Openness without Persecution: Preventing Attacks on Archivists
—Nikita Mazurov, data liberation enthusiast, post-doctoral researcher in the Living Archives project, Malmö University, Sweden
(Anticopyright/anticopyleft 2015. No license.)
The Shadow Archive
Traditional archives today have been eclipsed by shadow archives—free-form digital collections of cultural products which exist in-between any given legalities accompanying or hindering data exchange and characterised by an openness borne of peer, or guerrilla, curation. Within the shadow archive, everything from the self-selection of content to its actual storage and distribution may be orchestrated by open, voluntary collaboration, with guerrilla curators consisting of individual, self-designated contributors, as opposed to institutionally created and sanctioned job forms. To give but one example of the sheer magnitude of the shadow archive, the British Film Institute’s (BFI) archive—described by the Guardian as “reputedly the largest in the world, including 230,000 films and 675,000 TV programmes”—pales in comparison to LimeTorrents, a directory which indexes more than 1.2 million movie torrent files and more than one million TV show torrents.
So-called torrent directories such as LimeTorrents neither store the actual content of the ensuing archive—the film, audio, text, and other digital media; nor do they store the file pointers to this content, the digital .torrent files which contain metadata—information about the files and helpful pointers for specialised BitTorrent software clients to be able to download their particular contents. Instead, torrent indexers such as LimeTorrents provide links to these torrent files on other websites, which can then be downloaded and loaded into the torrent software, and then the particular file is finally retrieved by downloading it from other users who already have the file. In other words, LimeTorrents constitutes a collection of links to torrent files, which are themselves a sort of digital pointer to the actual file that one wants to share by way of loading the torrent file into the aforementioned torrent software.
The decision of choosing which files to index on LimeTorrents thus falls not on a predefined team of digital curators, but on anyone wanting to add content to the directory. Similarly, individuals have the option to share or not share particular files, as the content itself is also not stored on any centralised archive project server, but rather, is located on individuals’ computers.
The shadow archive thus contests the very notion of a centralised archive space, functioning instead as a dispersed, distributed phenomenon.
The openness of the shadow archive is, above all, predicated on voluntary user participation and cooperation. This openness is precisely what makes these archives a powerful force for the distribution and preservation of cultural production. However, this same openness is also precisely what makes shadow archives extraordinarily vulnerable to attacks by adversaries who want to shut down the distribution of the world’s cultural heritage that is preserved and disseminated outside of official state or corporate channels. This article will look at a particular manifestation of citizen curation—bootleg film recording—and will examine the risks posed to guerrilla archivists and curators engaged in such practices, in addition to putting forth various practical counter-measures that may be taken to help ensure their safety from persecution.
Cams
In the vocabulary of the shadow archive, a cam is a digital recording of, most commonly, a motion picture (although it can also be a recording of, for instance, a play or an art installation) made with a camcorder. The practice of engaging in producing cams is hence the act of camming, and those who engage in this practice are known as cammers. The usual technique for producing a cam—as graciously explained by the Motion Picture Association of America—is taking a modified camcorder and tripod into a movie theatre, recording the film as it plays on the screen, transferring the resulting video file from the camcorder, and then distributing it via physical media as well as online. A potential, typical workflow of a shadow archive guerrilla curator may thus proceed as follows: cam a copy of a film not yet available on the archive in the theatre, transfer the resulting digital copy of the film from the camera to the computer, make a torrent file using BitTorrent software and start sharing the file, upload the torrent file to a torrent repository website, and finally, add a link to the torrent file on any number of selected shadow archives.
The decision of which film to add to the archive, as well as the allocation of resources across the entire workflow itself, is made entirely on the curator level. This is in contrast to the formal, top-down control over curation found in traditional archives, where it would indeed be highly unlikely that anyone could come in off the street, add an item to the archive, and instantly have the item be accepted and available to the public at large—imagine the (im)possibility of such a workflow at the aforementioned BFI!
Persecution
So far so good, but what if one day upon returning to the same theatre where one had already made numerous cams of films, the curator suddenly finds oneself in handcuffs? The crucial point here is that the darker flipside to the openness of contribution is the openness to attack. It is precisely because these cultural products can be shared in an open archive, accessible to all, that those with adversarial interests may gain access and seek to harm the archive by neutralising its contributors. Here, an adversary is any agent (individual or collective; state, corporate or independent) who seeks to inhibit the operations of a shadow archive and any of its curators.
To give a recent example of precisely such an occurrence amidst an unfortunate and lamentably all-too-sizable pool to choose from—additional examples including a person being arrested on suspicion of recording films in Nottinghamshire in September, 2015, and another person being caged for 33 months due, in part, to recording a film in a theatre in Wolverhampton—we turn to the case of Sujay and Challa Pandi Ravi. The two were alleged cammers who had allegedly recorded and distributed some 300 films, both digitally and physically, over the last two years before making one last trip to the Q Cinema in Bangalore on July 24, 2015, for the ten o’clock matinee screening of a James Bond movie. Their arrest means that whatever particular archives they may have contributed to, whether directly or else indirectly by a copy of a cam they had made ending up on LimeTorrents, for instance, would now (or at least for the time being) be deprived of future contributions from them. As a result, all those who may have benefited from Sujay and Ravi’s curatorial additions—for example, those unable to go to the cinema, whether due to geographical distance or physical disability, or perhaps due to simply not being able to find a babysitter or any other salient reasons—would, for the time being, be deprived of content.
As Sujay and Ravi would languish in the confines of a prison cell for engaging in archival practices deemed impermissible by a repressive econo-legal structure predicated on violently enforcing the myth of intellectual property at the expense of cultural livelihood, so too would the films languish in the confines of the theatre as products of a global culture and be further denied to those not in the vicinity of the theatre with money to spare. The archive would likewise be all the poorer for their absence. The apprehension of Sujay and Ravi can thus be seen not as mere prosecution, but indeed as an act of persecution, or ill-treatment precisely due to the underlying political beliefs—regarding unsanctioned sharing of cultural products in this case—of the curators.
Apprehension
But let’s back up—how on earth did Sujay and Ravi end up arrested in the first place? How were their curatorial practices abridged by the aforementioned adversaries? Reports of archival repression and curator persecution are notoriously glib on the details of the circumstances that lead to arrests. From official reports, we only know that the two were apprehended, but not how. There are many overt mechanisms for apprehension—stemming from a theatre employee noticing the use of a camcorder and calling the authorities, to perhaps one of the cammers themselves inadvertently letting slip that they were going to be recording a particular film, but one particular vector of target neutralisation perhaps necessitates more time and explanation due to its covert operation—latent cinematic watermarks.
When discussing technological impediments to open guerrilla archival practices, we can group them into three broad categories: Digital Rights Management (DRM), overt watermarking, and finally, covert watermarking. DRM is a manifestation of access control: managing the ability of someone to use, for instance, a piece of software or play a movie. If you purchased a digital download of a movie from Amazon, for instance, you will be unable to play back the file on more than two devices; if you try to play the file on a third device, your video player will display an error saying that it cannot decode the file. This is one evident example of DRM-based access control: access to a given file may be overtly denied or granted by the seller.
On the other hand, let’s say that you download a copy of a journal article from an academic database. You may be able to open the article on as many devices as you want, but you may notice a message which appears in the margin of the article saying that it was downloaded at such-and-such time from such-and-such place. You may thus not be inclined to upload the article to an open shadow archive, as perhaps the publisher will find it and then know it was you who shared it and instigate legal proceedings against you. This is then an example of overt watermarking: the given cultural product is explicitly or visibly marked to identity the source (the purchaser), leading to potential apprehension of the source should the source decide to engage in such archival practices as that which are unauthorised by the seller of the content.
The flipside for publishers is then, of course, that if others can clearly see the overt watermark, they may also be able to remove it. This is why some content controllers have switched to using latent or covert watermarks which are not immediately apparent to the unknowing viewer. This may be as simple as having the same message as in the previous example of overt marking, but now the message is in the same color as the rest of the page—invisible unless one knows to look for it. Of course, content controllers have much more elaborate ways of potentially hiding identifying marks in the given cultural product as well. Given that, unlike with either DRM or overt watermarking, we are unable to see right away whether a particular item has been marked, covert marks are far more dangerous to the open sharing of cultural production.
Discovery
Let us then return to the case of Sujay and Ravi and speculate on how the two guerrilla curators may have been apprehended. To make things challenging, let’s assume that what led to their apprehension was perhaps the most nefarious form of technological impediment to open archiving—the aforementioned covert watermark. Let’s assume that an adversary, in this instance, an outfit known as Anti Piracy Solutions (MAPS)—who have previously collaborated with the police to neutralise many guerrilla archive curators—found out that Sujay and Ravi were recording films from a particular theatre at particular times of day and then simply waited for them to show up the next time. But how could MAPS have come to possess knowledge of Sujay and Ravi’s curatorial habits?
A modern film essentially has two key components, two streams which join together to produce a motion picture: video and audio. Both are potential venues for covert watermarking. Turning to the video first, a feature-length film is composed of thousands of still images known as frames displayed one after the other at playback speeds such as 24 frames per second. Given that a 90-minute film may thus have roughly 129,000 individual frames, and that the human eye cannot usually pick out the differences in every single frame as they are being displayed, what if tiny markers or dots could be imbedded in select frames? If frames numbered 37; 75,006; 101,183; and 122,894 all had unique formations of dots in them, chances are that most viewers of the film would not notice one bit.
Now let us imagine that the frame, a still image composing the film, can be parceled out into sections as big or small as necessary, with each section in turn receiving a numerical value. In other words, we could say that if a dot appears in the top left corner of the frame, it receives the value ‘01’, and if it appears somewhere towards the bottom right of the frame it receives the value ‘32’. If all the values of all the dot positions respective to their particular frame were combined to form a string or line of numbers, we could get a serial number such as 01322012. It then stands to reason that an adversary such as a film studio could then assign a unique serial number to every individual print it sends out to different theatres.
If an adversary were to then analyse a cammed copy of a film procured from an open archive, locate the marked frames, and calculate the positions of the dots, they would thus extract the latent serial number covertly implanted in that particular copy of the film. If MAPS did this and then found out that many copies of films all had serial numbers that led to the same theatre, they could deduce that this particular theatre was the one to concentrate surveillance efforts on for any future camming activity.
Likewise, the same thinking can be applied to the audio stream of a film. If unique sounds are implanted at unique intervals in the audio track, for example, if there’s a barely-perceptible ‘bleep bloop’ at one hour, two minutes, and three seconds and a ‘bloop beep’ at one hour, thirteen minutes, and one second, and so on as necessary to reconstitute a unique serial number, then the audio track could also readily betray the location where the cam (or at least the recording of the audio stream) was made. Knowing the location, adversaries could once again concentrate extended surveillance efforts around that particular theatre, leading to a greater likelihood of apprehension and, in turn, the prosecution of the cammers.
Prevention
Given these briefly outlined potential attacks on open archivists, what then are some potential preventive counter-measures? With regard to the visual watermarks imbedded in the frames of the film, each frame can be scrutinised separately for any potentially identifying markers, and then these frames could be replaced with a ‘dummy frame’ identical to the one immediately preceding or following it (alternatively, the frame could also be simply deleted with no replacement) to effectively excise the troublesome watermark. Similarly, the audio track may be closely scrutinised for any faintly perceptible irregularities which can then be removed using basic audio editing software.
In other words, the guerrilla archivist must consistently and effectively engage in counter-forensic practices to ensure both their own safety and the longevity of the shadow archive.
Before engaging in the actual sharing and distribution of procured cultural products, successful curation of the content must begin with the careful scrutiny of the given item and the removal of any potentially identifiable information. In order to protect the archive, we must protect the same guerrilla curators who make it what it is. The first step lies in awareness-raising: covert watermarking is successful precisely because it remains unseen. By highlighting its operating practices, we neutralise its key primary line of effectiveness—lying in a latent state waiting to be discovered by adversaries exploiting the public openness of the shadow archive. By making the covert overt, we pave the way for the next step: the removal of the marks to facilitate the successful and continuous supply of materials to the archive without fear of persecution for engaging in the distribution of cultural products.
The job of today’s citizen curators is thus the protection of their fellow curators and the archive in general against attacks by adversaries who would love nothing more than to see the shadow archive fall so that they can maintain their tenuous grasp over control of what are deemed to be acceptable archival and curatorial practices around cultural production.
About the author
Nikita Mazurov is interested in investigating modes of guerrilla archival practices, the threats thereto, and the protection thereof. He is further interested in exploring potential modes of ensuring archival longevity and availability of information.
