Data stream, database, timeline: the forms of social media
Lev Manovich
[Part 1]
The interiors of Buddhist temples and Medieval cathedrals, store displays, books, newspapers, magazines, films, motion graphics, manga, museums, car dashboards, wayfinding systems in airports, email, chat, blogs, social networks and every other communication media organize information in particular ways, creating distinct user experiences. These information design patterns [ http://en.wikipedia.org/wiki/Design_pattern ] specify how information is presented visually and/or spatially, how it is updated over time, and how users can interact with it. Some of these design patterns are unique to particular communication technologies; others are more general and shared by multiple technologies. We can call such shared patterns the forms of information. In this article, I will discuss three common information forms of web-based social network services: a data stream, a database, and a timeline.
Any discussion of the interfaces of the social network services has to be done carefully. First, Twitter, Facebook, Google+ and other companies periodically change the interfaces of their products (think of Facebook Timeline interface introduced at the end of 2011). Second, the companies are expected to provide access to these services on a number of platforms – full web sites, mobile web sites, the apps for iOS Android, and smart TVs. [ http://en.wikipedia.org/wiki/Mobile_website ] Because of the differences in screen size, types of files which can be accessed, and other technical details, the interface for each platform will have some differences. Third, numerous third party web sites and mobile apps designed for interacting with the social platforms may also offer alternative interfaces. For example, instead of a single timeline as offered by the native Twitter interface, TweetDeck allows the user to curate multiple columns each displaying different type of content. (In 2011, over one million applications and clients for Twitter were registered.) [ http://en.wikipedia.org/wiki/List_of_Twitter_services_and_applications ] Forth, users also use third-party applications which aggregate the data from many social networks, providing a single commmand center interface (for example, HootSuite). Because of these considerations, it is not meaningful to talk about a single “Twitter interface” or “Facebook interface” as seen by all users. So when I discuss below the details of how these services organize information, these details only refer to the services’ own web sites and mobile clients, as they are implemented currently (10/2012). However, the general information forms of a data stream, a database, and a timeline are central to both
Before digital computers, the data was typically recorded in some permanent medium. This meant that the format in which it was presented was also fixed. The invention of interactive graphical computing in the 1960s enabled displaying the same data in various ways on the computer display. The user experience of the data was no longer dependent on how it was stored (files, relational databases, object-oriented databases, etc.) (More precisely, we should say that the physical representation of data, its logical representation, and its user representation became separated.) In addition, the display could now be updated dynamically in real-time. This added more possibilities for displaying data.
When the first Macs in 1984 brought a graphical user interface to the masses, they popularized this new independence. The Finder allowed users to view the files and applications as icons or as a list of items, and sort any view in four different ways.[ http://www.guidebookgallery.org/tutorials/mac1984/whatelseisinthemenus ]
The rise of the World Wide Web in the 1990s, and social media in 2000s expanded this principle to other collections of data. As an implementation of a hypertext system, the design of the web and the graphical web browsers emphasized a particular information form - hyperlinks between separate pages (thus the logical model of the web and the interface view were closely aligned). In many popular illustrations of the web at that time, it was similarly shown as a network of single linked documents. However, the users were free to link web pages n whatever way they preferred. This led to the emergence of certain common patterns for organizing the data that were not originally planned by computer scientists. Rather than being sets of pages all linked to each other, the actual web sites created by users and companies in the 1990s often followed a different organization: a single page presenting a large collection of linked documents, i.e., a curated catalog of data objects. The examples included the list of “favorites” (other sites a user liked), a collection of personal photographs, separate radio shows archived on the site of a radio station, etc. Such digital catalogs were also very common in stand-alone digital media products such as CD-ROMs presenting artworks from museum collections.
In my 1998 article Database as a Symbolic Form I called this information form a “database” and opposed it to the historically dominant way of organization information – a narrative. [ http://manovich.net/articles.html ] I used the word “database” to describe a catalog of objects that does not have a default sort order. (Metaphorically, we can say that in a catalog the objects are organized in space rather than in time.)
As I explained in the beginning of that article, the actual computer databases are anything but simple catalogs of objects. The databases allow users to compute and retrieve many kinds of information about the stored data; generate and combine subsets of data; create different views of the data; and perform many other operations. These operations are enabled by database languages such as SQL. Similar to the separation between how data is stored and how it is represented to the user in general computer interfaces, database design also differentiate between three levels or data organization: external (specific database views designed for end users), conceptual (the global view only available to the database administrator), and internal (database implementation). [ http://en.wikipedia.org/wiki/Database#Database_architecture ] I wrote: “From the point of view of user's experience a large proportion of them [museum multimedia, personal web pages, company sites, and other types of 1990s new media] are databases in a more basic sense. They appear as collections of items on which the user can perform various operations: view, navigate, search. The user experience of such computerized collections is therefore quite distinct from reading a narrative or watching a film or navigating an architectural site.” To underlie the importance of a computerized collection, I called a database the “symbolic form” of our time.
In 2000s, the web was reshaped by new economic, social and technological forces: web commerce (e.g., Amazon, iTunes), blogs, social networks and social media (e.g., Orkut, Flickr, YouTube, Facebook, Twitter, Sina Weibo, and many other media sharing sites and social networks), and mobile computing (smart phones, tablets, ultraportables). So what happens to the database form in this decade? Is it still the key information form, co-existing with other forms (for instance, a spatial narrative in video games)?
I want to suggest that in social media, as it developed until now (2004-2012), database no longer rules. Instead, social media brings forward a new form: a data stream. Instead of browsing or searching a collection of objects, a user experiences the continuous flow of events. These events are typically presented as a single column. New events appearing on top push the earlier ones from the immediate view. The most important event is always the one that is about to appear next because it heightens the experience of the “data present.” All events below immediately become “old news” – still worth reading, but not in the same category.
The data streams of Facebook and Twitter are perfect examples of this information form. (In design patterns terminology, it has been called "activity stream" pattern). [ http://ui-patterns.com/patterns/ActivityStream ] In the center of Facebook is News Feed, featuring an updated list of user’s friends activities: conversations, status updates, profile changes, and other types of events. Even more immediate is Facebook Ticker that displays the updates instantly. [ https://blog.facebook.com/blog.php?post=10150286921207131 ]
Twitter’s design also puts forward the “social stream” experience. Depending on how many users you follow, your experience maybe ambient (infrequent updates) or very dynamic, with new tweets from different users appearing rapidly one after another. (According to the Twitter help article, currently each user is limited to 1000 tweets a day, and user is normally allowed to follow up to 2000 users). [ http://support.twitter.com/articles/15364 ]
In Facebook and Twitter interfaces, individual broadcasts from spatially distributed users are formatted into a single constantly growing montage. (However, since no single author organized this montage, the events often have no connection to each other, so “montage” maybe is not the best term. We also can’t compare this with a surrealist intentional juxtapositions of completely unrelated objects; if you have many friends with similar backgrounds and interests, at least parts of your stream are likely to refer to similar topics and experiences.)
Watching the collective data stream formatted into a single column can be fascinating and mesmerizing. There is a pleasure in being "in the stream,” in watching rapidly growing conversation or a series of comments, in expectations about what new messages will appear next. And if you are switching your attention back and forth between the data stream and other social activities such as walking, talking with a friend, or doing homework, nothing important is lost because you can always scroll down to see the recent events you missed.
Data stream can be a called a quintessential modern experience (“Make it New”), but intensified and speeded up. But comparing data streams generated by hundreds of millions of people at the same time to navigating a modern metropolis or reading a newspaper a hundred years ago is as useful as comparing 2012 feature films (shot at 4K and put through the software where you can adjust every pixel) to first flicks by Edison and Lumière brothers. What they share pales in comparison to all their differences. You also can’t call a stream user a “voyeur” either since s/he actively participates in stream construction, posting and responding to posts by others.
If you do want to evoke some modern phenomenon, it maybe more meaningful to compare the data stream experience to that flaneur, as described by Balzac, Charles Baudelaire, Walter Benjamin, and many other modern writers. The flaneur navigates through the flows of passerbys and the city streets, enjoying the density of stimuli and information provided by the modern metropolis. He can intensify his experience of “being in the flow” by choosing particular places and times of days. (My own favorite place is Garosugil area in Gangnam part of Seoul that is busy the whole day, seven days a week.) [ http://en.wikipedia.org/wiki/Garosugil ] Although Twitter or Facebook user experience of the data stream may appear passive - just watching from the outside the flow of updates on the screen – in fact she is like a flaneur because she actively seeks information density. Like flaneur, she also can control it. By subscribing and subscribing to different people, groups and lists, choosing what kinds of events will appear, and controlling her stream in other ways, she can adjust the density of the experience and how predictable or unpredictable the information will be.
At the end, all such comparisons to 19th or 20th century modern figures have limited usefulness. Because social networks are used by people for many different purposes and in different ways, with the patterns of use varying between age groups and genders, no single figure (voyeur, flaneur, etc.) can capture it all. (For example, the survey of U.S. Facebook users by Pew Research Center found that "Women average 21 updates to their Facebook status per month while men average 6." They also found that the average number of friends a user has varies dramatically between age groups. [ http://pewresearch.org/pubs/2262/facebook-ipo-friends-profile-social-networking-habits-privacy-online-behavior; http://pewinternet.org/Reports/2012/Facebook-users.aspx ]
Moving to the art, the first artistic representation of the collective web data stream was the amazing installation Listening Post by Mark Hansen and Ben Rubin (2002). [ http://earstudio.com/2010/09/29/listening-post/ ] In this installation, the bits of conversations pulled from multiple Internet chat rooms and bulletin boards forums were displayed simultaneously in six dynamic layouts across a large display wall made from 300+ small screens. Listening Post anticipated data flow interfaces of Facebook and Twitter by about five years – and today it keeps reminding us that these interfaces are not the only possible ways to format the data streams.
[End of Part 1]