3 ms·
Yes, the morass of software licensing sucks. But unfortunately another side effect of that is that not choosing a license or making uninformed decisions about l
by spacelizard 13y ago
Yes, the morass of software licensing sucks. But unfortunately another side effect of that is that not choosing a license or making uninformed decisions about licensing creates situations that end up with users becoming extremely susceptible to copyright trolling. It is true though that the solution is obviously to ease up on some of the powers granted by copyrights.
- nkurz 13y agosituations that end up with users becoming extremely susceptible to copyright trolling I worry the trolling is largely extra-legal, and that the actual terms of the license matter little. Currently, anyone can create an account on GitHub and without proof-of-positive-ID can apply any license to any code they choose to upload. Whether they have the right to issue this license is unknown to the user other than by direct due diligence. Consider: I take a small function out of a piece of proprietary or GPL code, and without permission put it into a library which I publish on GitHub. You use it in your successful commercial venture. Is your legal exposure different depending on whether I chose "BSD" from the drop-down or left it unlicensed? Is my liability different? What about GitHub's? Does there exist a software equivalent of real-estate title insurance? I like the SQLite approach to this: http://www.hwaci.com/cgi-bin/license-step1 http://www.hwaci.com/cgi-bin/license-step1
- tjaerv 13y ago> Currently, anyone can create an account on GitHub and without proof-of-positive-ID can apply any license to any code they choose to upload. Whether they have the right to issue this license is unknown to the user other than by direct due diligence. A good exemplar of this from just the other day: http://lists.gpl-violations.org/pipermail/legal/2013-July/004100.html http://lists.gpl-violations.org/pipermail/legal/2013-July/00...
- nkurz 13y agoYes, that is a gem. With liberty, I've submitted it hoping it will also be a conversation starter: https://news.ycombinator.com/item?id=6094223 https://news.ycombinator.com/item?id=6094223 Reading one of the linked comments, it's even more interesting because the "Samsung proprietary code" looks clearly copied from the Linux kernel: http://phoronix.com/forums/showthread.php?82576-Samsung-Accidentally-Leaked-The-exFAT-Linux-Driver&p=344872#post344872 http://phoronix.com/forums/showthread.php?82576-Samsung-Acci... Sometimes I wonder if much of the corporate fear of open source is that it would slow development and restrict code reuse, since rather than just swiping the code and releasing the driver, once it's out in the open they'd also have to have endless discussion of what is and is not a derived work.