Re: RFC: criteria A4 should be a C-class criteria

[Aaron Wolf]

There seems a confusion. A host can distribute non-free software legally
because their terms of service describe their permission to do so. When
the public downloads such software, they are subject to the terms of the
host's distribution and whatever terms the software is under.

Issue 1: a host is letting people download software which they are not
actually legally allowed to have or engage with (this is a serious problem)

Issue 2: a host is letting people download software under some default
terms-of-service that the host offers which doesn't actually license the
software for use by the downloader (still questionable for sure)

You compare to soundcloud etc, but a better comparison might be a forum
where people post things, text and whatever. The forum does not block
anyone from copying and pasting the test people post. That does not mean
people have the legal right to do anything with the contents of the
forum posts necessarily.

These things are indeed messy and there's some big issues to untangle.
But this gets into really making sense of the somewhat uncharted
territories of the practical application of hosts' terms of service

On 2021-03-12 4:31 a.m., bill-auger wrote:
> On Thu, 11 Mar 2021 20:59:55 -0800 Aaron wrote:
>> It's another to suggest it's wrong for free software
>> to be included there.
> 
>> The fact that some repos
>> also *separately* provide non-free software
> 
> you must have mis-understood me - i was not implying anything
> about the proximity to other non-free software
> 
> the analogy with distributions does not reflect the issue i am
> raising - distributions clearly identify, the licenses of the
> software they distribute; and they do not distribute source code
> for which they have no license to do so - the issue here, is
> that forges do not identify unlicensed software as such; yet
> they offer copies anyways, prominently, and with no warning that
> pressing this shiny "download me now" button, is usually a
> copyright violation, and is not actually permissible, neither by
> the authors nor service operators
> 
> i realize now, why my proposal was misunderstood - i should not
> have suggested removing A4 - C7 has a very different meaning -
> perhaps i could have worded it better - A4 could remain without
> redundancy or conflict with C7; and it has value of it own
> 
>   A4 is: "Does not permit nonfree licenses"
>   C7 is: "Does not offer public downloads of unlicensed works"
> 
> 
> On Thu, 11 Mar 2021 20:59:55 -0800 Aaron wrote:
>> We don't want a situation where any host that won't exclude non-free
>> software effectively decides there's no incentive
> 
> the proposed criteria would not require the exclusion of
> non-free software - it would require that non-free software be
> either private-access, or at least clearly presented as "not
> licensed for distribution", _and_ at the very least, to abstain
> from offering downloads of un-distributable works, so glibly and
> prominently as they do now
> 
> i would not be concerned about alienating anyone - this is
> merely suggesting that the forge does not entice their users
> into violating the copyrights of their other users - i think
> that any forge could easily meet the proposed C7 criteria, with
> a trivial patch or two; and that their operators are
> irresponsible at best, if they do not recognize this as a
> liability, and a dis-service to their own users - also, if this
> list featured more libre forges, and if the changes to meet C7
> got into the upstream forge code, they would most likely be
> taken by service operators without complaint - after-all there
> is no legitimate argument to oppose it
> 
> 
> On Thu, 11 Mar 2021 20:59:55 -0800 Aaron wrote:
>> The line we would draw is that if the distribution itself 100%
>> respects the license
> 
> they should also respect the copyright of works without a
> license; and avoid enticing their users to disrespect it -
> software freedom depends on a general respect for copyright -
> disrespect for copyright is disrespect for software freedom, as
> we know it
> 
> if someone publishes something, without a free license, that
> should indicate the intention that it is for presentation only,
> and should _not_ be distributed freely (just as a movie theater
> is authorized to present films; but not to give away copies) -
> it is the default expectation, that the author of some work,
> should have some confidence, that the authorized presenter will
> not re-distribute copies without permission - it would be very
> clearly "unethical" then, if the host (or anyone) offers copies
> to the public
> 
> this is the same default expectation of copyright protection,
> that people have when publishing to youtube, soundcloud, and
> similar services - if those services freely offered copies of
> everything, without explicit permission to do so, their users
> would be crying about it in droves, and the hub-bub would be
> widely reported as a scandal
> 
> if the "ethical repository criteria" were focused on software
> freedom alone; then perhaps this would be less relevant - i
> agree that no one's software freedom is directly impeded by this
> - GPL violations do not impede anyone's software freedom either;
> but if these websites featured a prominent "violate the GPL"
> button, which somehow assisted in doing so, i suspect that there
> would be a criteria rejecting it, not as optional, but at the
> essential level
> 
> these criteria are not presented as software freedom concerns
> though - they are concerned with the ethical practices of server
> operators - so, here is a counter proposal: drop the word
> "ethical" from the description of the criteria - this practice
> of entrapment is obviously unethical; so the only way to ignore
> it, would be removing "ethics" as the primary concern - it is
> the equivalent of a movie theater with a heap of unauthorized
> copies of the present matinee in the lobby, under a large sign:
> "Free - Take One"
> 
> more importantly though, it _is_ significantly detrimental to
> software freedom; because in practice, it displaces any
> incentive for developers to license their code
> 
> people who publish their code on these forges, want only for
> other people to see it, perhaps to use it, and perhaps to
> collaborate on it - proper licensing is of absolutely no concern,
> especially to beginners - if the forge does not require the
> proper distribution permission, then there is little incentive
> for anyone to give it - they are gratis, automatic "E-Z"
> distribution channels; and that satisfies many people, with the
> false belief that they are allowing their code to be used and
> shared freely
> 
> this practice robs all arguments for software freedom of their
> convincing power; because in practice, all software on a typical
> forge is presented to the naive, as freely distributable; and it
> _is_ very literally "distributed freely", regardless of the
> license
>