5 ms·
Applying the Free Software Criteria
- benwerd 11y agoI don't think this is a realistic request on the modern web - particularly when it comes to JavaScript code, even if it's a technically accurate interpretation of how JS is transmitted and executed. (Licensing is an interesting thing to think about here.) But hey, for what it's worth, every piece of JavaScript we run is free software compatible. Hooray! https://github.com/idno/known https://github.com/idno/known
- dikaiosune 11y agoI don't get why it's ok in this context (if you agree with the inherent unethicality of nonfree software) to have compatibility and support for interoperability with nonfree software. It might not purposefully lead someone to a less free situation, but when you're speaking as absolutely as Stallman does, I don't see how you can condone actions that may prolong someone's lack of freedom. I'm genuinely confused by this - if there's anyone who could lay this out without dogma and hyperbole I'd be much obliged.
- Retra 11y agoTwo reasons: 1. If nobody uses your free software because it doesn't interoperate with what other people use, then you're really not accomplishing much by making your software free. 2. If you want to empower users, you can't very well tell them that they aren't allowed to use certain kinds of software just because you don't want them to. That would be quite hypocritical.
- tedks 11y agoStallman in fact is quite a centrist and a pragmatist. The problem is that the software freedom landscape is so incredibly shifted towards the non-free end of the spectrum that he appears to be a radical. We can easily conceive of a radical splinter group of the FSF, the SLF (software liberation front), advocating: 1. Only the use of the AGPL license. No other license is considered free, because all other licenses can be used in some way to limit the freedom of users. As soon as a license more free than the AGPL comes along, only that will be the free license. (In contrast, the FSF promotes the Apache license, the LGPL, and similar, when it is strategic to the free software agenda.) 2. Never interoperate with non-free software. To do so is to condone actions that may prolong someone's lack of freedom. The fact this may be necessary for their work or personal life is irrelevant. (In contrast, most GNU software runs on Windows, which Stallman probably sees as no different from it running on the proprietary Unices he wrote it on.) 3. Do not allow the execution of non-free code, if your free program executes code. (In contrast, gNewSense or IceCat will execute any x86 or javascript you throw at them, even if it doesn't have licensing metadata that identifies it as free.) The FSF and Stallman are in fact quite tolerant in an absolute sense compared to our hypothetical SLF. This doesn't even go into the other strategies or tactics such an extreme group could engage in, such as firebombing the Microsoft offices or forcibly liberating source code. Instead, the FSF restricts itself to very tame boycott politics. You in particular have a gross misunderstanding of Stallman's strategies and tactics, even if you might understand his goals. Stallman is a moderate. He thinks pragmatically, even if he's sometimes wrong (protesting DRM at apple stores is probably a waste, as is protesting furriers, but sometimes you can't help but want to speak out).
- mankash666 11y agoSo your argument is: Stallman is a moderate when compared to a hypothetical software terrorist? In the real world, Stallman is still a radical. The English meaning of the word freedom cannot be twisted for suiting one's philosophy. True freedom lies in allowing your work to be modified without restrictions - like the BSD or Apache licenses offer. Daemonizing software released under those licenses, and proselytizing against their use IS radical.
- vezzy-fnord 11y agoThe FSF does not demonize nor proselytize against the BSD licenses. They are free software licenses and thus FSF-approved.
- mankash666 11y agoFSF's most vocal proponent has gone on record with his opinion on the BSD/Apache licensed software. http://developers.slashdot.org/story/15/02/08/210241/rms-objects-to-support-for-llvms-debugger-in-gnu-emacss-gudel http://developers.slashdot.org/story/15/02/08/210241/rms-obj...
- adrusi 11y agoThat is not true freedom, because it still works within the confines of the law. No, true freedom would mean it it impossible to have anything be anything other than public domain because there would be no one to enforce IP. The free software movement with its copyleft licenses strives for this kind of freedom, but having been yet unable to eradicate IP law, essentially uses it to undo itself. "My creation rejects intellectual property and so must everyone who uses my creation." It seems contradictory because it is, and it is very self aware of that. Because licenses like the gpl aren't an end in themselves, they are simply a crude attempt to mimic the idyllic end game: no copyright law covering software at all.
- tedks 11y ago1. You are factually wrong in saying that Stallman demonizes software released under permissive free licenses, and factually wrong in saying that Stallman proselytizes against their use. I will not link you to any of the numerous examples of Stallman doing so, because you must find your own truth, and you must make your own DuckDuckGo searches. If you have any honor, you will do this. If you do not, you have blinded yourself and chosen, willfully, ignorance over knowledge. Nothing can forgive this. 2. The point of the parable of the SLF is to pose an actual radical group in the software freedom space similar to the actually radical groups in other political spheres such as animal liberation (ALF), environmentalism (ELF), feminism (Wimmen's Fire Brigade), socialism (RAF), anti-racism (Black Panther Party), or queer liberation (Bash Back!). The outlook of a radical is fundamentally different from the outlook of a liberal. A radical does not negotiate. A radical does not not make strategic compromises. This is rarely a strategic or tactical choice and more often an ideological one, and it seems mostly orthogonal to the effectiveness of the radical group (history is rife with radicals who have made enormous gains, often being entirely transformative on the societies they set out to change). Stallman does not demonstrate any of the qualities of a radical other than that his views are very different from most others. But this alone does not a radical make. 3. Arguing from a definition is a weak stance indeed. It is an appeal to the most meaningless of authority. Stallman disagrees that freedom lies in allowing your work to be modified without restrictions; he rather posits that freedom lies in the ability of a user to have total self-determination over the code that they execute. Neither of these definitions is in a dictionary, and a dictionary cannot offer any source of philosophical, ethical, or moral truth, nor should it. This is a deeper matter than that.
- brudgers 11y agoDisclaimer: I don't agree with the proposition that non-free software is inherently unethical. This doesn't change my admiration for Stallman nor does it lessen my respect for his position. My speculation would be that the position is deemed ethical because Stallman's view [as I understand it] has always been based on the primacy of freedom for the user. Stallman has not [so far as I am aware, though admittedly I do not approximate a one-eyed king on HN] ever suggested that users should be absolutely prohibited from installing and using non-free software. Only that it not be included under the "Good Houskeeping Seal of Approval". What Stallman has done here, as elsewhere, is a pair of tasks. One is to carefully define a set of terms and the second is to state his position with respect to those terms. Agree or disagree (and it's more important in disagreement I suppose) one can reasonably know more or less exactly what one is agreeing or disagreeing with...e.g. this exactness is the reason that the issue of interoperability as a possible contradiction can be raised and clearly articulated. Stallman speaks absolutely in the sense of clarity. He knows his position and his beliefs. My opinion is that they have perhaps done more to shape the current software ecosystem than anything else. We live in a world where a common moderate position for sober business women is an Apache type license not a 30 day trial binary, cripple ware, or even shareware...let alone something sold by a four-legged meat eating sales team. Great wealth has been built upon a shared commons, and programmers and developers have vastly benefitted in their economic lives.
- mankash666 11y agoI'm a huge user, believer and propagater of QUALITY free software, or quality anything. It's preposterous to suggest that non-free programs limit individual freedom and liberty. Quid-pro-quo: the undeniable biological, evolutionary instinct to receive something (money) for a rendered service/product (software) simply cannot be deemed criminal. There's very little incentive to charge money for something whose blueprint is free. RedHat is the exception in the software world, not the rule. Very few profitable free-software based companies thrive - take canonical for instance, they're barely profitable despite putting out a VERY high quality Linux distro Plus, why stop at software - why not release the blueprints of EVERYTHING ever made? Oh - and please tell your ambitious school going kid to NOT aim to finish first, because, when he grows up to be a successful inventor, Stallman and co. would have brainwashed the masses into daemonizing inventions that aren't made public property!
- yarrel 11y agoPatents are a means of ensuring the release of inventions to the public. Patents are bad, right?
- paulryanrogers 11y agoPatents are a tradeoff: short term power granted (to patent holder) in exchange for long term freedom (for everyone).
- gillianseed 11y ago>Quid-pro-quo: the undeniable biological, evolutionary instinct to receive something (money) for a rendered service/product (software) simply cannot be deemed criminal. FSF/Stallman in way claims that charging for software is criminal or even unethical. The whole point of Free Software is to give and preserve rights for end users, rights which enable end users to do what they want with the software they recieve, like examining it to make sure it doesn't do anything bad (for the user), modifying it to do what the end user wants, copy it, etc. Now obviously this makes it incompatible with the traditional way of commercial software, which is proprietary distribution of binaries, which often also comes with DRM schemes further limiting what the end user can do with said software, and worst case, exploiting the end user by enganging in hidden or poorly-disclosed unwanted actions (spyware, rootkits etc).