't think that's wild? More significantly, this entire line of argument has an angels on the head of a pin quality. The GPL and the CDDL are the licenses that have already been chosen by Canonical, because Canonical didn't, like you, read an incompatibility into this CDDL-GPL combination. So -- fair enough that Canonical didn't feel the need to cast about for some theoretical alternative license, if it doesn't feel that the CDDL and the GPL are incompatible. > I think the issue here is the use of CDDL licensed code As discussed -- I don't think it matters much how the code is licensed. If it makes it easier, imagine the CDDL sourced code is actually a closed source, proprietary filesystem. So -- Linux provides a module interface, but it could be any interface. Just as in Sega, IMHO even proprietary code can use GPL licensed code, which provides an interface, for the purposes of interoperability. Why? Because "fair use".

Y_Y 2y ago

> As has been said, being a FOSS "advocate" may put you on many sides of any copyright issue. Who's an advocate here? > Your comparison fails to understand how different these claims are. Comparisons don't understand anything. I mean to draw attention to aspects shared by two different things. > I don't think anyone imagines either is readily available given the number of rights holders Me neither, but I don't think that's pertinent. Why do you? Finally, you make an argument about interfaces, but I think the issue here is the use of CDDL licensed code, rather than the (re-)implementation of an interface. To me it seems like the finding in Sega v Accolade aren't relevant here.

mustache_kimono 2y ago
> I mean to draw attention to aspects shared by two different things. Yes, I understand you meant to reason by analogy, and I told you that I think your analogy is flawed. When you reason by analogy, perhaps you should be prepared for someone to say, "That analogy doesn't make sense in this context." Mostly because -- your analogy doesn't make sense in this context. One, is an actual damages case, and the other is statutory damages case. When we are strictly discussing damages, your comparison is an apples to oranges comparison. You said: >>> If you receive software under a license that is too restrictive for you, then there is always the option to pay the rightholder (sic) to grant you a more permissive license You suggest "there is always the option to pay the rightholder (sic) to grant you a more permissive license" and I'm saying that it's impossible in this circumstance? There is no one stop shop for the licensee. There is no 50 stop shop. In your scenario, where we are to imagine an incompatibility exists, you also imagine a theoretical alternative license would be relevant to a court. And I'm saying, just as relevant, is the actual impossibility of obtaining the consent of either all ZFS licensors or all Linux licensors. We've created effectively perpetually licensed software via our shared use of FOSS. And now one party we imagine saying -- actually these two works will be incompatible forever. You think a court won't think that's wild? More significantly, this entire line of argument has an angels on the head of a pin quality. The GPL and the CDDL are the licenses that have already been chosen by Canonical, because Canonical didn't, like you, read an incompatibility into this CDDL-GPL combination. So -- fair enough that Canonical didn't feel the need to cast about for some theoretical alternative license, if it doesn't feel that the CDDL and the GPL are incompatible. > I think the issue here is the use of CDDL licensed code As discussed -- I don't think it matters much how the code is licensed. If it makes it easier, imagine the CDDL sourced code is actually a closed source, proprietary filesystem. So -- Linux provides a module interface, but it could be any interface. Just as in Sega, IMHO even proprietary code can use GPL licensed code, which provides an interface, for the purposes of interoperability. Why? Because "fair use".