Navigating YouTube

Constituting a Hybrid Information Management System

“We no longer watch film or TV; we watch databases.” It’s in these terms that the Dutch new media scholar Geert Lovink addresses the shift emblematised by, among others, YouTube in the introduction to the Video Vortex Reader.[1] This is the “database turn” that Geert Lovink presents as a fundamental shift in the way in which moving images are being experienced today. Talking about YouTube in terms of a database is without doubt an adequate description of the technological basis allowing users to upload, search, find, and retrieve moving image files on the YouTube website. This, however, is not the only way in which users (or scholars) conceive of YouTube, and probably not the one that intuitively comes to mind, as the digital objects that one deals with are in fact not perceived as data sets, but rather as films, video clips, TV shows etc. - in other words: as moving images. So before we continue our discussion of the functioning of this platform as a database, we will have a brief look at some other conceptualisations that try to consider YouTube in analogy to other cultural institutions that deal with the collecting and making accessible sounds and images.

Constituting a Hybrid Information Management System

In another article from the Video Vortex Reader Thomas Thiel discusses an installation by the Polish artist Wilhelm Sasnal consisting of a 16 mm loop projection showing various video clips filmed from the screen of a laptop presenting “the historic and diverse contents or the media archive YouTube”.
The authors would like to thank the members of the Utrecht Media and Performance Seminar for their valuable comments on an earlier version of this text. [2] Referring to a video sharing website as an archive highlights, on the one hand, the fact that among the millions of clips that can be found there, a non-negligible number present, and thus make accessible, historic material. Such a point of view is perfectly illustrated by a post on the McGill Tribune by someone called Bryant, who states: “Now YouTube is an archive; just the other day I watched an educational film that was made in the 1950's, that, without youtube I would've never seen.” [3] On the other hand, the archive analogy stresses the possibility for users to find material there, which they can re-use for their own purposes, as is the case with the installation by Sasnal mentioned above. Similarly, Henry Jenkins declares that “YouTube functions as a media archive where amateur curators scan the media environment, searching for meaningful content, and bringing them to a larger public (through legal and illegal means.” [4] The term ‘archive’ is used in both cases in a rather straightforward way, pointing towards a collection of audio-visual material that is stored and can be retrieved through appropriate search operations, rather than, for instance, in the more epistemological sense in which Foucault uses it. [5] The term ‘archive’ here is furthermore associated with the general possibility of storing data collections, and does not refer to the traditional understanding of archiving as an institutionalised practice. Online data collection labelled as archives could in fact, as we shall see, be better characterised as perpetual transmission than as permanent storage. Some moving-image archivists, therefore, clearly reject the analogy because of at least one fundamental difference emphasized by Leo Enticknap:

"I don't see any evidence that YouTube is attempting to undertake long-term preservation of any of the material it hosts, which is surely a core function of a archive; one which distinguishes an archive from other types of document or media collection. Indeed, it'll be interesting to see what happens to the less frequently viewed content once YouTube's server capacity is filled.  As far as I can see, YouTube is essentially an infrastructure for the distribution of video content for end user viewing." [6]

Enticknap marks here a decisive difference, indeed: questions of preservation do not play any role on such a video sharing website, neither does the question of precisely identifying the status of a document, or the issue of different versions of a film, its registration and cataloguing according to certain standards etc. This leads Rick Prelinger to conceiving of YouTube rather as a library:

"Actually, I think your description of what YouTube does shows that it's more of a library than an archive. I understand the difference to be that an archive ('archos' - 'first') is charged with the permanent preservation of original documents, whereas a library simply exists to make copies available for access." [7]

But this proposition, too, is rejected by another AMIA member, Andrea Leigh, because of yet another important aspect that lacks on YouTube, namely the rules and regulations, that is the ethics, with regard to the material aspects of a document that govern the work of archivists and librarians alike:

"Libraries are oriented around a code of ethics (see ALA Code of Ethics […])and a core set of values (see ALA Core Values of Librarianship […]) to provide communities with comprehensive access to both information and entertainment resources, not entertainment resources only that lack selection criteria, principled organizational methods based on over 100 years of practice and tradition, and a high service orientation.
So not only is YouTube not an archive, it is not a library, either." [8]

While such debates might be seen as traditional archives setting up defence lines against new practices and especially new organisational forms appearing on the internet, there also are obvious differences between both as regards goals, procedures, and ethical commitments. (And there are obvious differences between web-based projects as well, for instance between archive.org and YouTube.) But clearly, archives and libraries are institutions that function according to relatively strictly codified lines of conduct, that have to observe standards defined by professional associations, often at an international level. More importantly for our analysis here, however, what this discussion shows is that whatever analogy is drawn with existing institutions or functions (this might even be valid for the more neutral term “repository” that is also used regularly to describe YouTube), will fall short at one level or another, covering only in part the rather specific way in which such a video sharing facility functions. [9] It thus may indeed be more productive to let go of such comparisons and start from the technological foundations of the platform, that is, as suggested by Geert Lovink, its being a database.

YouTube as a Database

Lev Manovich has identified the database as a crucial aspect of digital media as such. Going beyond the computer sciences’ definition of the database as a structured collection of data, Manovich considers it as a cultural form that follows its own logic and exceeds operations such as the storing and retrieving of data. “They appear as a collection of items on which the user can perform various operations – view, navigate, search.” [10] In addition to this, YouTube, as well as other services generally referred to as Web 2.0, offers the possibility to add items to databases, improve the information management through user-generated meta-information as well as synchronise them through so-called Application Programming Interfaces (API). [11]
In this respect, YouTube as a database is in fact more accurately described as an infrastructure, as its scope goes well beyond the YouTube internet site proper. The website “The programmable web” lists more than 330 so-called mash-up sites employing video feeds and other data from YouTube. These facilities make accessible specific selections from the YouTube database that YouTube itself does not propose to its users, combining them sometimes with other web applications such as, for instance, Google Maps, Flickr, but also other video sharing sites or music distribution services such as Last FM.

Another option YouTube offers are the so-called “embedded links” that create the possibility to easily integrate YouTube videos into all types of other environments, from personal websites and amateur or professional blogs to the online services of traditional media such as newspapers, magazines or television channels. YouTube even explicitly encourages such embeddings, as is evidenced by their proposing links to several other Web 2.0 platforms. [12] The YouTube database, in other words, is accessible not only through the one interface that they manage themselves. While surfing the Internet, a user can encounter moving images branded with the company’s logo almost anywhere. When a video has been watched through an embedded link, the viewer is offered the possibility to look at so-called related material, too. The user can thus navigate the database also from an external site, albeit with reduced options.

The YouTube database, however, does not only consist of video files, but also contains titles, brief descriptions called “info”, tags, hyperlinks to the uploader’s site or to related material, as well as user comments of variable, but sometimes quite extensive, proportions. In addition, it stores data concerning the number of views, popularity ratings, flagging rates, recursive links and other kinds of statistical information. In fact, video retrieval and management depend fundamentally upon such user-generated input provided as text. Since moving image files are not machine-readable – meaning that the program cannot identify the semantic content of this kind of files –, information management relies on metadata that name, describe or categorize whatever there is to be seen. This is, in other words, an essentially hybrid constellation, since users provide semantic input, which the machine then processes algorithmically, producing different types of clustering with a corresponding organisation of video files and metadata. [13]
Ultimately, this technological infrastructure can be seen as a specific affordance enabling new forms of media practice. In a way, thus, understanding YouTube means to describe it in terms of a “hybrid interaction” where humans and machines – users and information management systems – are inextricably linked. Here one could also refer to the approach formulated by the so-called Actor-Network theory, according to which human and non-human actors have to be considered as equally important in the constitution of social interaction. [14] As the functioning of YouTube and other Web 2.0 applications such as Flickr, Facebook and others depends fundamentally on the way in which they succeed in channelling user activities into software design, one could describe what O’Reilly addressed as “architecture of participation” also in a way akin to Bruno Latour’s analyses of translations of social protocols into technological design. [15] Similarly, Web 2.0 applications thrive on stimulating user participation on various levels, which subsequently is translated into input feeding the information management system.

YouTubeing: View, Navigate & Search

Doing YouTube can mean a number of things: one can simply watch one (specific) or a whole series of clips, one can rate, flag, or comment upon videos, and one can upload, categorize, annotate and tag one’s own moving images (either self-produced or found and appropriated elsewhere). These operations imply different levels of activity from the user, but even a simple viewing (on the YouTube site as well as with regard to material embedded somewhere else) leads to an invitation, or proposition, to watch more. Right from the start, the YouTube interface offers various choices. In addition to the search facility it shows what videos are being played at that very moment, it presents a number of “promoted videos” (proposed by the YouTube company) as well as “featured videos” that are highlighted for their qualities (having been selected for this category is one of the “honours” that subsequently is flaunted under the “Statistics & Data” information for the clip).

Viewing, in other words, is but a default aspect of navigation. The act of watching YouTube is, in such a perspective, but the practice of navigating through the database's contents, exactly as claimed by Lovink. This, however, fundamentally differs from the activity of zapping from channel to channel in front of a traditional TV set, since the various television programs are not linked to each other by any semantic relations, but are simply related by the fact of their being simultaneously broadcast. The YouTube suggestions do not really compare to the program structures of early cinema, neither. Thematically organised programs would also present some kind of overarching narrative, whereas otherwise the dominant format would rather present some sort of structured variety. [16] By contrast, viewing YouTube actually consists of navigating from one video to other, semantically related ones. This practice, as we have seen, is not confined to the YouTube.com domain, but has been implemented into many other web services and websites by means of open Application Programming Interfaces that allow users to stream contents from the YouTube database into different applications, and where every viewing leads to a list of related clips.

In order to search the database for video clips, the YouTube interface, to begin with, includes a common search bar. However, the Application Programming Interface offers the possibility to also automatize search processes and retrieve videos according to a certain search string in order to implement them into a different application. [17]
As mentioned above, the YouTube information management relies on machine-readable information describing the video clips, in order to retrieve them according to search key-terms. A videoclip of Madonna, in other words, can only be recognised as such when there is an explicit textual marker. This so-called meta-information is initially generated by the users uploading the videos, and consists of the title users give to a clip, the information added to an info box in order to provide background information or a summary of the clip, and the tags, that is a number of keywords one can select freely according to what one assumes to be appropriate labels for these images. Users viewing videos also implicitly provide meta-information, since the viewing rate is a criterion for the order of videos that match a given keyword. [18] The activities of users, either those supplying information about videos they upload in the form of a title, additional information and tags, or those viewing, rating, commenting upon or flagging videos, do affect the responses of YouTube to search requests. Navigating the YouTube database is therefore also intrinsically related to the activities of the numerous other users providing the necessary meta-information for an efficient information management. Furthermore, a specific media practice emerges here, where users create meta-information in order to receive more views of the material they have uploaded, and by the same token they improve information retrieval processes within the database. Meta-information, in other words, is crucial for the information management on web platforms that host non-machine-readable content such as videos or images. In order to function YouTube, but also Flickr and other services of that type, in fact ‘crowdsource’ the labour that is necessary to supply the meta-information, benefiting from the various ways in which users willingly or unwillingly, explicitly or implicitly provide them with input.
The success of searching moving image files thus relies upon the different types of metadata provided by the person that uploads a clip as well as by other users. Search results consist of a selection of videos that match the request in a presumed order of relevance, but may in fact not include the item one has looked for. Users then can either renew the search or click on one of the suggested clips in the hope that the “related videos” listings then will get them closer to their goal. These kinds of operations would be utterly inefficient for a traditional moving image archive, where search criteria are defined as precisely as possible, where categories and keywords are fixed in thesauri, often as a result of cross-institutional or even international agreements, and where catalogues contain similar information for each item. [19] In other words, the kind of metadata produced and used by archival institutions aim for maximum clarity and efficiency, whereas the search-as-navigation procedures characteristic for YouTube and similar platforms derive from a form of media practice that follows a different rationale.

Creating YouTube: Upload, Tag, Comment and Flag

YouTubeing in many ways exceeds the activity of merely watching videos. As we have seen, the interface on the YouTube.com domain already offers many possibilities for users to partake in its functioning. The ensuing activities can be divided into operations resulting in either explicit or implicit participation. Users participate explicitly by, for instance, starting a channel and uploading videos to the database. This activity includes the creation of meta-information, adding title, tags, and other information to the uploaded video clip. Users watching those clips can either react to them by posting a video response or commenting upon them in writing. Both forms of comments will be intrinsically related to the initial video, but may also offer additional possibilities to navigating to other user sites and videos. When uploading a video, users are requested by the YouTube interface to select a category that fits the video clip. [20] Since these rather broad categories do not sufficiently describe the video clips to allow efficient search operations, users can additionally add tags to formulate a more specific categorisation of their content. [21]

The choice of tags is supposed to be related to what the images present, but, as will be shown further on, the meta-information can also be employed strategically for other means. Obviously, YouTube can be used in many different ways. As indicated above it can serve as some sort of a repository, enabling users for instance to copy and upload material from TV and add them to an already existing, rather incoherent collection of memorable TV moments, featuring incidents from TV shows such as The Jerry Springer Show, Idols, The Colbert Report, Late Show with David Letterman, etc. Here YouTube constitutes something like a dense compilation of television highlights. Tags refer then to the title of the TV program as well as to the noteworthy aspect of the fragment in question (e.g. tags as Jerry Springer, bizarre, cheater). Another function YouTube fulfils is to serve as a channel to explain the way in which one needs to proceed in order to achieve a given result. This takes the form of videos illustrating various do-it-yourself practices. Such files are then associated with the Howto category, but additionally tags refer to the practice in question and even to related issues (e.g. tags like: Little Brother book, How to make a shirt print). Yet another way of using YouTube has been labeled as self-presentation. [22] Often such videos provoke, or explicitly ask for video responses, which constitute an additional set of video-comments to the original video. This area of self-presentation could also be understood as some kind of commentary users make on popular culture, political trends, society-wide debated issues etc.

Especially for political reasons tags are often chosen with the strategic goal to lure others to view a given video files. In such cases the practice of tagging is in a way appropriated and turned into a form of deliberate misinformation. The clip entitled “XXX PORN XXX” by user AbolishTheSenateOrg, for instance, is tagged with all kind of keywords that can refer to pornography, while the video itself is a plea to abolish the US Senate. User scottstone567 tries to attract views by adding pornography-related tags, title and info to videos that display instances of rahter unskilled painting activities. [24] It appears that the use of tags, title and description is actually quite frequently appropriated in order to increase the number of viewing for such videos, which show content that does not at all correspond to what the metadata suggest. Referring to complaints about this practice, user Redsoul76 comments in the info box of his video “Paris Hilton new Sex Video”: “RoflMonkeysCopter stop complaining and searching for porn on youtube and next blame for not finding it lol....” [23].

Users take ample advantage of the option to comment upon uploaded material. In some cases, and not always in direct proportion to the number of views, clips may provoke tens of thousands of reactions. Sometimes mere statements of approval or disapproval (in many cases just a simple “lol” or “wtf”), there also are strings of discussions between users, which by the way do not always concern the content of the clip in question. In some cases, it is an aspect of a comment that leads to a reaction and leads to a discussions on an entirely unrelated issue. In practice, there is probably but relatively small number of viewers that will read through all these comments, especially if there are thousands of them. However, the commentary section is an integral part of the database and generates metadata that are relevant to the overall functioning of YouTube, even though the content of individual comments may be of interest only to the commentator herself or himself. As for the generation of metadata, the amount of comments feeds into indications of popularity and relevance. When searching for a video, the search engine seems to favour those that have obtained high numbers of views and also received many comments. [24]

YouTube also offers the possibility for a very specify type of user comment, namely the so-called flag button allowing to “report video as inappropriate”. While making available such an option to their users may originally have been inspired by the idea that this might help to efficiently and rapidly remove pornographic, racist or otherwise extremist content from the site, it can also lead to various forms of censorship, when people declare to feel “offended” by a clip, for whatever reason. As the YouTube staff has the final authority in this question without having to argue or justify such decisions, there are some concerns about hidden forms of censorship resulting from abusive flagging. [25]

All these operations that YouTube offers as a possibility for its users – or rather: which YouTube needs to be activated in order to generate the metatdata necessary for its functioning – are at first sight ancillary options and additional services. Quite on the contrary, however, they actually provide the indispensable basis for the database’s information management. In part, this requires acts of deliberate participation – uploading, tagging, commenting, flagging – where users choose to actively contribute to the YouTube website. But in addition to such actions, which one could label as “explicit participation”, any such operation also is a contribution to the database, even though this latter aspect may be for users just a side-effect of what they consider being their main purpose, namely to interact directly or indirectly with others. In fact, every single click on one of the links to a clip, however random or accidental this choice may be, does feed the database as well. Every interaction with the YouTube site leads to a trace in the system and becomes a record relevant to the statistics that can be read at the surface as an indicator for “popularity”. Such acts of “implicit participation”, of which most users are probably unaware, are actually the backbone of the entire operation. The participation, in other words, is implemented into the software design. [26]

YouTube as Resource: Mashups and Spin-off Services

Given the enormous amount of uploaded video clips as well as the specific software design, which can be described, with reference to O'Reilly, as an architecture of participation, YouTube functions in many respects as an infrastructure and cultural resource. This is the case for the artists or other users for whom, according to Jenkins or Thiel, YouTube functions as an archive, that is as a reservoir of material they can appropriate and re-use according to their own needs. But in addition, as O'Reilly points out, it is crucial for web-based services to provide synchronizable databases in order to have their service implemented in as many third party applications as possible. The activities that we have labelled “Creating YouTube” in fact also include implementing the contents of its database into other web applications. The YouTube Application Programming Interface is frequently used for building so-called mashup websites. This term designates sites that ‘mash’ various data streams from different web applications together to create a new format. [27]

Mashups such as Tagbulb use the Application Programming Interfaces of various web services such as YouTube, Yahoo Video; Flickr, Google, del.icio.us and others in order to present related content from those different sources, or in this case rather: resources, according a given key term. [28] Mashups, in other words make available combinations of material from various services organised according to specific zones of interest in ways that the original platforms do not provide. Reversing, in a certain sense, the practice of commercial enterprises opening their own channels on YouTube as an additional distribution outlet, mash-up websites comb through all sorts of collections of material and make selections targeted towards relatively well-defined groups of users.

The permanent circulation of YouTube material, together with the aforementioned fact that identification of particular clips depends on the meta-information added by uploaders and other uses, makes that the database can be described as a repository for audiovisual material that is at the same time stable and unstable. Instability results, to begin with, from the rather incoherent and unorganised way in which videos are indexed, labelled and filed. But at the same time, and in fact by the same token, stability is generated through the organisation of content in the form of a dense layer of meta-information consisting of titles, tags and descriptions. Another factor characterising YouTube as an instable repository is the uncertainty whether or not uploaded videos remain on YouTube. Any video may disappear from the domain for a variety of reasons: material can be removed by the person, who posted it originally, or it can be deleted by YouTube staff members for violating the terms of use. But popular videos that have been removed may re-appear either on YouTube, or on other video sharing platforms. Here an unorganized, rather anarchic and accidental practice of users copying and re-uploading videos results in redundant storage. [29] This practice creates at least some kind of stability countering the instable nature of digital repositories. The inherent instability, in other words, seems to stimulate practices for compensation. One example here is the MIT project YouTomb that monitors videos uploaded to YouTube and identifies those that have been removed for violating terms of use. [30]
YouTube is therefore more than a mere web platform providing possibilities to upload and view videos. It rather appears as an infrastructure and a cultural resource that can be used in numerous ways. It constitutes the raw material for a new media practice of perpetual uploading, viewing, and deleting material, as well as streaming it into a variety of other web services and sites.

Conclusion: YouTube as a New Media Practice

While Geert Lovink clearly identifies a central and crucial point of the phenomenon YouTube when he claims that “we no longer watch film or TV; we watch databases”, this statement in a way short-circuits the basic infrastructure of YouTube and the variety of ways in which it functions for its users at its surface and even beyond, as is the case with mash-up sites or the embedding of clips within other sites. As we have tried to show. YouTube constitutes an intrinsically hybrid system of information management, where users provide all sorts of input, among which the uploading of audiovisual material most certainly is the raison d’être for the site. However, making available video clips to others is not sufficient for YouTube to function. The material has to be described, indexed and categorized in various ways in order to be storable, identifiable, retrievable and thus viewable or, in a literal sense, to become visible. Once made visible in this emphatic sense, thanks to the software design a clip gets related to other videos, ranked in terms its of relative popularity according to the number of views, it maybe even become a “featured video” because of whatever qualities have been ascribed to it, others comment upon it, it may trigger approval, disapproval or debates, it may also get flagged and consequently removed for being judged inappropriate.

While description, indexing and categorization are a standard operating procedure for traditional archives as well, albeit in a different and much more systematic and normative way, the different acts of explicit and implicit participation, the generation of metadata by the various kinds of user activities, constitute a new media practice that represents a challenge to our established conceptions of media use. In order to analyse a phenomenon such as YouTube, one needs to take into account its fundamental heterogeneity and hybridity, its technological infrastructure as well as what is happening on its surface or its interface.

The multifunctional interfaces of YouTube, providing numerous possibilities of use and re-use, form a perpetual stream of data that go well beyond the YouTube web platform and appear in a multitude of other websites and services. Using YouTube is a practice of navigating through the data base of stored contents, either by direct search requests on the YouTube website, or by ‘clicking’ through provided lists of videos on YouTube or any other website that streams videos from its data base. The hybrid information management is crucial, as it determines the retrieval and the relational presentation of videos. Users apparently quickly learned how to affect search results through keywords (tags) and additional text-based information. The versatility visible in the technical design meets the miscellaneous practices to employ YouTube as a platform and a channel for ‘shameless’ self-representation, but also educational videos, political propaganda, informative documentaries, commercial programs, as well as grassroots journalism and alternative news services and political debate.
YouTube obviously provides a platform for viewing all sorts of audiovisual clips, but also a forum for various kinds of interaction between humans, and, perhaps even more importantly, an infrastructure for generating data that can be treated as metadata. To understand YouTube, one needs to go deep into YouTube.

Endnotes

The authors would like to thank the members of the Utrecht Media and Performance Seminar for their valuable comments on an earlier version of this text.

[1] Geert Lovink, “The Art of Watching Databases. Introduction to the Video Vortex Reader”. In: The Video Vortex Reader. Responses to YouTube. Ed. by Geert Lovink & Sabine Niederer. Amsterdam: Institute of Network Cultures, 2008, 9.
[2] Thomas Thiel, “Curator as Filter/User as Creator”. In: The Video Vortex Reader, op. cit., 184.
[3] Post by Bryant dated 01/07/08 on http://www.mcgilltribune.com/home/index.cfm?event=displayArticleComments&ustory_id=28bb1450-9089-4f57-97c5-811bed08aa77 [last checked February 15 2009].
[4] Henry Jenkins, Convergence Culture. Where Old and New Media Collide (Revised Edition). New York, London: New York University Press, 2008, p. 275.
[5] See Michel Foucault, L’archéologie du savoir. Paris: Gallimard, 1969, 169-173. See also Wolfgang Ernst, Das Rumoren der Archive. Berlin: Merve, 2002.
[6] Post by Leo Enticknap on the Association of Moving Image Archivsts (AMIA) discussion list, thread “The YouTube issue”. http://palimpsest.stanford.edu/byform/mailing-lists/amia-l/2006/10/msg00274.html [last checked February 15 2009].
[7] Post by Rick Prelinger, ibid.
[8] Post by Andrea Leigh, ibid. This position, however is nuanced by another librarian: “A library is simply a collection of materials made for use by a particular population - that's the dictionary definition. […] Does YouTube hold a collection of materials? Yes. Do they provide comprehensive access? Yes (to those who have computers). Is there any organization? Probably very little, but I haven't spent enough time on there to figure that out. Do they have selection criteria? No I don't think so, but in this case it makes it all the more interesting...which is the objective I believe. The point is...I think there is definitely strong argument that it is not an archive...but not a library? Hmmmm.” Post by Brena Smith, ibid.
[9] For YouTube itself, however, such analogies are rather profitable in terms of its cultural legitimation. Hence the comment of one AMIA member: “I’m sure Google would be thrilled to know that professionals are spending their days discussing whether or not YouTube is an archive, library, or neither. $1.65 billion well spent. Ha ha!” Post by Brena Smith, ibid.
[10] Lev Manovich, The Language of New Media. Cambridge, Mass., London: The MIT Press, 2001, 219.
[11] See Tim O’Reilly, “What is Web 2.0”, 2005, http://www.oreillynet.com/pub/a/oreilly/tim/news/2005/09/30/what-is-web-20.html [last checked February 15 2009].
[12] YouTube and other video websites stream their videos in the Adobe flash video format. Various additional applications, so called add-ons, for the frequently used Mozilla Firefox Browser allow to download video clips and save them as flv files. Videos downloaded from YouTube thus constitute an 'unknown' data collection that is extending the YouTube data base into the hard drives of a multitude of users, where videos might be stored even long after their removal from the YouTube data base itself.
[13] Among others, these metadata offers the possibility to place advertisement that correspond to the content of a clip and thus generate additional income for YouTube’s mother company Google.
[14] See Bruno Latour, Reassembling the Social. An Introduction to Actor-Network-Theory. Oxford: Oxford University Press, 2005.
[15] See Bruno Latour, “Technology is Society Made Durable”. In: A Sociology of Monsters. Essays on Power, Technology and Domination. Ed. by John Law. London: Routledge, 1991, 103-131.
[16] See contributions to KINtop. Jahrbuch zur Erforschung des frühen Films 11. Kinematographen-Programme. Ed. by Frank Kessler, Sabine Lenk & Martin Loiperdinger. Basel, Frankfurt am Main: Stroemfeld Verlag, 2003.
[17] This also goes for the already mentioned so-called mash-up websites ‘mixing’ content from YouTube with other databases or streaming selected videos into an interface different from the YouTube interface.
[18] The ranking of videos provided as results for any given search request are also affected by the popularity of the video, ranking more popular videos higher. As Ann-Sophie Lehman has pointed out in her review of the category of How-To videos on YouTube, the search phrase “How to iron a shirt” retrieves a list of videos showing how to iron a shirt, but also ranked within the the top five results the video “Hillary Heckled 'Iron My Shirt'” of an incident during the Hillary Clinton campagne. The video has had a high number of viewers, the title contains the keywords “iron” and “shirt”, which in combination with the high viewing rate (99,313 views as of January 2009) and a high number of comments (988 as of January 2009) probably identified it as a video relevant to the search request “How to iron a shirt”. See also Ann-Sophie Lehmann, “How to YouTube. Aneignung und Repräsentation kreativer Prozesse als Performanz von tacit knowledge.” Paper presented at Jahrestagung der Gesellschaft für Medienwissenschaft 2008, Bochum.
[19] See for instance Harriet W. Harrison (ed.), The FIAF Cataloguing Rules for Film Archives. München etc.: K. G. Saur, 1991.
[20] YouTube makes users choose between the following fifteen categories: Cars & Vehicles, Comedy, Education, Entertainment, Film & Animation, Gaming, Howto & Style, Music, News & Politics, Non-profits & Activism, People & Blogs, Pets & Animals, Science & Technology, Sport, Travel & Events. While there most certainly is some kind of logic behind this, the categories definitely read as a rather random attempt to create a classification.
[21] Frequently used tags are keywords such as video, sexy, sex, music, rock, rap, funny, news, pop, dance, film, short, TV. The majority of the uploaded videos are categorised as Music, Entertainement, People / Blogs, and Comedy. See a survey on the most recent uploaded videos of March 12th 2007, conducted by Micheal Wesch at Kansas State University, http://ksudigg.wetpaint.com/page/YouTube+Statistics. [last checked February 15 2009]
[22] See for instance Birgit Richard, “Media Masters and Grassroot Art 2.0 on YouTube”. In: The Video Vortex Reader. Responses to YouTube. Ed. by Geert Lovink & Sabine Niederer. Amsterdam: Institute of Network Cultures, 2008, 145-146.
[23] The video does not display any pornographic content but non-nudity pictures of Paris Hilton and footage of monkeys. It is tagged as “Hot, XXX, Adult, Action!, Extended, version, real, though including, paris, hilton, and, pamela, anderson, sex, tape, clips, uncensored“, YouTube eventually removed the video due to violation of terms. User Redsoul76 reacted by uploading another clip entitled “Paris Hilton New Sex Video” in which he comments on YouTube censorship http://www.youtube.com/watch?v=e1IIL2oYHTo [last checked February 15 2009].
[24] For a quantitative study on the construction of popularity on YouTube videos and the way this is carried out by the information management system see Meeyoung Cha et al., “I Tube, You Tube, Everybody Tubes: Analyzing the World’s Largest User Generated Content Video System.” In: Proceedings IMC 2007. San 
Diego: ACM Internet Measurement Conference, 2007, 1-14.
[25] With regard to censorship concerns regarding material relating to homosexuality see Minke Kampman, “Flagging or Fagging. (Self-)Censorship of Gay Content om YouTube.” In: The Video Vortex Reader. Responses to YouTube. Ed. by Geert Lovink & Sabine Niederer. Amsterdam: Institute of Network Cultures, 2008, 153-160.
[26] For a discussion of this and similar phenomena as aspects of an “extended cultural industry” see Mirko Tobias Schäfer, Bastard Culture! User participation and the extension of cultural industries. Utrecht: PhD Diss., 2008.
[27] At the time this article was written, i. e. in December 2008, the website The Programmable Web lists 334 mash-ups for the YouTube API. http://www.programmableweb.com/api/youtube [last checked February 15 2009].
[28] See Tagbulb, http://www.tagbulb.com/ [last checked February 15 2009].
[29] This, however, has side-effects with regard to the automatic ranking by the information management system. As several identical clips can be uploaded, each of which is considered separately, the popularity status of the same video may vary considerably. See Meeyung Cha et al., op. cit., 13.
[30] See YouTomb, http://youtomb.mit.edu/ [last checked February 15 2009].