Arch Linux Has Been Living In A Legal Grey Area

[@windunursetyadi]

The Arch community: we like to live dangerously

[@__vyre]

describes us incredibly well actually

[@budgetarms]

And we love it

[@Lizard-o1m]

Me

[@theexile4694]

Arch Users: Unknown author with a package for a cooler looking discord client in the AUR? It's safe.

[@evoltaocao5078]

lol "try me, mofos".

[@insu_na]

I think the easiest way would be to require licensing every new PKGBUILD going forward as well as asking every maintainer to change the license of their current PKGBUILDs, then after the 4 months are over make all of the packages that did not get responses from maintainers public so that someone else in the community can create a new PKGBUILD without copying the original. It'd probably be good to see which maintainers have gone inactive or died in the meantime anyway, so having an overall picture of the maintainership should be useful in any case.

[@__Brandon__]

Don't think you realize how many packages there really are. 4 months isn't that much time

[@insu_na]

@@__Brandon__ for automated emails to all maintainers? ehhh

[@Clooger-]

​​@__Brandon__ Then you don't need a damn package if it's horrendously outdated. No one needs that much software, and if you do there's plenty of other options to build with.

[@felixfourcolor]

@@__Brandon__ But every package must be updated eventually, otherwise it would be considered unmaintained and they have to find a new maintainer or remove it from the repo. So isn't it easy to ask that the next time the maintainer updates the package, they add a license to the script? Eventually all packages will have a license, maybe not in 4 months, but it should be a smooth process.

[@SirLightfire]

it's not really possible to make a PKGBUILD "without copying the original" most of them basically do the same thing download code/binary from site extract/compile it put files in places

[@animehaider]

We talked about copyright, now let's talk about copywrong

[@theevilcottonball]

Why copyright? I want originalright.

[@Tooney712]

Nintendo

[@Roxor128]

I prefer copyleft over copyright.

[@Sandeepan]

What about pasteleft

[@theevilcottonball]

@@Sandeepan That would require clipboards support. Also cutleft is missing

[@ACuteAura]

Glue code (generally anything that makes two other things compatible) like this is not considered under copyright in the EU, just like the viral piece of the GPL is not enforceable in the EU. This is been discussed a lot on the EUPL blog, but the basis is Directive EC 2009/24.

[@arturpaivads]

I think that the automated emails can be sent and awaited for response. You apply the license for those that respond or any new project. And those old that keep will be slowly replaced anyway. Its far better then a consent by no reply. Its not that they have a problem right now about this.

[@howling-wolf]

Consent by silence, is actually a thing in germanny. Here it is mostly used in business contracts where someone offers you someting, you give them money and take the offered goods. This is scalable to fairly large transactions but is used by businesses less often since it can be quite ambigous.

[@szaszm_]

The "implicit consent" sounds like a big enough problem to make this approach infeasible IMO. It shouldn't be too difficult to script automated consent emails to all contributors, and only apply 0BSD to those files that are fully in the clear. Then the regular contributors can over time rewrite the PKGBUILDs or their parts that didn't get full relicensing consent.

[@sharp14x]

Most (if not all) of a PKGBUILD is trivial boilerplate, and not copyrightable.

[@iotku]

Big not a lawyer territory, but with implicit copyright to the PKGBUILDs that didn't get relicensed (in the USA anyways) wouldn't that further enforce the idea that something is copyrightable? I don't think any of the authors are realistically going to sue over it, but it does seem to be a contradiction there. I'm not sure you can force a copyright/license reassignment by lack of response to an email (unless there were already preexisting terms stating this). For something seemingly simple enough where you'd still have to argue that it's not copyrightable I don't see how you can recreate an equivalent PKGBUILD without technically infringing copyright if it were deemed to be copyrightable in some regions (even if it wasn't in all of the regions). OTOH we live in the age of mass generative AI copyright infringement so I'm not sure who really cares at this point.

[@szaszm_]

@@sharp14x Maybe. It's up to courts to decide in expensive lawsuits. I wouldn't wanna risk it for anything other than the most trivial snippets. I guess if it's just a template with ./configure; make; make install, then it's probably not copyrightable. But IANAL.

[@szaszm_]

@@iotku My point was intended to be that yes, you can't force a relicense by lack of response, and at that point, it's better not to put a misleading license tag on it. Arch Linux can't relicense work they don't own. For the second part: yeah, it's questionable whether they're copyrightable at all, but that's for courts to judge, and I'd rather not risk it. wrt AI, only human-made work is copyrightable, but when AI-generated stuff looks like a prior human work, that looks a lot like infringement by the AI tool or its user.

[@cheebadigga4092]

Germnan here: In Germany, "public domain" translates to "no copyright protection applies". I'm commenting because I was confused about the way you explained it xD

[@lx2222x]

Actually most things in Germany under "public domain" still have copyright protection

[@thingsiplay]

Public Domain is a right in the US, that is defined. We can only translate into German language, but the copyright system does not have the Public Domain. You can't put out work in German public without being the copyright holder.

[@makuru.42]

Also, in Germany, you could argue, it's not above the threshold of originality (Schöpfungshöhe).

[@cheebadigga4092]

ahhh that's what he meant thanks!

[@thingsiplay]

@@makuru.42 Depends case by case. Some package build scripts are fairly complex and basically programs on their own. For the simple ones where only a few lines exist and only one way of doing, then there is not much originality and I agree with you.

[@thingsiplay]

The problem is not the license itself, but licensing it after the upload without the consent of the original author. I would much prefer that every author updates the packages with accepting the new license. The is the only correct way.

[@budgetarms]

But what if the author is dead and the package is extremely famous?

[@thingsiplay]

@@budgetarms Rewrite it, because you cannot get the permission anymore. Just because someone died does not mean the licensing and the rights system changes. (Besides after 70 years I think.)

[@felixfourcolor]

@@budgetarms if it's a famous package, it should be easy to find a new maintainer

[@lucolesco]

​@@budgetarms If the program is extremely famous, it should have an upstream maintainer. If it doesn't, it is unmaintained and you shouldn't be using it anyway.

[@solbu-1973]

Hi Brodie. This is also why the Linux kernel _cannot_ be relicenced from GPLv2 to GPLv3+. As far as i understand, they had an internal discussion about it just for the sake of the argument, and they found that because several of the developers are _dead,_ they cannot ask everyone for permission to relicense. Torvalds personally _strongly disagrees_ with teh changes in the GPLv3, so even if he could relicense, he _will_ not. The Arch build script issue has the same legal problem. Not everyone can be contacted anymore for various reasons. They are therefore bound by the current legal terms, which as the lawyer explained is that the individual contributor retain _All Rights Reserved._

[@fb39ca4]

An example of why it's ridiculous for copyright to apply after the death of the author.

[@jakubrogacz6829]

@@fb39ca4 as much as I would loose good income if that were a case - sofware shouldnt be copyrightable - it's mathematics, you shouldnt copyright maths. If we must make it copyrightable - then it needs reconsideration if copyright for books is adequate at all ( and in general reconsider copyright in tech era ). For books and programs at least - the demand of possibility to acquire license legally should apply - if book is nolonger sold - in five or ten years and after attempting to contact author - everyone should be able to make a copy - otherwise we loose older media cause its not medieval with 10 authors per country.

[@XxZeldaxXXxLinkxX]

@@fb39ca4 I don't think it's ridiculous at all. If author has a spouse/children, then they should be entitled (with consent of the author of course), to benefit from that copyright. Having the copyright expire on death of the author opens up ... questionable means of gaining access to copyrighted material for those will ill morals.

[@kittenhero568]

​@@XxZeldaxXXxLinkxXthere already are means of gaining access to copyrighted material And even if as an author wishes to give their family ownership of their IP adter their passing, isnt 70 years a bit long? Might as well make it 100

[@Edde1018]

The estate of the copyrighted code could still consent to a relicense though.

[@rGunti]

The quickest way to get started is to require new entries to have a license (basically you're not allowed to post a script without selecting a license). For existing ones: 1. Ask contributors to specify a license of their choosing (while providing the preference for 0BSD, MIT-0, or another "good" license). 2. If the maintainer does not respond within timeframe X, mark the build script as "unlicensed" / "all rights reserved" and maybe even "deprecated". I think for that reason, any online platform as a clause in their ToS that grants them a license to the content you post there.

[@Finkelfunk]

Not specifically Brodie: Most legal systems actually have laws about the "it's common sense" part. Like if Google were to suddenly try and Trademark the phrase "to google something", it is commonly understood that this does not associate with Google and thus they'd fail to establish that. If something has always worked in a certain way there is a general expectation that it will continue to do so, at least in a lot of legal systems across the world. This is a point that has been argued many times over too.

[@BrodieRobertson]

That's fair but it sort of goes back to the most issues with the public domain, it's probably best to not rely on most countries allowing something and instead making it clear

[@Nina-cd2eh]

I don't think this applies here though. Or at least it could be argued otherwise. I think a proprietary pkgbuild could exist and be successfully defended for, depending on its complexity. We're not talking about a single specific pkgbuild that has been copied and actively used in a way it would violate a potential proprietary license, but about a collection of uncopyrighted works where new works keep being added to it. One of those works could come with a copyright notice, which you'd be liable and responsible for if you ignored it.

[@AndrossUT]

​@@BrodieRobertson it's not licensable content because it's too generic

[@BrodieRobertson]

@@AndrossUT that'll depend on the jurisdiction

[@SodaWithoutSparkles]

There have even been companies in China suing other companies for not obeying robot.txt, and they won. The reason given was not obeying "standard industry practices" constitutes unfair competition. Therefore using "standard industry practices" as an argument can be held up in court, in some cases.

[@rocstar3000]

It should be clear that any update that is done to the pkgbuild from now on means that you agree to use BSD-0, when uploading to the AUR it should prompt you this message that you need to agree to. That way, any new/updated/used package would have a proper agreement over it. And I guess that should give Arch more legal ground.

[@TheMadHatter626]

That's what I was thinking, it would be a whole lot simpler and more reliable.

[@rocstar3000]

@TheMadHatter626 yeah, and that's literally how every big tech updates their ToS, they throw you a popup and a yes, no button. In the case of the big techs iirc there is precedent of the jury saying that this agreement is not valid cuz many times you need to go to an outside link to read it all, most ppl don't read it and most ppl don't understand what it says cuz of the way its written (lawyers vocabulary). But in our case, ppl who are editing a PKGBUILD know what a permissive license is, so that shouldn't be a problem.

[@xard64]

"On the flip side" if you use the phrase twice you'll end up where you started from.

[@kittenhero568]

^me trying to convince the universe that the Rubik's cube shouldve been solved

[@Ganerrr]

Not in ℂ!

[@SeoFer]

Thanks for the subtitles!

[@guss77]

US public domain dedication is a gray area in the sense that it is covered in fog and heavily mined. The 1976 copyright act allowed authors to dedicate their work to the public domain, but 1988 Berne Convention Implementation Act superseded it and did not leave that option, also the current statute allows an author to cancel a transfer of ownership, and by extension also a PD dedication (if you can convince a judge that the 1988 act didn't supersede that part of the 1976 act) - so you cannot rely on PD dedication to be in effect forever, OTOH there's case law about software that was made available through a public free exchange to be automatically considered PD - so the AUR has that going for it. But as I said - it's a set of hidden legal land mines and the best IANAL advice I can give - if it says "public domain", run away as fast as you can.

[@josefa9671]

I actually had a discussion with the legal people in my organisation regarding close topics, it was for configurations, simple scripts and dockerfiles etc, and i would put buildscripts here too - here is what they said: A) Configurations and alike is not considered reacing "Verkshöjd" översätts "threshold of originality", "level of originality", "creative work of authorship". "modicum of creativity" eller "the author's own intellectual creation". (you also touch this in this video). B) We dont have public domain-concept in law in Europe , but we can use the CC0 (creative commons zero) to achieve roughly the same thing.

[@rantaviiva5218]

In certain experiences, eg. we have lots o' trade of cargo btwn 🇫🇮 / 🇸🇪 / 🇳🇴 / 🇩🇰 to this day also w/🇩🇪--sure this has resulted in Nordic geeks to have many experiences of using localized software programs even in case of peripherals and their utility program. Pointimg 👉to examples like that of a sw-manufacturer (i. e. sw house) can produce driver and utility sw for Windows OS, maby HP's the only one doing Linux driver as well. BUT, peripherals like DYMO -sticker creator is often a specific factory-made model that may have no localised version available for the next Windows version. The sw-house might cut costs and only create one localised version (for that ageing model of their DYMO -device) not the whole region). Given the controversy over home users' adoption of Windows 11 (also recentlly under heavy scrutiny over its sw and license like Recall, and massive lot of bloatware and other underllying privacy issues). However, “ZeroBSD" -lic. sounds cool or even a fair bet!

[@jakubrogacz6829]

there is public domain - just not a concept of relinquishing rights - at the very least the attribution must be made to author in copies. Dunno if dervatives have to do so too.

[@jefftp]

This is a super valuable video. I had no idea the Unlicense was so problematic.

[@t0m5k1]

Copyright, The legal framework that destroys everything we like, allows lawyers to fight needlessly to get big Corps money and ultimately creates soooo much admin you wish big Corps were not so damn greedy in the first place to make this an issue.

[@Outfrost]

Copyright as a concept is necessary, you don't want those same big corps just stealing people's work and profiting off of it, but yeah the current way it's set up is massively rigged towards corporations (especially evil publishers) and rich people

[@RanCham727]

You act like it doesn't also prevent big corps from stealing your work and reselling it.

[@t0m5k1]

​@@RanCham727 yes, because it was never designed for that. We've come along way since sheet music! Perhaps this law needs to be redone from the ground up to actually work correctly and not allow for so many damn loopholes to be used as legal devices as you and I (the consumer).

[@nephatrine]

If one of the original contributors of an existing pgkbuild isn't reachable for relicensing and they need to re-write it "from scratch"... in like 99% of cases the resulting package source will just be exactly the same as it was before since that's just how you build the thing. This seems very silly.

[@At-Dawn-We-Ride]

Well, as mentioned in the video, common sense is not really a thing when laws are concerned. Especially in a global context. 😐

[@beefyjoe]

Was wondering when you were gonna cover this, saw it pop up in my email inbox and immediately thought "Brody is going to make a video on this..." 😂

[@Kannaaaa]

The notable exception. Arch Linux, my beloved.

[@undefinedchannel9916]

Arch Linux more like "break my trackpad after updating" Linux.

[@aqua-bery]

​@@undefinedchannel9916 oh well

[@Lizard-o1m]

Nice pfp :)

[@fahadahaf]

@@undefinedchannel9916 nah more like break my trackpad 4 months before other distros do 😂

[@ForeverZer0]

@@undefinedchannel9916 Skill issues. Jokes aside, their is nothing Arch-specific regarding trackpad, this sounds more like an issue with your desktop environment and/or configuration, which actually installs trackpad-related stuff as dependencies. With the exception of pacman, its related libraries, and a few utilities like arch-chroot, archlinux-java, etc , there is absolutely nothing in Arch that is unique to or maintained by the distro.

[@lePirateMan]

I would like to inform everyone that the operating system I use is the one discussed in the video

[@SugarBeetMC]

btw

[@Dimkar3000]

if they solve it now, then they can avoid a pressing issue later. Its like a land mine under their feet. I think its a good idea to get it over with now, since there are no issue yet and most people will be positive to it. The only thing that feels weird is the 4 month timeframe. They coud make it mandatory for all package updates and new packages. Within a years most, if not all of them, would have switched organically.

[@max-mr5xf]

In Germany you could even make the argument that a license in English may not be legally binding. But I don’t know the answer to that question.

[@makuru.42]

No you can't, you can write a legal document in any living language you choose.

[@KasasagiWad3]

code copyright _generally_ works like it does for spreadsheets, the raw data/functionality of a program is not copyrightable, but the formatting and non-functional structuring of the code is, US and EU are generally pretty clear on this, it's elsewhere you run into issues.

[@AmirHosseinHonardust]

If public domain is in trouble, we all are! Sqlite is public domain!

[@LinusBerglund]

In sweden recipes are not copyrightable. This could be the same over here. We also dont really have public domain.

[@electricindigoball1244]

Disclaimer: I'm not a lawyer and this is not legal advise. IMO I think it makes sense to take care of this now rather than to wait for this to potentially become an issue. The earlier you can clear up legal ambiguities like this the better.

[@lorenzo42p]

I figured the build files would fall under the same copyright as the project itself. but then if it's not created and maintained by the project creator then it complicates things. in that case it would be copyright owned by whoever wrote it and whatever license they give it. same as any other project, the build files themselves are copyrightable.

[@deivedux9342]

The public domain doesn't even exist in software. The closest that we have is the Unlicense license, which, I think, is better for international enforcement because it doesn't rely on individual jurisdictions' interpretation of public domain, but rather has explicitly written out what is and isn't allowed by the license itself.

[@jasger_]

I'm actually pretty surprised that no one asked Rankin/he didn't look at it before this got merged, since he is a very prominent and active person on the arch mailing list. It kinda seems like an obvious thing to do before sending out that change notice.

[@autarchprinceps]

What about copyleft or the licenses of Elastic, MongoDB or more recently Redis? Wouldn’t they force things like pkgbuilds to be in their license? That’s probably the interpretation that lead to OpenSUSE licensing each in the packages license

[@nobodyimportant7804]

If things like PKGBUILD are copyrightable, we are very lost.

[@AsgerJon]

Legal gray areas are there for lawyers to get gain

[@ForeverZer0]

I don't personally like the entire concept of build scripts needing licensed. There has to be some commonsense line where we don't bother licensing code. If I share a line or two of Bash with someone to help them accomplish something, should it now be licensed? Of course not, but the same logic stands. At some point we need to just use commonsense. I never even considered "licensing" my build files: they are essentially just instructions on how to use the licensed code.

[@JG-nm9zk]

Can't have opensource without long discussions of software licenses where nobody involved is a lawyer

[@FeatherP]

Not grandfathering in the old packages seems like a no brainer in this case tbh, this gets rid of 'consent by lack of objection' issue, and eventually all packages will get migrated or deprecated over time

[@Jaymal10]

As has been said before doing it on future packages makes more sense. I also think it might be more interesting to just define package build.

[@Pelicanzzz]

I suspect SteamOS is the one most at risk from unlicensed PKGBUILDs. The other Arch based distro mostly aren't that important. But Valve is a target. At the least, Arch should specify going forward that contributions need to be BSD0. The "implied consent" argument might fail for someone saying "i contributed to Arch, not SteamOS".

[@unknownhours]

0BSD seems to be about as close to public domain as you can get without actually being public domain. I like it. I think it should be on new build scripts going forward, but too much work to do for all existing build scripts.

[@JSiuDev]

I think the issue also depends on where the org register. If it is registered in US and no official branch in other country, then it is not a "real", but a "technical" legal issue. And for the future, add the requirement that all contributed build script must be under certain license, and after a few years, as build scripts get updated along with the packages, the issue will auto phase out itself.

[@gigaherz_]

Zero-clause BSD may not be a good choice. IIRC, in some places a license can be retracted by the authors, so a license that wants to simulate public domain would need to be more elaborate to ensure the receivers of the copies continue having rights if the author decides to take things back and leave the open source world.

[@tutacat]

I am surprised the Un-license doesn't poly-fill the license with legalese for _the closest you can get to what public domain does_

[@mnoxman]

This is why CC and GNU EULAs exist. They are very permissive license. Public domain in the US means anyone can now pick it up and use it. They can even re-copyright it. See also: Carol Highsmith and Getty Images. A CC/GNU/OSI or other permissive license means the end user is given a number of rights but that the "ownership" of the material is still that of the author or the assigns or inheritors.

[@AlucardNoir]

The problem is that there isn't just one way to make/build a program, at least for programs that have complex build instructions. There is no PEMDAS for building software. That being said, arch has long since been on the edge when it comes to copyright. There's a reason companis like RedHat stick well away from Nvidia, while compsnies like davinci don't use work arounds for viseo decoding and most distros wint tuch ZFS with a 10 foot pole.

[@kelvinnkat]

There are a couple of countries with no concept of copyright.

[@-ism8153]

I suspect that a good number of countries would have stricter law that at least has the possibility of counting this as intellectual property if it ever gets settled, so it is good to tie it down just in case. In the U.S., intellectual property requires a minimal "spark of creativity". It doesn't matter if there are multiple ways to do it if there isn't a real idea behind doing it one way or the other. A build script is also so derivative of the software it builds that it could probably only count as IP if it counted as part of the software. Most of how to use and set up a piece of software is inherent to how it's made, so I'm not sure that it's really adding much information beyond the package.

[@robw3610]

Im sure some sort of automated process could be built to reduce the man hours needed for obtaining explicit concent. Honestly, the best time to have required licenses with these build scripts would have been when the AUR was first created... that being said, the second best time to do so would be now. If it were me making the decision, I'd make some sort of automated system to collect the needed approvals to relicense the scripts, and give a deadline to approve or reject. And any build scripts that dont get approval just get removed and remade by contributors. I cant imagine they'd get that many rejections. These programmers already submitted this code for free use. The potential issue comes in what percentage of the the contact information for these contributors is still good or not. This could be a lot of built scripts, or it could be only a few. Either way, many of these are so simple that they could be rewritten quickly by active contributors. But leaving them as having no license is not a great solution because it could put the Arch project in legal jeopardy. This really should have been a thing a long time ago...

[@esoqueexiste]

The need for licencies might have to do something with a private entity wanting to use and distribute the content of the repositories in some way (most likely valve)

[@TheMadHatter626]

Would it not have been possible to add a pop-up box to agree to add this license when the maintainer uploads an update their packages? That would force every maintainer to acknowledge and agreed to the new license without having to rely maintainers keeping up-to-date on their emails.

[@moarjank]

Or they could use the reuse lint tool and automatically reject package builds that don't properly declare licences

[@Krmpfpks]

In Germanys defense: public domain does exist, it’s just not for software. Also there is a difference between your rights as a creator (which are always with the creator, at least until death, i think it’s also inheritable to some extend) and the usage rights. A creator has an exclusive usage right that the creator can transfer to anyone they choose to (with or without the exclusivity). This is effectively handing over all rights just with the distinction that whenever the license expires, gets declared invalid etc all rights fall back to the person creating the work, not a company or anybody else. And nobody - besides the creator - can claim to be the creator.

[@Linuxdirk]

I always go the easy route when licensing my content. If it’s code, it’s MIT, if it’s media it’s CC0.

[@uuu12343]

I reckon the license for a PKGBUILD script should be/is the LICENSE of the original project you are building for Because the build script is just a wrapper that automates the building process defines the project maintainer/developer/owner

[@BogdanTestsSoftware]

I would guess requiring further packagers to either pick a license 0-BSD or some other is a sound procedure. It´s unlikely someone claims copyright infringement from AUR / Arch, but it pays to be on the safe side. Consent by default is not a thing in most jurisdictions, afaik, NOT A LAWYER here. If some parts do get contentious -- they do need be rewritten tho. Imagine someone with a *patent troll ethics* coming to Arch seeing a goldmine in there.

[@Kalimangard]

Why didn't they just send an email to every aur author with a link to a form confirming that their packages will be licensed under BSD-0, MIT or whatever and all packages by people who did not respond within a month get taken down, just as abandoned packages already are being taken down regularly. Once the form is done (a few hours of work at most) the rest can be automated.

[@At-Dawn-We-Ride]

Were you not paying attention when the video mentioned that many of the contributors' email addresses are no longer active? Some people change their addresses like underwear.

[@kupferdrachevideosfurdich8733]

In Germany this issue could be one of the very rare silence as consent between commercial actors. But too be sure you would have to sue a non answering Dev and present that to a court in a couple of years. in about 5 years or so a Dev that didn't respond and desides to might run int limit of claim issues. As always, only after atleast one judge saw this on their desk everything is speculation.

[@iodreamify]

I wonder if this could theoretically spell trouble for Valve if they're using Arch packages for the Deck

[@muskyoxes]

"Legally, common sense isn't a thing." It kinda is - all sorts of obvious absurdities aren't spelled out, like banning dogs from playing basketball

[@BrodieRobertson]

I don't think there would be anything wrong with a dog basketball league, it just wouldn't be very productive. I think what you're trying to say is that somethings are not explicitly illegal however it is similar enough to something else that it is a crime, this however is based on case law. A good example of this is the crypto world, much of it remains unregulated however a lot of what people have been charged for is based on existing law regarding the stock market.

[@Smethead]

Why not just require new additions/updates to have a free licence like the mentioned BSD 0. Send out monthly emails to those that havn't added a license to their existing scripts. After a predetermined time, hide/deactivate them on AUR until someone creates a new one or the original author notices it's inactive. Ofc this will end up being a purge of the AUR. But is that really a bad thing to make sure the script have an active maintainer and for those applications that users find missing, they or someone nice they know can prob whip one together :)

[@privacyvalued4134]

12:56 You can't legally change the software license of someone else's project. This is akin to someone saying, "I can take Apple's Mac OSX and change the license to MIT and publish the OS on GitHub." Or a video game from Nintendo. Or a movie from Disney. (Three of the most litigious companies on Earth.) How long do you think it will be before such a repo gets DMCA'ed? You can't just assign a license of your choosing when someone else holds the copyright to the software. That's the whole point of a default "All rights reserved." All rights. That includes changing the license!

[@Poldovico]

You misunderstand the proposal. This isn't about the actual software projects in the packages, this is about the scripts that create the packages. Say I want to install the GNU coreutils with pacman: the software that ends up on my computer is licensed under the GPLv3, and nobody is changing that. However, in order for pacman to grab the software and install it on my computer, the Arch maintainers had to download the sources, compile them into binaries, put them in some kind of archive, and include a script that tells pacman where to copy all the files. This is about licensing the code used by the Arch people to do all that, not about the code for the GNU coreutils themselves.

[@Sebanisu]

I didn't know about Zero BSD. I'd probably use that over Unlicense.

[@NFvidoJagg2]

INAL - I would suggest adding a script-license option to PGKBUILD and make it a requirement going forward. after four month, make any script without it auto assume BSD-0. and state, any author can opt-out at anytime by adding the script-license option. This still has the implied consent, but if someone brings up a legal challenge, it might give a better leg to stand on, by saying if you don't like then you can opt-out by tell us what you want to use. edit:note while I'm think of it: It may also stand that, BSD-0 most closely reflects the de facto understanding within the community for how the scripts are license.

[@coladict]

Sounds like the obvious solution would be that the license of the software being built with that script is to be applied.

[@AM-yk5yd]

Wait for SCO #2 or patent troll and they'll do something stupid eventually. IMO idea of opensuse (reuse the license) maybe good as on one hand you don't need to dance around CYA licenses like GPL/AGPL which go like "you can't use or link this software anyway except under our license" and installer steps away for couple of minutes. But otoh installer copy-pastes code with other packages, and some licenses are really contagious, so if something was done under GPL for GPL package first and then reused for apache project, apache project should be relicensed too.

[@jaythefruit_]

How does 0BSD compare to Blue Oak Model License 1.0.0?

[@mikehensley78]

i would think a build script would be considered code now that you've mentioned it. (never really thought about it before). from this point on maybe ppl should put a little #//note with the license they want to use at the end of their scripts.

[@bobthecannibal1]

They in fact aren't. Lists of instructions are not copyrightable. This is why recipes posted on websites always have a fricking 12.5 page TED talk before the fricking recipe: They're trying to prevent lazy people from cloning their site wholesale and deriving profit. They can't copyright the recipe, but they *_can_* copyright the TED talk. TL;DR: ingredients and steps _AREN'T_ copyrightable, unique style and artistic expression _IS._

[@Dosenwerfer]

There is a legal concept of tradition though. If something gets done for years by a vast majority of people and a silently accepted norm or expectation, that can have very real implications on how cases like this can get handled. Because I can hardly imagine a lawsuit that is exploiting this situation without explicit malicious intent. And that's why we've got humans judging this and the concept of reasonable person and their expectations.

[@JohnWilliams-gy5yc]

I agree with Tuchman's view. Focusing on who legally own build scripts simply profits nobody except greedy law firm shareholders.

[@ElliotWeishaar]

Subtle dig at nix, nice. I kind of like how the nix derivations are like the arch pkgbuild system.

[@o0shad0oo]

Automatically relicensing shouldn't be a problem IMO. Basically, you take ownership of the build scripts and relicense them as BSD. If the author objects he might sue for damages, but it'd be really difficult to quantify any damages for a script that was released to the public unlicensed and practically unattributed. So long as the script was withdrawn immediately I don't see how Arch could be held liable for anything.

[@TrabberShir]

For some corporate and governmental users, these scripts being used without an explicit license have been an accidental policy violation and, in at least one case I can think of, technically criminal violation, unless they caught this and clean roomed new build scripts, which I doubt.

[@BottleOfCoke]

I'm only one year into Arch, still learning the basics... is this important?

[@stefanalecu9532]

For you, no

[@uuu12343]

Nah, nobody truly gives a crap about this, well, unless you're a maintainer of a project then you gotta be alittle careful about your project's licensing However, even if you use the AUR/PKGBUILD, most people wont really care

[@No-mq5lw]

So we're talking about the PKGBUILD format the AUR uses? Why not start requiring a license when updating a PKGBUILD, emailing the author asking to add a license, and officially add a warning header for abandonware or ambiguous licenses? I can also see the argument being made that since the maintainer contributed the PKGBUILD script in the first place to the AUR, the maintainer gave the AUR a non exclusive license to redistribute the work almost like uploading to YT. I also don't really see any reason why the AUR can't also integrate ebuilds, since they are extremely similar in concept to each other anyways.

[@AbteilungsleiterinBeiAntifaEV]

It's not just the "format that the AUR uses", any package shipped by Arch has its corresponding PKGBUILD.

[@rexthewild]

I wander, can't the places where the pkgbuilds are stored just have a clause in the T&C that every unlicensed piece of code automatically falls under the US's definition of "pubilc domain"? Then it would get the other countries actually comply with this definition and the majority of the loopholes would be broken

[@joseoncrack]

Actually, I would assume that anything *intentionally made publicly available* and without any license or copyright mention would be considered public domain.

[@Poldovico]

@@joseoncrack surprisingly, no. To my understanding, in most countries it would be all rights reserved. Consider this: is a large mural painted overnight in a public place "intentionally publicly available"? By that logic, a bunch of Banksy's work would be public domain, but it is not. When it comes to copyright, just assume the meanest, greediest way it could work and you probably won't be THAT far off.

[@joseoncrack]

@@Poldovico Right, I probably poorly phrased my thought though: let me add to my sentence that no name would be present or any other means of identifying the author(s). Then it couldn't be considered anything other than public domain, what do you think? But anyway, this isn't the case for PKGBUILD files in Arch, so that's pointless. The names and emails of the maintainers and contributors are all listed. So, yes, without any license info, that should imply "all rights reserved" to the explicitely listed people. Which can be seen as kind of a "grey area", if you consider that a problem. Let's imagine one of the prolific contributors suddenly decided to take down all of their contributions (which copyright would allow), then it could take a long time for Arch to replace all their contributions and recover. So certainly not a safe situation. Although, while you can see typical Arch contributors as a bit "rough", i have never seen any such nasty behavior (the kind that would scr*w others up) from any of them.

[@jjjannes]

Germany dissallows abandoning copyright, it's just not possible.

[@joseoncrack]

@@jjjannes That sounds funky.

[@user-eq8mq6zw3e]

Public domain is why the CC0 license is ridiculously long and complicated

[@Burgo361]

I wouldn't underestimate how petty someone can be, I can see someone going after them as retaliation to arch doing something they don't like. Whether a court would take them seriously is another matter for the lawyers to argue about.

[@TheCommunistRabbit]

WHY AM I GETTING ADS ON BRAVE

[@veritanuda]

How is i t hard to jut give a consent form when someone submits a pkgbuild script to the AUR in the first place ? Do you agree to submit this script under the BSD-0 licence, tick or not.

[@svaira]

I don't really think that these build scripts need to be super "public domain"-like, more than other software, so I don't really see the value in this discussion. It would make a lot more sense to just ask the script authors to put it under MIT. This would make it a) easier to integrate with repos in MIT or GPL (which would be good, if the original source authors want to see that it works) and b) reduce possible legal issues, as licenses like MIT, GPL or CC actually have stood the test of time, but 0-BSD, afaik, is not really much used, and could have possibly issues later.

[@quintuple_lained]

“They’re not readily available” Laughs in gentoo

[@bigpod]

Im not sure it leaves licebse of sources up for interpretation someone created them therefore they own copyright, and if there is nothing saying otherwise they didnt grant anyone any rights. Im not a lawyer so what i say is not legal advice please consult your lawyer for that

[@ruroruro]

Nixpkgs mentioned. Let's gooooo!

[@stefanalecu9532]

Nix who?

[@20NewJourney23]

You advertised for the big green video site but you didn't provide your username there. I searched your name and your channel didn't snow up.

[@trumpetpunk42]

Gentoo has build scripts as well - Portage, modelled after BSD Ports. Edit: ah, he mentioned it at 10:20

[@glennmcgurrin8397]

Do people writing these scripts really expect others to sell the script for profit or to file off their names? The more extream designed to be as if public domain licenses give away more than I think people are semi implicitly giving away by making/publishing such a script, at the minimum I would think most would be thinking more along the lines of creative commons attribution.

[@defnlife1683]

Everyone else: “reeeeeeeeee this needs a license!” Arch Users: “dude, chill…..…. ……….. ………. Arch btw.” What a time to be alive.

[@TatharNuar]

isn't the pkgbuild part of the package?

[@At-Dawn-We-Ride]

No. I am the author of a variety of Python packages, all of which work without requiring Arch Linux. They in fact work cross-platform, which is why I publish them via PyPI, the Python Package Index. Pretty typical stuff. For several of my Python packages, related Arch Linux packages were later created; some by myself, some by others. Only at that point were the Arch specific PKGBUILD files added.

[@DePhoegonIsle]

2nd comment. The issue isn't the 'use' or 'intended use', it is the enforcement of a license without consent of the creator. (assuming it's even able to be copyrighted, and my opinion is the only copyright there is that of the build script language in use by the OS). Here is the thing... an all rights reserved license can be included, and even allow modification of the scripting as the point is to build the program & not claim credit, nor bei taken away from the source files its meant to build from. I've a feeling that a few someones will get into legally hot water for stripping the implied All rights reserved, and applying a 0BSD one which they had no idea existed.

[@raughboy188]

Hmm zero clause bsd might be good solution for arch if the unlicence becomes problem.

[@SnorreSelmer]

Probably a terrible idea, but what about setting a cut-off date, and if the package is updated after that, it's given the new license dealie. No retroactive licensing, and only maintained packages are part of the is change. If people disagree, have them pull packaging-scripts or something. Come to think of it, some idiot with a bunch of really core packages is probably going to go Super Sayan, pull build-scripts, and cripple Arch for a while until they can implement a fix themselves... This will be an interesting time no matter what. We already know that there is plenty of drama in the Linux-world (b2cache, Pottering, "just use distro X, it's so much better", etc...

[@archlinuxsys]

Arch always has been living in my life, can't get rid of it.

[@stefanalecu9532]

No way, you? Using Arch? With an Arch Linux profile picture and archlinux in your username? Nah, you must be using FreeBSD instead...

[@someone01233]

​@@stefanalecu9532 nah he is using plan 9

[@stephenreaves3205]

Once again, Arch is playing catch-up with Gentoo

[@THEMithrandir09]

Well then legally my usage of the pc should be licensed as well, since all I'm doing by touching rhe keyboard and mouse is writing and executing a one-time used script, like in a repl.

[@loyeyoung1068]

"Legally, common sense isn't a thing." Having been a practicing lawyer before I was an engineer, I can attest that Brodie is absolutely correct that common sense is not a thing.

[@howling-wolf]

Copyright law is so broken in most places of the world...

[@orbatos]

Just going to say this now, the automatic license change part is *not* legal, it is a direct and literal violation of copyright in all countries that have automatic attribution, e.g. Europe, the Americas, Australia, Japan, Korea, etc. They should know this is the reason license changes are such a pain.

[@no_name4796]

bro mispelled in the description lol

[@egar-k3h]

"Consent by no-answer or no-response isn't thing" Not a lawyer, but I am wary of lawyer speaking in absolute rather than in generalities. Many jurisdiction have things that can look a lot like a form of "consent by no-answer". Like a substituted service or a default judgment, neither are consent per say, but they are in the same ballpark.