GitHub takes down YouTube-dl with DMCA notice.
October 25, 2020 1:06 PM   Subscribe

 
This was rather striking to see roll out, I have something built off of Youtube-DL, and it's not even touching Youtube!

The legal details are also interesting, there's better analysis than myself floating around but I think the big notable detail that could be missed is that it wasn't a DMCA-notice as we generally know them (S.512), but S.1201, "the tool itself is illegal as an anti-circumvention measure".

That the unit-tests included VEVO-protected videos as examples didn't *help* their case any, but I'm fairly certain if it wasn't that it'd be something else, so I'm not down on the Y-DL devs for that.
posted by CrystalDave at 1:18 PM on October 25, 2020 [8 favorites]


Well, that sucks.
posted by confluency at 1:21 PM on October 25, 2020 [3 favorites]


This was rather striking to see roll out, I have something built off of Youtube-DL, and it's not even touching Youtube!

I've been working to build up my personal GitHub page out of boredom and to help out my freelance career. I wanted to simply download all the NFL teams and maybe delayed score. While these are facts under law, scraping any sites that utilize NFL APIs is illegal. I'm talking about simply trying to get a list of teams with conferences. Don't believe me? Try finding a simple JSON feed that lets you use it to just get teams and conferences for a given year.

On a personal note I recently started jogging with an iPhone that didn't have mobile access because I've dropped too many so I decided to use an old one. I found out that all ad driven apps (Spotify, SoundCloud) don't work without an account. To get an MP3 to an iPhone, I needed to setup iTunes which meant an account, etc. Despite having AirDrop and all kinds of useful things to easily transfer between devices if you don't want to go through the approved route it ends up being a mess. I even setup a script that allowed me to curate what I wanted, input the name and had youtube-dl download them then sent them to listensound.

Between this and shows jumping around expensive streaming services I'm about to jump ship back to downloading my own. I dropped doing so due to a short magical time when streaming services made it more convenient then running a server, but it looks like there's been a lot of progress in offloading an even reasonably sized collection to the cloud. Using rclone for cloud storage that looks local, Radarr to download and organize movies. There's also a vibrant community to automate this all with Docker and a collection of bash scripts so you can have a local cache, cloud storage and it all automated (looking of this, will post when I find it).

I might add a Github cache too to the list of things I store. Sucks as I didn't really want to be maintaining a private mirror of Github but I also don't want all my projects failing because of a DMCA notice. Hopefully like the above links someone makes it as easy as downloading a Docker image and uploading it to AWS.
posted by geoff. at 1:27 PM on October 25, 2020 [16 favorites]


Ah here is one, Ultimate Media Server I believe there's another one that takes this a step further and setups up local caching plus cloud services so you don't have to really think of it beyond putting a bunch of hard drives attached to a RasberryPi. But the point being that adding a git repository to this would be trivial at this point they've more or less created a giant file storage ecosystem in the cloud.
posted by geoff. at 1:31 PM on October 25, 2020 [2 favorites]


... I'm really tired of the RIAA. We have a pandemic. Can that not be enough? I feel like the RIAA is kind of de trop at the moment.
posted by trig at 1:42 PM on October 25, 2020 [10 favorites]


As someone who lives in a country where many regions don't have access to reliable internet*, the assumption that you either have fast, unlimited data access or you don't matter has been exhausting.

I can't always stream, which means that I often don't have full access to services I pay for. Netflix, for example - I can download some shows, but only on my phone, and not all shows. Because it's sooooo important to prevent me from having access to the files.

And then there are problems with buffering. I'd watch your stupid YT video if I could buffer more of it in advance. I did grow up in the nineties, I'm used to waiting. But nooooo.

I've downloaded videos from YT because I couldn't stream them. I've downloaded TV shows that I've paid to stream. Seriously, I'm tired of being collateral damage in the "fight against piracy." You're driving me to piracy, corporate shitheads. I'm about to sign up for a VPN and go back to torrenting because you won't let me pay a fair price.

(I know this isn't the most topical rant, but these are the feelings I have about the background context here)

*The USA
posted by Kutsuwamushi at 1:55 PM on October 25, 2020 [66 favorites]


A pull request was opened on github's dmca repository (which contains records of dmca takedowns received by github), replacing the original contents with the youtube-dl code.
posted by polytope subirb enby-of-piano-dice at 2:15 PM on October 25, 2020 [6 favorites]


Nooooooo
posted by Going To Maine at 2:33 PM on October 25, 2020


Good opportunity to finally review the source code of a very excellent python project, at least.
posted by Going To Maine at 2:39 PM on October 25, 2020 [6 favorites]


What angers me about all this is that Google is arguably the largest copyright violator of them all. I've found not only youtube-dl backups but actual full movies through a Google search. RIAA simply targets anyone who is not financially beneficial to them. If they were just about protecting published works they wouldn't go after projects without a legal teams, they'd go after anyone facilitating misuse of copyright.

Stallman uses the term "copying not approved by the publisher" and not piracy, I thought that was just neckbeard logic but I realize how stupid I've been. Framing it as piracy already assumes that you're doing something wrong and a genius move by MPAA/RIAA. It is similar to how Breonna Taylor lived in a suspected "drug house" (though I noticed the media quickly stopped using that term), which already frames the conversation in a way where the accused have to bizarrely prove they didn't do anything wrong then the other way around.
posted by geoff. at 2:56 PM on October 25, 2020 [15 favorites]


Ctrl-F circumvent—yep, the whole circumvention device bit in the DMCA is a fusillade that's been coming down the barrels at us for a quarter-century now.

To lay out the basic context for anyone who needs it, youtube-dl is a piece of software which, while it can do a variety of things, at its core is something which lets you watch a Youtube video or videos from other web sites. So it's a fundamental building block of many other pieces of software that let you watch videos like the one CrystalDave mentions. But the 1998 Digital Millennium Copyright Act is a U.S. law that governs aspects of copyright and, among other things, allows companies to control exactly how you consume copyrighted material, and to declare a piece of electronics or software an illegal circumvention device if it allows people to consume content in an unauthorized way.

youtube-dl, in addition to simply letting you watch a Youtube video, can save a copy of a video. So the reaper's scythe swings down. (Of course, the phone in your pocket can also save a copy of a Youtube video if you take it out and point it at a computer or another phone's screen while the video is playing, but why all smartphones don't also fall afoul of the DMCA, or why only an infinitesimally small percentage of the copyright owners of Youtube content will ever see any cash for the licensing fees that would supposedly justify the disparity, is left as an exercise for the reader.)

This kind of thing always reminds me of how, in the only-slightly-more-corporate-dominated future world of the 1980s film and television series Max Headroom it was illegal for televisions to have off switches.

Happy Halloween!
posted by XMLicious at 2:59 PM on October 25, 2020 [13 favorites]


As annoying as it would be the dev's should use a distribution model like Calibre where there is a general plugin interface. Then if some crazed radical developer gave a link to an isolated bit of code to enable backups of certain types of files then that small bit would not disable a full package.

For the non-devs, the code is in python, every downloaded version is ready to be sent anywhere and used on any computer that can run python. It's not going away. This is the slightest speedbump, mostly imperceptible for most folks.
posted by sammyo at 3:03 PM on October 25, 2020 [3 favorites]


I mean, on the surface: youtube-dl exists because youtube is rife with illegal and grey-market movie and tv show uploads. You can blame a service that uses the affordances of the internet or you can take action on the company that doesn’t police its service correctly.
posted by Going To Maine at 3:05 PM on October 25, 2020 [2 favorites]


The legal arguments against such tools require folks to not understand technology at all.

Under the old broadcast paradigm, you could broadcast a signal to TV or radio and a device would convert it to entertainment. Saving a copy was adjacent to broadcast.

Under the streaming media paradigm, a user necessarily must download a copy to their computer (or other device which is itself of course a computer), and said copy is already saved to either RAM or the hard drive (as a cache). "Saving a copy" is simply a matter of telling your computer to NOT delete the downloaded copy.

There is no way to stream media without letting users download it (aka "save it"), so I'm always really skeptical of corporations' attempts to control their media. They're inevitably either applying the law wrong, or trying to set law precedent that will have horrible repercussions.

Edit: I mean, on the surface: youtube-dl exists because youtube is rife with illegal and grey-market movie and tv show uploads.

youtube-dl is the tupperware for your leftovers because sometimes videos are very long and you want to save the video to watch later. You've already made the meal (downloaded the video), you're just using a tool to keep it fresh and edible.
posted by explosion at 3:08 PM on October 25, 2020 [22 favorites]


lets you watch a Youtube video or videos from other web sites.

Well, lets you download a video from those websites. The downloading is key, the watching not so much.
posted by Going To Maine at 3:09 PM on October 25, 2020


It's not going away. This is the slightest speedbump, mostly imperceptible for most folks.

Eh, I wouldn't be so sure. Source control access is quickly being centralized into a few large tech companies. While it is trivial for me to host the source control elsewhere the real value is the ease and reliability of getting to the code and to the developers an easy to use interface to interact with other developers and contributors via pull requests, issues and things that make GitHub, GitHub. Saying the source code of the project is all that was taken down really doesn't do the GitHub platform justice. It is getting worse (better?) the more OSS utilizes propietary GitHub features like workflow Actions. Again, yes, you can recreate those but the higher the barrier to entry the more you alienate all but the biggest projects. It is the long tail that allowed super popular OSS like wget, Linux and the like to survive but would kill things like youtube-dl.

To be honest if I go back to the pre-GitHub hell of SourceForge + maybe a poorly written website + newsgroup to figure out what needs to be done to install something as simple as youtube-dl I'll probably just recreate the functionality and not have as well done an end product.
posted by geoff. at 3:10 PM on October 25, 2020 [2 favorites]


The downloading is key, the watching not so much.

Seriously, what's the difference? Whether I stream something or download it? That a handle to a resource is closed before reading something? I mean it is "downloaded" in memory or a persistent cache and that seems rather more of a technical detail. If I created a giant YouTubeContext object in memory would that be different? We could take out all IO functionality in youtube-dl and let the reader implement anything that actually saves.
posted by geoff. at 3:12 PM on October 25, 2020 [4 favorites]


Yeah I tried to upgrade via homebrew and got:

==> Upgrading 1 outdated package:
youtube-dl 2020.03.24 -> 2020.09.20
==> Upgrading youtube-dl 2020.03.24 -> 2020.09.20
==> Downloading https://github.com/ytdl-org/youtube-dl/releases/download/2020.09
#=#=#
curl: (22) The requested URL returned error: 451 Unavailable for Legal Reasons


That's a first.
posted by gwint at 3:16 PM on October 25, 2020 [19 favorites]


The downloading is key, the watching not so much.

Seriously, what’s the difference?

Well, in this case what I mean is that youtube-dl doesn’t incorporate any viewing tools - it extracts a file from a URL. You could conceivably just use youtube-dl to make a private archive of YouTube and not view any of the content. It has no function other than to divorce a file from its original context and save it to disk.
posted by Going To Maine at 3:23 PM on October 25, 2020 [2 favorites]


The "pre-github hell" of installing youtube-dl was: sudo apt-get install youtube-dl

(Github was founded in 2008 and youtube-dl became available that way in 2006)
posted by joeyh at 3:23 PM on October 25, 2020 [12 favorites]


The website itself is still up (hopefully their host won't come down on them).
posted by trig at 3:25 PM on October 25, 2020 [3 favorites]


Fundamentally, this is about controlling the flow of money from of the work of artists into the coffers of the corporations that hold the copyrights. The music industry was upon its inception a criminal enterprise, and remains so. The money comes from somewhere, who knows where now given the streaming revenue, and usually circumvents the artists altogether. Anything that can be done to eliminate corporate control of them music industry should be done. They are worse than useless for the artists they exploit.
posted by hilberseimer at 3:46 PM on October 25, 2020 [3 favorites]


For those not familiar with the tool, don't let the name fool you. It might have been born for downloading things from YouTube, but it supports countless sources now. I use it all the time for work purposes: rarely ever with YouTube and never with copyright restricted content. It isn't as important as ffmpeg, but it's right up there among the swiss army knives of modern media content tools.
posted by bcd at 3:49 PM on October 25, 2020 [32 favorites]


A quick look at the source suggests my 'countless' above is currently 766 different extractors, which translates as different categories of sites. Many different sites often share the same underlying engine.
posted by bcd at 3:53 PM on October 25, 2020 [7 favorites]


I too use youtube-dl to download video content to watch when my wireless speed is too slow for streaming to be watchable. So very sucky news.

But won't the makers just find alternative sites to be host? Or will the RIAA scare any site out of it?
posted by bertran at 4:29 PM on October 25, 2020 [2 favorites]


Know how I know that I want this?

The RIAA is suing them.

This means that, unlike umpteen tools for capturing YouTube content that I've tried, this one WORKS.
posted by delfin at 4:29 PM on October 25, 2020 [24 favorites]


I should say: Github has categorized the RIAA letter as a DMCA notice, but it's not really — that term generally describes communications sent pursuant to §512. This letter is about the §1201, the controversial anticircumvention rules.

The closest example to this RIAA/youtube-dl letter is the one MPAA sent Github to try to shut down Popcorn Time back in 2014. That one cited §512, but relied under the hood on a fairly extraordinary tertiary liability theory
posted by geoff. at 4:41 PM on October 25, 2020 [1 favorite]


The supportedsites.md file in the source code is about 1200 lines long, so, yeah, not just for youtube.com. Although a lot of those sites look really sketchy.
posted by jabah at 4:42 PM on October 25, 2020


Well, in this case what I mean is that youtube-dl doesn’t incorporate any viewing tools - it extracts a file from a URL. You could conceivably just use youtube-dl to make a private archive of YouTube and not view any of the content. It has no function other than to divorce a file from its original context and save it to disk.

In 1998 web browsers did not incorporate video viewing tools either, nor did they when Youtube launched in 2005: they used “plugins” for video, in browsers and platforms where that actually worked; but the browser itself ran just fine on Windows NT or Linux without any video support because it wasn't integral to the browser.

youtube-dl has a parameter --exec for calling a video player through the command line environment they share, just like browsers would call a plugin through the interfaces they shared with the plugin. Or for example the Linux manpage contains a section “How do I stream directly to media player?”

The distinction you're trying to make is just as hair-splitting and artificial as the ones the DMCA tries to make, and does not work logically any better. How exactly is any other method of watching a Youtube video less “divorcing” from the “original context” of a video file in a replicated fault-tolerant file system on a load-balancing caching server, or perhaps a value in an exotic binary datatype column in an Oracle database, or whatever Youtube uses on the back end, than youtube-dl and an exec'd file viewer? Viewing it on an ARM-powered smartphone via a buffer in flash memory, or on a Roku and television set or something, actually seems more distant from the original context. (And if you were trying to imply by “context” something like “you have to see all the brilliant comments and web page stuff at the same time”, of course you don't see those things on a Roku or in phone apps either.)

Okay, maybe “just fine” is a slight exaggeration but you know what I mean.
posted by XMLicious at 4:50 PM on October 25, 2020 [4 favorites]


This means that, unlike umpteen tools for capturing YouTube content that I've tried, this one WORKS.

It works and works well (and is open-source and non-sketchy and works for a million other sites), but since youtube changes things constantly on its own end, yt-dl also needs frequent updates to keep up.
posted by trig at 4:53 PM on October 25, 2020 [2 favorites]


The distinction you're trying to make is just as hair-splitting and artificial as the ones the DMCA tries to make, and does not work logically any better.

I -- ok? I'm not sure why arguing that because youtube-dl lets you execute some code on a downloded file the distinction between something that downloads something and something that lets you view media is meaningless.

This is a piece of software that, on its website, characterizes the software as “a command-line program to download videos from YouTube.com and a few more sites. ” I’m not a lawyer, but once I took a class on IT law by a copyright lawyer. His general take on these fine-grained logical distinctions that, because you moved bits in just such a way is that they were delightful for CS majors but they cut no ice in courts because judges are not idiots and know when you are fucking with them. That’s greatly to the judges’ credit, I think.
posted by Going To Maine at 5:13 PM on October 25, 2020 [2 favorites]


is that they were delightful for CS majors but they cut no ice in courts because judges are not idiots and know when you are fucking with them. That’s greatly to the judges’ credit, I think.

I agree with your point and I'm continually impressed by lawyers ability to distill complex technical topics and not be swayed by smart people. Part of this is the thinking like a lawyer they teach you in law school that distills events to the relevant facts.

That said this isn't just flipping bits. This is more like Sony v Columbia in limiting on what I can do because it might possibly might not be the way a publisher wants me to consume it. Besides the very valid reason of content creators downloading their own content (!), there's also ISP caching. I'm guessing per the DMCA request this is also illegal which is a dynamic cache for squid. Looks like it doesn't work well but there's plenty of people who don't have unlimited bandwidth. It seems like youtube-dl's biggest violation right now is being called something straightforward and not "dynamic content adaptation cache for a squid proxy."
posted by geoff. at 5:47 PM on October 25, 2020 [3 favorites]


The point that content creators sometimes need to use youtube-dl to download their own content is something that surprised me until I had to help someone figure out how to save a local copy of video they had live-streamed themselves to Facebook. It certainly appears youtube-dl is the easiest way to do that.
posted by bcd at 6:02 PM on October 25, 2020 [1 favorite]


Doc Edward Morbius has published some more details on this: The RIAA’s fraudulent attack on youtube-dl is not a DMCA §512 infringement/safe-harbour, and the reality is weird.
posted by Lanark at 7:07 PM on October 25, 2020 [8 favorites]


I'm not sure why arguing that because youtube-dl lets you execute some code on a downloded file the distinction between something that downloads something and something that lets you view media is meaningless.

Yeah, I guess we are just worlds apart on this if you really think your logic holds up to scrutiny, because it seems pretty obvious to me that the distinction is basically meaningless in the realm of concepts like to divorce a file from its original context, much less anything like the copyright clause of the constitution which the DMCA and other copyright laws would ostensibly be derived from.

(...to allow, by the way, the rights of the RIAA as copyright holders to in practicality extinguish the otherwise-copyright-protected rights of the youtube-dl authors to publish their work, through what appear to me to be a Rube Goldberg contraption of concepts like the aforementioned one accreted over the last century or so, at some point involving player piano rolls. I'm not a lawyer myself, and haven't taken the one class on IT law either, so I'd welcome specific explanation and citations from a legal point of view, but no matter how deep I've ever dug into this I never get to a point where it coherently makes sense: it remains looking like a Divine-Right-of-Kings-esque, inconsistently-applied and inconsistently-enforced set of pretexts that, like so many other things, work out to these people get all the money and the power, and these people don't.)

In most of the media viewers I've used in the twenty-first century (most of which are using operating system components that directly interface with video hardware to go the last mile to render video at this point in history anyways, rather than doing it themselves, as I understand it) you can hit Open in a menu and paste a file URL into a dialog that appears and the program will engage libraries and OS components to do the downloading; and in most utilities that focus on downloading there will be a hook like --exec to engage a media player at the end of the download if you want.

(Also... for anyone who hasn't traced through the kinds of things youtube-dl does in the download part of the process, streaming pretty often just means downloading a text file that contains a bunch of file fragment URLs first, then downloading those fragments. The media and infrastructure companies will try to obfuscate this by doing HTTP redirects and breaking the digest text files into pieces with their own separate redirects, and mix in browser cookies and tokens handed out by the server in its original URL or which have to be calculated in javascript, so that everything breaks if it's not done perfectly; but that's really all that's happening and dealing with the obfuscating crap is what youtube-dl accomplishes.)

So no, all of these word games that arrive at youtube-dl somehow being the illicit controlled substance of the copyright world, while your phone or an open-source browser with an open-source video plugin (for the moment) are the government-approved tobacco and alcohol, do not seem legit to me.

On preview: ooh, some juicy stuff in the link Lanark posts.
posted by XMLicious at 7:20 PM on October 25, 2020 [1 favorite]


curl: (22) The requested URL returned error: 451 Unavailable for Legal Reasons

That's a first.


The status code number is some A+ shade, though.
posted by aws17576 at 7:56 PM on October 25, 2020 [28 favorites]


Fahrenheit 451 reference to boot!

Is there a way to incorporate GitHub issues and various errata into the source itself? Or is there a way of standardizing GitHub issues and comments into something that can be stored in git? Or a git extension? Two repos storing separate issues would be a huge merge headache, but beyond that simply moving youtube-dl's source to another remote repository doesn't really help now that source control itself has become a lot more rich since git was created.

I guess really only issue and action are the things OSS projects are using and I was worried for awhile now that GitHub might create an environment that is so proprietary and essential we'll run into a situation where the underlying git while open source and free is just a small fraction of what make up a project.

To be honest I don't blame GitHub or Microsoft, anyone getting a legal notice from someone as powerful as RIAA is likely to behave the same way. We should have ways to contribute to the open source community and not be beholden to GitHub. I'd feel better if GitHub just turned into a pretty interface with some social features. It isn't youtube-dl I'm worried about but the myriad of projects that have very few users and that might get removed that no one will care about.
posted by geoff. at 8:06 PM on October 25, 2020


So no, all of these word games that arrive at youtube-dl somehow being the illicit controlled substance of the copyright world, while your phone or an open-source browser with an open-source video plugin (for the moment) are the government-approved tobacco and alcohol, do not seem legit to me.

Speaking more generally, the anti-circumvention measure, is viewed by RIAA lawyers et al as a simple one - do you end up with unencrypted copies of 'their' music at the end of it?

So that's why they issued this notice, they see people using youtube-dl to grab copies of music tracks and/or music videos, and *somebody* has to be punished.

Of course, the anti-circumvention provision of the DMCA is written (effectively by RIAA lobbyists and the like), and intended, so that they could go after distributors of tools that could break the various harebrained encryption schemes they dreamed up in the 90s and 00s; Sony's Minidisc, XCP/accidental rootkit, iTunes DRM, HDCP etc etc

In 2020, all such DRM schemes have largely been dropped because a) they don't work, b) they cost money and c) users overwhelming prefer music they can download - legally - and put on any device they currently own without jumping through giant hoops.

The remnents (outside video gaming) are largely just streaming obfustication as you see in, for example, youtube. Now, to a software engineer, the difference between a browser and something like youtube-dl is effectively nil, because they both jump through the same hoops to access the same content, from the same source, legally.

To a RIAA lawyer, one gets recorded on youtube as a view, has adverts applied to it*, and doesn't leave the user with an unencrypted copy to do with as they wish; youtube-dl does not. That such potentially infringing behaviour lies with the user, not the tool, doesn't matter.

As the DMCA says though, the following would need to apply to youtube-dl - "has only limited commercially significant purpose or use other than to circumvent a technological measure that effectively controls access to a work protected under this title". Youtube's obfustication is not DRM, i.e. an effective technical measure, as interpreted by the courts - it's mainly been applied to things like mod-chips for consoles. From memory, compatible 3rd party garage door openers and compatible printer cartridges both won when the same tactic was applied against them for merely reverse-engineering a basic communication scheme. It also seems youtube-dl does other have other substantial uses that don't involve protected works at all.

It's also very questionable that the RIAA labels even have standing, as opposed to youtube themselves.

So it's unlikely the RIAA labels would win. However, the intent is likely to be to use the threat of defending a costly court case to get them to self-censor themselves off the internet without that messy business of actually proving youtube-dl did anything wrong. Hopefully they will not succeed.

* TV networks sued Dish a decade ago, on the basis that the ability to skip adverts was 'stealing', and a tool to do so was contributory infringement. Dish eventually settled, and removed the functionality.
posted by Absolutely No You-Know-What at 8:09 PM on October 25, 2020 [6 favorites]


Well, in this case what I mean is that youtube-dl Betamax doesn’t incorporate any viewing tools - it extracts a file from a URL broadcast. You could conceivably just use youtube-dl Betamax to make a private archive of YouTube broadcast tv and not view any of the content. It has no function other than to divorce a file broadcast from its original context and save it to disk tape.

IE it is defacto time shifting which was determined to be non-infringing and fair use way back when the tool dejure was Betamax. For the youngins out there Betamax was an analogue tape standard that let one play and record movies on analogue TVs. It lost out to VHS (another analogue video tape standard) which was itself superseded by DVD and then Blu-Ray.
posted by Mitheral at 8:44 PM on October 25, 2020 [10 favorites]


Tim Bray's RFC 7725, introducing HTTP response code 451: Unavailable For Legal Reasons.

Interesting side note on this, given that the 4xx class is for client-side errors/etc. and the 5xx error class is for server-side errors/etc.:
“This is considered a client-side error even though the request is well formed and the legal requirement exists on the server side. After all, that representation was censored for a reason. There must be something wrong with you, citizen.” -RESTful Web APIs, Leonard Richardson & Mike Amundsen
posted by brainwane at 9:30 PM on October 25, 2020 [10 favorites]


I mean, on the surface: youtube-dl exists because youtube is rife with illegal and grey-market movie and tv show uploads.

Wouldn’t it be more appropriate to say that youtube-dl can be construed as a piracy tool primarily because YouTube is rife with piracy? I think the premise that it exists for that purpose is rather dubious. It would be a useful tool no matter what kind of content was on YouTube.

Of course it can presumably also be used to download, say, music that record labels put up themselves for streaming. I don’t have a very strong grasp of the legal nuances of saving stream-on-demand media that doesn’t have real copy protection but does have some tricks making it difficult to save.
posted by atoxyl at 9:52 PM on October 25, 2020


Wouldn’t it be more appropriate to say that youtube-dl can be construed as a piracy tool primarily because YouTube is rife with piracy? I think the premise that it exists for that purpose is rather dubious. It would be a useful tool no matter what kind of content was on YouTube.

Is YouTube rife with piracy? Do we know this? This is not something I think even a reasonably informed person can determine. We know that there's a level of bulk analysis of YouTube videos to determine copyright attribution. YouTube passively collects analytics and serves up ads based on their own algorithm in determining when we should be shown ad pre-rolls. Even ad supported content has in-video product placement at times. We have no idea what revenue sharing agreements are setup with YouTube or what copyright holders consider proper use of their content.

Schindler's List was shown without advertising on network television many years ago and watching that was not privacy.

I'm not being willfully obtuse, determining copyright infringement is a complex topic with legal use way beyond buying a ticket or physical media.
posted by geoff. at 10:12 PM on October 25, 2020 [2 favorites]


Count this another recommendation for that article posted by Lanark. It covers all the reasons why this action is not valid. (Which sadly doesn't mean it will be squashed.)
posted by bcd at 10:21 PM on October 25, 2020


IE it is defacto time shifting

This would be a more compelling argument if YouTube didn’t have five nines of uptime and wasn’t available everywhere.
posted by Going To Maine at 12:23 AM on October 26, 2020


His general take on these fine-grained logical distinctions that, because you moved bits in just such a way is that they were delightful for CS majors but they cut no ice in courts because judges are not idiots and know when you are fucking with them. That’s greatly to the judges’ credit, I think.

I just read this again and noticed—it seems backwards to me, Going to Maine, though maybe I'm misunderstanding you. I'm arguing that the way the bits are moving from Youtube to your computer is essentially the same whether you use youtube-dl or another method to view a video, and that this is why it makes no real sense to say that viewing the video is divorc[ing] a file from its original context in one case but not the others.

Whereas from my point of view, it seems like you and the RIAA are the ones saying that when you compare a video being downloaded to view in a web browser or on a Roku, etc. to downloading through youtube-dl with --exec 'vlc {}' for example, you can tell that the bits move in just such a way that it's illegal in the latter case but not in the former cases.
posted by XMLicious at 1:53 AM on October 26, 2020 [2 favorites]


I wonder how much of YouTube itself runs on youtube-dl.
posted by rhamphorhynchus at 2:03 AM on October 26, 2020


So my grandfather died in the 1990's, a miserable affair altogether, delirious with brain cancer in a dirt-poor country in the tropics. It is said they found him in his home, surrounded by lottery tickets. The family organized a burial and had a recording made for the relatives in Europe. I got the video tape, as well as some poems that my grandfather also wrote, delirious with brain cancer, surrounded by lottery tickets.

A while ago my dad asked about the video. So I digitized the tape and uploaded it to YouTube. It was hard to watch it again after all this time. The quality was atrocious, all smudged blotches of color and tape warble, like what you'd expect from a decades old video tape from a dirt-poor country in the tropics. The burial itself was possibly even more dispiriting. A decrepit building, a lackluster ceremony, a hole in the ground. There was a small band, playing renditions of various spirituals. It wasn't a good band, and it would have sounded bad even without the tape warble. While they played, the family said their goodbyes. They were so far away, decades in time, thousands of miles, and painfully, so much poorer. So much was stolen from them. I felt guilty and angry, and so many things. But at least it was on YouTube. It wouldn't be forgotten.

I sent the video link to my dad. He said he couldn't see it. So I check it out. Turns out there's a problem with the video. Apparently one of the songs played at my grandfather's funeral, one of these dirt-poor, warbling renditions for dented horn, somehow triggered their copyright detection algorithm. I have to check a box saying I can't "monetize" the video before they will show it. It feels like another thing stolen from my family, a plunder over my dead grandfather's body.

Copyrights, IP, I get it. Reward risk, encourage culture, necessary evil, fine. But also honestly fuck these vultures.
posted by dmh at 4:36 AM on October 26, 2020 [24 favorites]


In the past, devices and/or programs that have fallen afoul of the circumvention device provision of the DMCA have been things like DeCSS or console mod chips, whose entire reason for existence is to bypass some sort of DRM and have little actual use for other purposes.

That is not the case with youtube-dl, which does not strip DRM and in fact is often used for non-infringing purposes, many of which have been outlined in the comments on this here post. Those substantial non-infringing uses should invalidate any attempts to stop the distribution of youtube-dl, though it doesn't keep RIAA from suing its authors for money damages.
posted by wierdo at 4:54 AM on October 26, 2020 [3 favorites]


This would be a more compelling argument if YouTube didn’t have five nines of uptime and wasn’t available everywhere.

It doesn't matter how reliable youtube is if I don't have five 9s of internet connectivity. Heck I don't have one 9. There are lots of places where always on internet fast enough to stream youtube isn't available or isn't affordable.

Heck a classic use even for the insanely privileged would be to time shift youtube for viewing on an airplane.
posted by Mitheral at 4:56 AM on October 26, 2020 [17 favorites]


I wonder how much of YouTube itself runs on youtube-dl.

I'm not sure how serious a question this was intended to be, but none at all.
posted by bcd at 6:00 AM on October 26, 2020


I wonder how much of YouTube itself runs on youtube-dl.

I think what is meant by that is, given how often copyright content is taken down, or entire Youtube channels removed, most original uploads will eventually disappear. But if someone has downloaded a copy with youtube-dl, then it can be re-uploaded (hopefully with little loss in quality) by someone else. This happens A LOT.
So ironically one of the biggest losers should youtube-dl be permanently outlawed would be YouTube.

If the only content left was official Movie trailers and official VEVO music videos then YouTube would be a pale shadow of its current state.
posted by Lanark at 6:15 AM on October 26, 2020 [1 favorite]


If the only content left was official Movie trailers and official VEVO music videos then YouTube would be a pale shadow of its current state.

And the RIAA/MPAA would finally be able to rest on their giant piles of money with a contented smile (unlike now where they're scowling from atop their money piles).
posted by Dysk at 7:25 AM on October 26, 2020 [2 favorites]


Probably none of YouTube is dependent on youtube-dl.

But I would bet my pants that the RIAA and MPAA have technical teams that use it for their own content analysis.
posted by at by at 7:29 AM on October 26, 2020 [9 favorites]


@geoff. Is there a way to incorporate GitHub issues and various errata into the source itself? Or is there a way of standardizing GitHub issues and comments into something that can be stored in git?

There's another distributed version control system called Fossil (https://www.fossil-scm.org) which adds bug-tracking to the repository itself. There are also other Git repository hosting systems like Gitlab (https://gitlab.com) and Sourcehut (https://sourcehut.org) that are themselves open-source projects that you can host yourself.
posted by technodelic at 9:57 AM on October 26, 2020 [1 favorite]


The time-shifting argument is more complicated than I’d first allowed, it is entirely true.
posted by Going To Maine at 10:38 AM on October 26, 2020 [3 favorites]


Stallman uses the term "copying not approved by the publisher" and not piracy, I thought that was just neckbeard logic but I realize how stupid I've been.

I usually say download-sharing myself. It just feels wrong to use the word piracy, first because obviously you aren't stealing anything, you're making free copies (which anyone should be able to do with anything that can be copied, full stop, period, forever). Second, there are still actual pirates in this world doing piracy. Their activities are not comparable in the least.
posted by GoblinHoney at 11:13 AM on October 26, 2020 [5 favorites]


Music industry forces widely used journalist tool offline
Numerous reporters told Freedom of the Press Foundation that they rely on youtube-dl when reporting on extremist or controversial content.
Also, there are several video players that automatically use youtube-dl to stream video from youtube. For example "mpv https://youtube.com/whatever" automatically uses it. A user of one of these video players can easily use them this way with no intention or act of copyright infringement.
posted by joeyh at 2:09 PM on October 26, 2020 [3 favorites]


Is YouTube rife with piracy? Do we know this? This is not something I think even a reasonably informed person can determine.

I was granting another user’s premise in disputing their argument from that premise. Whether YouTube is actually “rife with piracy” depends largely on what “rife” means, and maybe a little bit on what “piracy” means. There’s certainly no shortage of technical IP violations, though, I’m comfortable asserting that.
posted by atoxyl at 4:48 PM on October 26, 2020




Is YouTube rife with piracy? Do we know this? This is not something I think even a reasonably informed person can determine.

“rife” is my word, and of course that‘s subjective. Let’s put it this way: YouTube has enough uploads of grey-market TV shows, movies, and other content that I am guaranteed easy access to entertainment under copyright any time I would log in. I would hazard that there is more grey market video content than there was/is Q anon and Nazi content, simply because there’s more demand for obscure video content than there is Nazi content. (There’s more than 500 hours of video uploaded to YT every minute. It’s rife with a lot of things.)
posted by Going To Maine at 5:16 PM on October 26, 2020


Music industry forces widely used journalist tool offline
“I have also used it to secure a good quality copy of video content from Youtube, Twitter, etc., in case the content gets taken down when we start reporting on it.” Skille pointed to a specific instance of videos connected to the terrorist murder of a Norwegian woman in Morocco. “Downloading the content does not necessarily mean we will re-publish it, but it is often important to secure it for documentation and further internal investigations.”

Similarly, a number of reporters described using youtube-dl for nuts-and-bolts workflows such as transcribing videos they’re covering. Jeremy Gray, a data scientist with The Globe and Mail, described a Slack tool he provides to journalists to allow them to automatically transcribe their own interviews and, until Friday, to transcribe Youtube videos from a URL. “It used youtube-dl, and now that part is broken.” Another journalist, who works at a “small-ish public media newsroom,” described a common situation where a reporter needs “a recording of a public meeting for a story but is on deadline and doesn’t want the hassle of recording the parts they want it in realtime or wants the full file for something like AI transcription.”
posted by geoff. at 7:23 PM on October 26, 2020 [5 favorites]


Privately-hosted gitea servers may not be as discoverable as github in 2020, but they do keep on ticking.
posted by rum-soaked space hobo at 1:20 PM on October 27, 2020


RIAA’s YouTube-DL Takedown Ticks Off Developers and GitHub’s CEO
Hosting service Uberspace was one of the recipients. The company hosts the official YouTube-DL site and still does so today. Instead of taking the website down, Uberspace replied to the notice through its own lawyer, who said that the hosting company hasn’t don’t anything wrong.

“The software itself wasn’t hosted on our systems anyway so, to be honest, I felt it to be quite ridiculous to involve us in this issue anyway – a lawyer specializing in IT laws should know better,” Jonas from Uberspace says.

Meanwhile, GitHub’s CEO Nat Friedman wasn’t sitting still either. While the Microsoft-owned developer platform had to respond to the takedown notice, Friedman himself actively reached out to YouTube-DL’s developers to help them get their project reinstated.

The CEO joined YouTube-DL’s IRC channel hoping to connect with the owner of the repository so he can help to get it unsuspended.

“GitHub exists to help developers. We never want to interfere with their work. We want to help the youtube-dl maintainers defeat the DMCA claim so that we can restore the repo,” Friedman told TorrentFreak, explaining his actions.
posted by geoff. at 4:36 PM on October 27, 2020 [4 favorites]


Of course the lawyer knows better; it is classic barratry of the sort the RIAA has engaged in off and on for decades in an attempt to scare the people who buy their product writ large.
posted by Mitheral at 6:58 PM on October 27, 2020 [2 favorites]


Having stonewalled the internal and external objections to Github's contract with ICE for over a year, Nat Friedman must have been desperate to find a chance to play the hero.

https://www.latimes.com/business/technology/story/2019-12-04/github-open-source-developers-ice-contract
posted by polytope subirb enby-of-piano-dice at 8:24 PM on October 27, 2020


v2020.11.01.1 available, youtube-dl -U works again
posted by Bangaioh at 9:55 AM on November 1, 2020 [2 favorites]


brew update && brew upgrade also working again.
posted by Going To Maine at 10:59 AM on November 1, 2020 [2 favorites]


Of course the lawyer knows better; it is classic barratry of the sort the RIAA has engaged in off and on for decades in an attempt to scare the people who buy their product writ large.

Then track down the lawyers involved or could have touched this and bargrieve them.
posted by rough ashlar at 9:00 AM on November 2, 2020


Some folks on reddit were saying YouTube-dlc is a current fork that might be available in places (honestly, not being coy, I know nothing more) to address server-side changes at YouTube.

I guess the source code is too voluminous to fit on a t-shirt, à la DeCSS.
posted by exogenous at 3:58 PM on November 2, 2020 [1 favorite]


Big thanks to gwint and Going To Maine for mentioning homebrew; I use Debian, which in its most recent major release disabled the youtube-dl update mechanism for security reasons, and I've been limping along using youtube-dl via an older server because I didn't get around to installing it to my home directory.

But I just found a package linuxbrew-wrapper and installed it, made some recommended .profile changes from the manpage, ran brew once and followed some further setup instructions, and now I'm off to the races. (evidently homebrew itself supports Linux now, hence “wrapper” I guess; I shuddered to see a Linux package dependency ruby-net-telnet installed but I'm going to pretend I didn't notice.)
posted by XMLicious at 1:19 AM on November 9, 2020 [2 favorites]


On my raspberry pi running Debian buster, I just installed the old, broken version in the buster repository with apt to get all the dependencies, then downloaded the latest version from here which I run from my home directory (since I am too lazy to do it a better way).

Anyway, it is still available, including source (just not from github).
posted by Dysk at 8:02 AM on November 9, 2020 [1 favorite]


"Origins of the youtube-dl project" by its founder. "All good stories need at least a villain so I have arbitrarily chosen copper thieves as the villains of the story that set in motion what youtube-dl is today."
posted by brainwane at 11:24 AM on November 10, 2020 [3 favorites]


Github blog: Standing up for developers: youtube-dl is back
As we explained, the key claim in the youtube-dl takedown is circumvention. Although we did initially take the project down, we understand that just because code can be used to access copyrighted works doesn’t mean it can’t also be used to access works in non-infringing ways. We also understood that this project’s code has many legitimate purposes, including changing playback speeds for accessibility, preserving evidence in the fight for human rights, aiding journalists in fact-checking, and downloading Creative Commons-licensed or public domain videos. When we see it is possible to modify a project to remove allegedly infringing content, we give the owners a chance to fix problems before we take content down. If not, they can always respond to the notification disabling the repository and offer to make changes, or file a counter notice.

That’s what happened in this case. First, we were able to reinstate a fork of youtube-dl after one of the fork owners applied a patch with changes in response to the notice.

Then, after we received new information that showed the youtube-dl project does not in fact violate the DMCA‘s anticircumvention prohibitions, we concluded that the allegations did not establish a violation of the law. In addition, the maintainer submitted a patch to the project addressing the allegations of infringement based on unit tests referencing copyrighted videos. Based on all of this, we reinstated the youtube-dl project and will be providing options for reinstatement to all of its forks.

posted by Catblack at 9:26 AM on November 16, 2020 [4 favorites]


« Older An "Arkell v. Pressdram" for our time   |   "The first step is recognizing that conspiracies... Newer »


This thread has been archived and is closed to new comments