Video killed

May 21, 2008 5:43 AM

YouTombMIT project that tracks youtube file deletions for aledged copyright infringement. They do not host the deleted files, fyi.via wired

Alongside a screenshot of each clip deemed in violation, YouTomb lets users see who posted the offending video, how many views it got before being pulled, when it was removed and by whom (for instance at the request of the user, a media company or third-party).

YouTomb, which launched about two months ago, is currently monitoring close to a quarter-million videos, and Jansen says the team is eager to expand its scope. Currently, the site only monitors popular YouTube videos, but Jansen hopes to span as many YouTube data sets as possible.

YouTomb is considering offering code to bloggers to use when embedding YouTube videos. That way, when videos were taken down, users could see stats and data instead of a simple error message.

posted by asok (16 comments total)

Great project! But how can they tell who requested the removal? I guess that must be on the page, but given the "site's opaque policies" what's to keep YouTube from simply not reporting that anymore?
posted by DU at 5:51 AM on May 21, 2008

This looks like an interesting (and perhaps quite valuable) reseach project... until I read the details:

"We're not interested in bootlegged videos of [anime show] Naruto that got taken down," said Jansen in a phone interview Tuesday. "It's about fair use, and remixes or performances forced off YouTube."

So the conclusions have already been written: "fair use" is being trampled by YouTube. It's just a matter of hacking the code until the data supports it.

Pity to see someone so young with such a closed mind.
posted by three blind mice at 6:02 AM on May 21, 2008 [1 favorite]

... alleged. It's all alleged.
posted by WalterMitty at 6:09 AM on May 21, 2008

But how can they tell who requested the removal? I guess that must be on the page, but given the "site's opaque policies" what's to keep YouTube from simply not reporting that anymore?

Often if a corporation throws a CI tantrum YouTube's way, they'll note it on the page. Usually along the lines of "This video has been removed by request of Viacom."

That's when you get to shake your fist at the screen and go "Viacoooooooom!!"
posted by Spatch at 6:27 AM on May 21, 2008

So the conclusions have already been written: "fair use" is being trampled by YouTube. It's just a matter of hacking the code until the data supports it.

Huh? How do you get that? They're not interested in blatant violation of copyright (putting up entire movies), they're interested in fair use. You can quote passages of reasonable length from copyrighted sources when you write; you should be able to do the same with video. As they say:

While many YouTube videos that contain non-original material are blatantly violating copyright (e.g. exact rips of TV shows), many others have a more complex legal status because of the fair use provision of copyright law. The sampling and remixing of non-original material have often led to great cultural accomplishments, so protecting this fragile aspect of copyright law is very important to us.

It's as if you were to say of a study of killings by cops: "Hah! They're deliberately ignoring all killings that aren't by cops! The conclusions have already been written!"
posted by languagehat at 6:27 AM on May 21, 2008 [6 favorites]

Huh? How do you get that?

By reading the post.

He's looking for something specific instead of simply collecting take-down data on a broader basis. Going into it with the attitude that "It's about fair use, and remixes or performances forced off YouTube" suggests to me that this project has a pre-ordained trajectory which makes it a lot less interesting as a research project.

At the very least Jansen's point of view should be: "we don't know what to expect, we're keeping an open mind, we'll look at the data and see what we find. Maybe we'll learn something."

A very dangerous point of view to have if you want to work for Google.....
posted by three blind mice at 6:41 AM on May 21, 2008

What's more interesting is who doesn't take stuff down. Like Fox seems to have no problem with full length American Idol performances being up.
posted by smackfu at 6:55 AM on May 21, 2008 [1 favorite]

He's looking for something specific instead of simply collecting take-down data on a broader basis.

Well, yeah. Aren't most researchers "looking for something specific"? They're looking for takedowns of material that seem like fair use; if there aren't any, they'll learn something, and if (as is pretty much certain) there are many such takedowns, that's important information. I still don't understand your objection.
posted by languagehat at 7:03 AM on May 21, 2008 [4 favorites]

"We'll look at the data and see what we find" is, sixth-grade science aside, a terrible approach for any sort of serious research. Scientists who set out to collect every piece of data possible and try to analyze it all for every variable drive themselves crazy, and essentially always end up with a poorly-focused work that's largely meaningless.

Good research is done by knowing what you expect to find, defining the scope of investigation and limiting data collection to the relevant portion. If you're interested in fair use, collecting and analyzing data on blatant infringement is a waste of your time and disk space, and simply adds noise with no signal. Keeping an open mind and being a good researcher means being willing to let the data prove you wrong.

Not that this has to be a strict research project, but even if it were, there's nothing wrong with going in under the assumption that fair use is likely being trampled, so long as you cop to the error if it's not.
posted by Dr.Enormous at 7:14 AM on May 21, 2008 [1 favorite]

"It's about fair use, and remixes or performances forced off YouTube."

Yeah, uhm, casual browsing showed more than one video that clearly was not a remix (as I have seen the videos), just a simple repost of copyrighted content (like the anime film example) in full. If the copyright holders wish to not have their stuff randomly posted on youtube, shouldn't they be able to have it removed?
posted by dabitch at 7:33 AM on May 21, 2008

we'll look at the data and see what we find.

Obviously YouTomb needs to get all the video data (both removed and unremoved) from every site on the Internet, past and future. Only THEN can they determine whether fairly used videos from YouTube were removed. Actually, to be on the safe and rigorous side, they better get audio and text data too.
posted by DU at 8:45 AM on May 21, 2008

Well, duh. That is not what this is about. This is about fair use.
posted by languagehat at 8:45 AM on May 21, 2008

(That was to dabitch.)
posted by languagehat at 8:46 AM on May 21, 2008

Yeah I can see everyone saying that, but posting full clips obviously isn't fair use. Unless I'm missing something.
posted by dabitch at 11:15 AM on May 21, 2008

Yeah I can see everyone saying that, but posting full clips obviously isn't fair use. Unless I'm missing something.

Nobody is equating full content upload with fair use.
posted by inigo2 at 1:53 PM on May 21, 2008

Sorry, I'm probably being unusually dense today, since Dean Jansen said "it's about the remixes or performances forced off YouTube", I thought for some reason that YouTomb would be showing the screenshots/info on such cases only, but it doesn't obviously. It shows all files deleted off youtube, period. I think I got confused further up in this thread, nevermindme.
posted by dabitch at 2:46 PM on May 21, 2008