Infoboxes and cleanup tags: Artifacts of Wikipedia newsmaking

Constructing news narratives

Journalism is in the process of being re-imagined as the number and variety of journalistic sources continues to multiply. This re-imagining is being driven both by traditional news organizations that are starting to incorporate participatory journalism aspects into their work (Deuze et al., 2007) and by amateur media workers, technologists, and institutions such as governments and non-governmental organizations becoming key players in the business of producing news (Powers, 2014). In this changing environment, there is a growing blurring of boundaries between what constitutes journalists, journalism, and the media (Mitchelstein and Boczkowski, 2009).

Participatory journalism, also known as networked (Jarvis, 2006), open-source (Deuze, 2001), or citizen journalism (Goode, 2009), is defined by Deuze et al. (2007) as ‘any kind of newswork at the hands of professionals and amateurs, of journalists and citizens, and of users and producers benchmarked by what Yochai Benkler calls commons-based peer production’ (p. 323).

The promise of participatory journalism is that it will enable ‘the people, formerly known as the audience’ (Rosen, 2006) to sidestep traditional gatekeepers of the news so that they can narrate their own stories. Wikipedia is often provided as a poster child of this promise. According to Dan Gillmor (2008), Wikipedia is an example of a wave of citizen journalism projects initiated at the turn of the century in which ‘news was being produced by regular people who had something to say and show, and not solely by the “official” news organizations that had traditionally decided how the first draft of history would look’ (Gillmor, 2008: x). For Benkler (2007), a project like Wikipedia

enables many more individuals to communicate their observations and their viewpoints to many others, and to do so in a way that cannot be controlled by media owners and is not as easily corruptible by money as were the mass media. (p. 11)

Although Wikipedia opens up its representations to the authorship of many more individuals, this does not necessarily result in a less restrictive environment for the development of diverse narratives about the world. Researchers have found that articles about cities, towns, and regions still reproduce traditional asymmetries in the representation of place (Graham, 2011), that Wikipedia’s coverage of history suffers from an overreliance on foreign government sources (Luyt, 2011) and that there are significant gender-associated imbalances in its topic coverage (Lam et al., 2011). These asymmetries are often attributed to a lack of editors from certain groups (Graham, 2011) or a culture that is alienating to certain groups (Reagle, 2013), but there is less analysis of how the tools and practices that Wikipedians employ may have a significant bearing on the narratives that dominate the encyclopedia.

In this article, I start to address this gap by examining the tools and practices that Wikipedians use to construct a single narrative about a political event (what journalism scholars have traditionally called ‘news’). I look, in particular, at two types of Wikipedia tools: infoboxes and cleanup tags that were used extensively in the ‘Egyptian Revolution of 2011’ Wikipedia English article. Infoboxes are summary tables on the right-hand side of an article that enable readability and quick reference, while cleanup tags are notices at the head of an article, warning readers and editors of specific problems with articles.

By discussing the use of these tools in the context of Bowker and Star’s (2000) theories of classification, I argue that these tools are not only material but also conceptual and symbolic. They facilitate collaboration by enabling users to fill in details according to a pre-defined set of categories and by catalyzing notices that alert others to the work that they believe needs to be done on the article. Their power, however, cannot only be seen in terms of their functional value. These artifacts are deployed and removed as acts of social and strategic power play among Wikipedia editors who each want to influence the narrative about what happened and why it happened. Infoboxes and tabular elements arise as clean, simple, well-referenced numbers out of the messiness and conflict that gave rise to them. When cleanup tags are removed, the article develops an implicit authority, appearing to rise above uncertainty, power struggles, and the impermanence of the compromise that it originated from.

Wikipedia, journalism, and non-routine work

The majority of Wikipedia research by journalism scholars has focused on traditional media’s experimentation with wiki technologies and processes (see, for instance, Bradshaw, 2007 and Glaser, 2004) or traditional media’s use of Wikipedia as a source (by Lih, 2004; Messner and South, 2011, for example), but there is still little research on Wikipedia’s coverage of breaking news as a serious journalistic endeavor. A notable exception is the work of Brian Keegan (2013) on the history of Wikipedia’s coverage of breaking news topics, the rapid growth of breaking news articles as a proportion of the entire corpus (Keegan et al., 2013), and the large-scale network dynamics of breaking news articles (Keegan et al., 2013). Although Keegan’s work effectively demonstrates a big picture perspective on Wikipedia breaking newswork, we still know little about the practices that structure the narratives it produces.

One of the defining features of the work processes related to Wikipedia articles is the vast quantity of automated and semi-automated tools that are used to coordinate work on the platform. Over the 13 years of Wikipedia’s existence, volunteer editors have created templates, bots, statistics, and snippets of code – some that are automatically triggered by a particular user action, others that become part of normative practice over time. These tools are used by Wikipedians under the framework of its ‘Neutral Point of View’ (NPOV) policy to form an assemblage of components that must be examined together to understand how Wikipedians are able to rapidly collaborate to produce a single narrative in environments of high variability.

Two articles published in 2010 emphasize the role of objects – particular automated objects or ‘bots’ (short for robots) in the smooth functioning of Wikipedia. Niederer and van Dijck (2010) expose the increasingly important role of bots in the rise of Wikipedia, arguing that it is impossible to understand Wikipedia’s response to vandalism without an appreciation of the encyclopedia as a socio-technical system driven by collaboration between users and bots. Geiger and Ribes (2010) demonstrate the role of non-human actors in the process of vandal banning, arguing that the decentralized activity enabled by automated and semi-automated tools is a type of ‘distributed cognition’ that delegates important moral decisions about what should be included and excluded from the encyclopedia to non-human actors.

More recently, Geiger (2014) has examined the role of ‘bespoke code’ on Wikipedia in supporting and structuring work and by algorithmically enacting a particular vision of what encyclopedia is and ought to be. Geiger (2014) defines bespoke code as ‘a software code that runs alongside a platform or system, in contrast to code that is integrated into server-side codebases and runs on the same servers that host the platform or system’ (p. 3) and describes a range of bespoke coded tools, such as bots, semi-automated tools, and templates, urging others to continue to emphasize the ‘concrete, material, local, and specific conditions that make projects like Wikipedia operate in the manner that they do’ (p. 12). It is two of these bespoke coded objects (infoboxes and cleanup tags) that this study focuses on.

This study

In order to understand how a narrative is constructed in the context of a rapidly evolving news article on Wikipedia, I look at the development of the ‘Egyptian Revolution of 2011’ article on English Wikipedia, with a particular focus on some of the bespoke coded objects that editors used in developing the narrative as events progressed. I chose this case because its subject is situated outside the main frame of reference of the majority of English Wikipedia’s mostly Western editors, creating the potential for a wide diversity of viewpoints to be exhibited in the article’s construction.

I employed a ‘trace ethnographic’ (Geiger and Ribes, 2011) approach to studying grounded practices of the ‘2011 Egyptian Revolution’ Wikipedia article. Geiger and Ribes call this technique ‘trace ethnography’ because the researcher is able to reconstruct a social situation by observing the traces left by Wikipedians in pursuing their work and craft these traces into rich, thick descriptions (Geertz, 1994) of the process of the article’s construction. Trace ethnography combines deep ethnographic understanding of practice through participation in a particular community with the ability to reassemble social action using the traces that participants leave when navigating Wikipedia’s socio-technical spaces.

I supplemented this detailed examination of traces by interviewing four frequent editors of the 2011 Egyptian Revolution article. I selected editors who were the most frequent contributors to the article and its talk pages and interviewed those who responded positively to my email invitations and to introductions by fellow interviewees. These interviews served not only to verify my perceptions about what happened during the time of the events in Egypt, but also added information necessary for me to develop a rich description of the article’s construction process. My use of trace ethnography in this instance is informed by about 10 years of involvement in the Wikipedia community as an advisory board member of the Wikimedia Foundation, educator, activist and, only more recently, as an editor.

In this study, I chose to focus on the objects used directly in the shaping of articles relating to breaking news of political events in order to understand how the Wikipedia assemblage resulted in the particular narrative that it did in the context of rapidly evolving information. Looking at these mundane artifacts provides an opportunity to understand what work these objects are doing for editors and for Wikipedia as a whole – an understanding of how ‘individuals and communities meet infrastructure’ (Bowker and Star, 2000: 22). Such an understanding was obtained by engaging in what Bowker and Star (2000) call ‘infrastructural inversion – learning to look closely at technologies and arrangements that, by design and by habit, tend to fade into the woodwork’ (p. 34).

This ‘close looking’ involved a detailed observation of the first 24 hours of edits of the 2011 Egyptian Revolution article, reconstructing an account of what happened by analyzing ‘diffs’ (two versions of an article with their differences highlighted), and following discussions about templates and other objects that were employed during this time. I then used grounded theory (Charmaz, 2006, 2013) to analyze the 100+ pages of ‘talk’ during the next 18 days leading up to the resignation of President Hosni Mubarak on the 11 February 2011. I did this by developing an iterative coding scheme according to questions about what was happening in the data and following up on related edits and related articles about the use and development of tools used in the article.

It was important not only to analyze data specifically within the frame of the article, but to follow up on discussions relating to the deployment and removal of particular artifacts, their relation to policy, and a host of other ‘process’ documents since these conversations form such an integral part of Wikipedia’s socio-technical space. I looked at the first 2 weeks because this is when editorial elements were chosen and stabilized – acting to frame the narrative of the page, classify it into a particular category of event, and display relevant warning signs.

In this study, I focus on two objects that have particularly important implications in the construction of articles relating to breaking news: cleanup tags and infoboxes. Cleanup tags and infoboxes are two types of Wikipedia’s bespoke coded ‘templates’. These templates consist of snippets of code that are used across a variety of pages and include boilerplate messages, standard warnings or notices, cleanup tags, infoboxes, and navigational boxes.

Cleanup tags are specialized templates that are appended to the head of an article by editors who wish to alert editors and readers about particular changes that need to be made to the article. An example is the ‘copyedit’ tag that appends a notice about the state of the article and a request for assistance:

This article may require copy editing for grammar, style, cohesion, tone, or spelling. You can assist by editing it.

In practice, cleanup tags have the dual function of warning users about potential weaknesses of the content and alerting editors in particular work groups or ‘WikiProjects’ (Morgan et al., 2013) about the existence of new content or new work to be done. According to best practice, cleanup tags should be temporary, not permanent, and they should be accompanied by explanations on the article’s talk page about what the problem is and how it might be fixed. Editors are discouraged from using tags to perform ‘tag bombing’ (adding numerous tags to pages or one tag to multiple pages unjustifiably) or from employing tags to show that they disagree with the article. In practice, however, editors do sometimes use these methods for subtly voicing their disagreement. Because any editor has the ability to add (or remove) a cleanup tag to an article, this act is the equivalent to a TV viewer being able to add warning messages about the accuracy of the content before others watch it.

An infobox template, such as the infobox template for a ‘news event’, is a type of template formatted around a particular set of categories that provide summary details of articles with a common subject. According to the infobox help page (Wikipedia authors, 2013a), Wikipedia’s infoboxes grew out of so-called taxoboxes (taxonomy infoboxes) that users developed to visually express the scientific classification of organisms. Today, infoboxes are used for a wide variety of topics – from mathematics to history and events, and from arts and culture to science and nature.

These elements played an important role in the development of the article dedicated to documenting the events, reasons, and responses of the protests that occurred in Egypt in early 2011 (Figure 1).

OPEN IN VIEWER

Figure 1.

An early version of the infobox used in the 2011 Egyptian Revolution article on Wikipedia (11 February 2011). ¹

‘The 2011 Egyptian Revolution’ article

Day 1: 25 January 2011 (first day of the protests)

The_Egyptian_Liberal published the article on English Wikipedia on the afternoon of what would become a wave of protests that would lead to the unseating of President Hosni Mubarak. A template was used to insert the ‘uprising’ infobox to house summarized information about the event including fields for its ‘characteristics’, the number of injuries, and fatalities. This template was chosen from a range of other infoboxes relating to history and events on Wikipedia, but has since been deleted in favor of the more recently developed ‘civil conflict’ infobox with fields for ‘causes’, ‘methods’, and ‘results’ (Figure 2).

OPEN IN VIEWER

Figure 2.

Infobox from the first version of the 2011 Egyptian Revolution (then ‘Protests’) article. ¹

The first draft included the terms ‘demonstration’, ‘riot’, and ‘self-immolation’ in the ‘characteristics’ field and was illustrated by the Latuff cartoon of Khaled Mohamed Saeed and Hosni Mubarak, with the caption ‘Khaled Mohamed Saeed holding up a tiny, flailing, stone-faced Hosni Mubarak’. Khaled Mohamed Saeed was a young Egyptian man who was beaten to death reportedly by Egyptian security forces and the subject of the Facebook group ‘We are all Khaled Said’ moderated by Wael Ghonim that contributed to the growing discontent in the weeks leading up to 25 January 2011. The infobox would ideally have been a filled by a photograph of the protests, but the cartoon was used because the article was uploaded so soon after the first protests began. It also has significant emotive power and clearly represented the perspective of the crowd of anti-Mubarak demonstrators in the first protests.

Upon its publication, three prominent cleanup tags were automatically appended to the head of the article. These included the ‘new unreviewed article’ tag, the ‘expert in politics needed’ tag, and the ‘current event’ tag, warning readers that information on the page may change rapidly as events progress (Figure 3). These three lines of code that constituted the cleanup tags initiated a complex distribution of tasks to different groups of users located in work groups throughout the site: page patrollers, subject experts, and those interested in current events.

OPEN IN VIEWER

Figure 3.

The three cleanup tags automatically appended to the article when it was published at 13:27 UTC on 25 January 2011. ¹

The first tag was automatically appended to the article when it was created using the ‘Article Wizard’, a series of forms that takes new editors through the process of creating an article. This was the first (and perhaps the most significant) challenge faced by the article. As soon as an article is published on Wikipedia, it is added to a list of unreviewed articles watched by page patrollers who then review the article and determine whether it qualifies for speedy deletion, whether it should be nominated for deletion (involving a more lengthy, consultative process), or whether it should be allowed to continue. Administrators are able to speedily delete an article without consensus on a number of grounds, including for copyright violations or ‘no indication of importance’. Over half of deleted articles are deleted via this process, with the rest determined by the more lengthy process where regular editors are able to nominate an article for deletion and have a discussion about whether it should remain (Geiger and Ford, 2011).

Editors working on Wikipedia’s ‘new pages patrol’ are presented with a list of new articles tagged with red alerts if an article has no categories, no citations, or is not linked to any other articles in Wikipedia. This article contained one reference to a mainstream news source, Agence France-Presse (AFP) that described how Egypt was bracing for a day of nationwide anti-government protests, ‘with organisers counting on the Tunisian uprising to inspire crowds to mobilise for political and economic reforms’ (AFP, 2011). The article also linked to other Wikipedia articles and contained a list of categories including ‘Riots in Egypt’ and ‘2011 in Egypt’.

It is entirely possible that this article could have been removed for attempting to ‘forecast’ the future by including a citation that reported only what was expected to happen and not what did, in fact, happen. Wikipedia attempts to prevent writing about events too soon and before their significance is recognized. This is expressed in the policies entitled ‘Wikipedia is not a crystal ball’ and ‘Wikipedia is not a newspaper’ (Wikipedia authors, 2014).

The second cleanup tag (that help was needed from an expert in politics) alerted users in the WikiProject Politics group to the existence of the article when it appeared in a list of other articles seeking assistance from WikiProject participants. The third cleanup tag (‘current related’) alerted users to the rapidly developing nature of information on the page and added the article to the category of pages entitled ‘current events’.

Despite the existence of these warning tags, the article endured, as mainstream media started to report on what was becoming a significant protest. As the day progressed, The_Egyptian_Liberal continued to fill out the ‘Background’ section of the article, listing individuals who had set themselves on fire in protest against worsening economic conditions, and adding to the growing list of references. About 2 hours later another editor joined him, removing the ‘new article’ template and ‘political expert’ needed tag, improving the grammar, adding links to other Wikipedia articles, and listing relevant categories. The actions of these two editors at the beginning stages of its development were crucial in solidifying the long-term future of the article.

Days 5–8: 29 January–1 February 2011

As widespread protests continued in Egypt and the death toll rose to at least 100, the position of the army continued to be ambiguous. Soldiers were ordered to use live ammunition but refused and protest groups gave their support to ElBaradei to negotiate the formation of a temporary national unity government.

Wikipedia editor, Knockledgekid87, nominated the article for a ‘POV (point of view) check’ by inserting the NPOV tag into the article. He alerted editors in the talk page as to the rationale for the tag by arguing that Carlos Latuff’s political cartoons, used to illustrate the article, ‘seem(ed) to side with the protesters’ (Wikipedia authors, 2013b). Other editors proceeded to argue against the presence of the tag and then to discuss possible compromises.

Ocaasi wrote that merely adding pro-Mubarak cartoons to balance the pro-protester images could not solve the problem since the format of political cartoons lent themselves to a particularly popular perspective. ‘Political cartoons’, he wrote, ‘don’t usually side with oppressive regimes’. After conceding to the request by replacing most of the cartoons with photographs of the protests (Figure 4), Ocaasi removed the tag at 04:55 UTC on the 29 February, writing in his edit summary,

removing NPOV check tag. most issues resolved. no ongoing disputes on talk. npov will continually be checked and improved, but no obvious issues warranting the tag.

OPEN IN VIEWER

Figure 4.

The existence of the ‘POV’ tag resulted in all cartoons (except for the cartoon of Mubarak ‘pulling the plug on the Internet’) to be removed. ¹

Later, the same day, Michaelzeng7 added the ‘news release tag’ warning (Figure 5) that the article ‘reads like a news release, or is otherwise written in an overly promotional tone’ and pointed to the ‘Wikipedia is not a newspaper’ section of the ‘What Wikipedia is not’ policy (Wikipedia authors, 2013b). This policy indicates that ‘Wikipedia should not offer first-hand news reports on breaking stories’ and that, although ‘editors are encouraged to include current and up-to-date information within its coverage, and to develop standalone articles on significant current events’, not all newsworthy events ‘qualify for inclusion’ (Wikipedia authors, 2014).

OPEN IN VIEWER

Figure 5.

News release tag warning. ¹

Ocaasi responded on the talk page that editors were already dealing with the problems indicated by the tag and that it was therefore unnecessary. He wrote that editors were ‘trying to integrate a ton of new information in an encyclopedic way already’ and that he would ‘vote to ignore all rules while they worked it out’. ‘Ignore all rules’ (IAR) is one of three Wikipedia ‘principles’ that states, ‘If a rule prevents you from improving or maintaining Wikipedia, ignore it’ (Wikipedia authors, 2013c). IAR has a controversial history vis-a-vis the encyclopedia because it has been used by some editors to disrupt Wikipedia in the past (Ford, 2013). Here, it was being used to try and counter what was seen as a rebuke to the work of editors.

Day 9: 2 February 2011

In central Cairo, Mubarak supporters on horses and camels armed with swords, clubs, stones attacked anti-government protesters in what was later known as the ‘Battle of the Camel’.

Keeping track of death counts in the article had been an issue of contention before this day but it came to a head when anonymous user, 94.246.150.168 (Internet Protocol (IP) address resolves to Poland), complained that numbers of dead in the article contradicted numbers in the table (Figure 6) and that the table should be removed. The editors argued about whether the table should be moved or kept in the main article, with some complaining that editors might be doing ‘original reporting’ by synthesizing data from a number of sources. ‘Original research’ is forbidden by Wikipedia policy and defined as ‘material – such as facts, allegations, and ideas – for which no reliable, published sources exist’ (Wikipedia authors, 2013d).

OPEN IN VIEWER

Figure 6.

A death count table from the 13:11 UTC, 31 January 2011 version of the ‘2011 Egyptian Revolution’ article. ¹

The Poland-based user 94.246.150.168 continued to complain about ‘outdated figures’ in the tables, the fact the numbers did not add up and that they were different from what reliable sources were reporting. The argument about the validity of the table reached breaking point when The_Egyptian_Liberal wrote that these death counts were more than just numbers in a table:

Listen dude, I honestly cant be bothered fighting with you. I have family in Egypt that I am worried sick about. The last time I checked the numbers it said the total was 149 in the table. We can make sure it stat(e)s the right numbers. A lot of people come to Wiki to know how many people have died where their family lives (due) to the lack of communication in Egypt. so keep that in mind.

The ‘fatalities’ category was a constant issue of contention since there was no one source that had a complete list of casualties and editors had to stitch together a number of sources as the numbers rapidly increased. Poland-based editor 94.246.150.168 argued against using the wording ‘confirmed death toll’ when sources were reporting ‘unofficial estimates’ due to ‘confusion on the streets’. S/he also noted the difficulty of deciding which deaths should be directly attributed to the protests since there were also deaths relating to looting and jailbreaks that occurred at the same time. Lihaas countered,

Absolutely it should be included, prison deaths too, because they are a result of the protests and what goes on now. Sure someone who gets a heart attack fishing on the nile wont be, but then again that would never be in any source so it wont be included.

As with the majority of disagreements that occurred in the writing of the article, these interactions centered on disagreements on how to classify: whether deaths attributed to looting, for example, should be included in the death toll for the protests as a whole. Disagreement was tense and extremely personal when the significance of these numbers surfaced; these were not only numbers on a page but deeply affected the lives of those who were personally affected by the story.

Discussion

Looking at the diffs in the first day of the article’s growth, it becomes clear that the article is by no means a ‘blank slate’ that editors fill progressively with prose. Much of the activity in the first stage of the article’s development consisted of editors inserting markers or frames in the article that acted to prioritize and distribute work. Cleanup tags alerted others about what they believed to be priorities (to improve weak sections or provide political expertise, for example), while infoboxes and tables provided frames for editors to fill in details iteratively as new information became available.

The act of inserting an infobox with categories that require updating as the event progresses, or cleanup tags alerting editors to the weaknesses of the article sets off a number of signals to editors about what work needs to be done on an article. The choice of the ‘uprising’ infobox, for example, provided the frame for other editors to understand the context of the event so that they could use their knowledge of this particular type of event to fill out the details of the article.

The modularity of infoboxes in particular enables faster distributed collaboration among editors with varying levels of subject expertise. Editors can look for specific missing information such as casualty numbers without knowing too much about the current affairs of Egypt, for example. In addition, infoboxes’ modularity has proven particularly useful in helping to coordinate translation into other language Wikipedias. These simply summarized tables have meant that ‘repeated items can be more easily translated, rather than having to interpret long sentences with complex verb and clause structures’ (Wikipedia authors, 2012). Translating an infobox that starts with the category ‘population’ requires a single noun translation – something that can be achieved much more easily than a complex sentence about the population of a particular country. This modularity has driven the recent Wikimedia Foundation initiative, WikiData, which aims to develop a semantic approach to Wikipedia infoboxes by being able to populate, alter, and translate data from a central location.

Infoboxes are examples of Wikipedia editing tools that enable the kind of rapid collaborative editing seen in articles relating to breaking news events. These tools are, however, not only material but also conceptual and symbolic (Bowker and Star, 2000). Their symbolic power is exposed when information resists categorization such as when editors were trying to combat the fact that composite figures were not yet available in the form the infobox or casualty tables required. Taken far from its origins in taxoboxes, the inadequacies of the infobox were exposed when The_Egyptian_Liberal revealed in a moment of frustration what this was really about: not numbers, but people – people who were dying and families who were worrying about those ‘back home in Egypt’.

Classification work, as Bowker and Star argue, is intensely political. And much of Wikipedia work, especially in the beginning stages, is classificatory. A key principle of Wikipedia editing is that each article be linked to other Wikipedia articles and to the Web. According to Bowker and Star, ‘every link in hypertext creates a category. That is, it reflects some judgment about two or more objects: they are the same, or alike, or functionally linked, or linked as part of an unfolding series’ (Bowker and Star, 2000: 7).

The editing of Wikipedia articles involves continuous linking and classifying. Alternative article titles are linked together by virtue of being automatically redirected to the primary title (‘2011 Egyptian Protests’, the earlier title, now redirects to the new title ‘Egyptian Revolution of 2011’, for example). The choice of words and phrases classifies the subject of the article according to particular classes of event (such as ‘protest’, ‘uprising’, ‘revolution’, or ‘military intervention’) and editors list other articles they believe are related in the ‘See also’ section (linking the article to the Algerian and Tunisian protests in the first version). They also link to categories of articles in the list of categories at the foot of the page (including ‘2011 riots’ and ‘riots in Egypt’, for example).

In the same way that classification work is embedded into the fabric of everyday life, classification work on Wikipedia is similarly embedded and thus ordinarily invisible. As Bowker and Star argue, these standards and classifications only usually become visible when they break down or become objects of contention. The title of the Egyptian Revolution article, for example, endured much debate, with editors arguing whether to call the events a ‘protest’, ‘uprising’, or ‘revolution’, among others. Editors also complained about the way in which editors were ‘forecasting the future’ with the use of the future tense in the article. One editor complained early on about the opening phrase of the article that stated protests took place ‘from January onwards’ when it was still January. S/he complained that the article was written as though it was predicting the future with phrases weighing up relative scale of events when the significance could not have been foreseen.

Transparency is an important axis of control in Wikipedia articles – especially in the first phase of their development. It is at this stage that editors leading the direction of the page were engaged in a contest with those who were questioning the quality of the article using cleanup tags. Editors continuously added cleanup tags relating to the instability of the article due to its current event status and what they believed to be attacks against NPOV at a time when the article was highly unstable. Cleanup tags were consistently removed because editors argued that they ‘were already working on the issues’ noted in the tags. Yet the removal of such tags to the casual reader may give the illusion of a return to stability before such stability had been reached. In the context of this article, battle over the visibility of cleanup tags could be seen as reflective of personal battles between editors, and less about warning readers of possible problems.

This illusion grows steadily with time as infoboxes and tabular elements with their clean, well-referenced numbers, and short, simple phrases develop an implicit authority – erasing the earlier resistance to the tabular form. Like Bowker and Star’s analysis of the International Classification of Diseases, these translated tables and infoboxes can be read as ‘a kind of treaty, a bloodless set of numbers obscuring the behind-the-scenes battles informing its creation’ and that ‘this dryness itself contains an implicit authority, appearing to rise above uncertainty, power struggles, and the impermanence of the compromises’ (Bowker and Star, 2000: 66).

When infoboxes and tables are translated, they lose their link to the messiness that gave rise to their initial development. This messiness is parallel to the struggle to turn competing voices (sources) into a single, clean narrative on Wikipedia. Such messiness resulted in some battles being waged and won – not necessarily because of any particular merit on grounds of truth or even accuracy, but because it made more sense for the rapid development of the article. In many ways, the opportunity for such battles is lost as time goes on since it is much more difficult to shift information when it becomes lodged in the tight tabular form.

The appearance and disappearance of cleanup tags and the development of infoboxes facilitate the kind of rapid collaboration Wikipedia is becoming known for but they have two important political effects. The first is the potential of obscuring the battles fought behind the scenes of an article when cleanup tags are removed and infoboxes translated. The second is that these objects also contain an implicit authority – appearing to rise above uncertainty, power struggles, and the impermanence of the compromise that originated the information contained by them.

Conclusion

Wikipedians use a number of technological tools to coordinate work on the encyclopedia. Cleanup tags set in motion a number of specialized work routines and warn users about the instability of the article, while infoboxes provide spaces for iterative work by editors who may not possess in-depth subject knowledge but who are able to more easily fill in the spaces provided by editors leading the overall direction of the article.

In addition to enabling rapid collaboration, however, these artifacts also have conceptual and symbolic effects. Cleanup tags, when removed, give an illusion of stability before such stability may have been achieved. Infoboxes lend information a degree of authority by their material form, legitimating certain knowledges at the expense of others. This has the effect of erasing information’s messiness as the ‘spill over’ of excluded knowledges loses their connection to the process that enabled their classification – especially in the case of translations.

When this happens in the realm of a rapidly developing article about a political event, it has a significant effect on framing the narrative of what happened and why. When the event is still ongoing and the stakes are high, these tools can have enormous political effects that remain invisible to both the average user who is unaware of the reasons that editorial elements are removed or added, as well as to the majority of editors who are looking for whether individual facts are accurate rather than whether they preclude or frame the narrative as a whole.

While we may look to bots and other automated forms to observe the delegation of power to technology, these more mundane classification technologies may be as powerful and less visible. Because they are deeply embedded in daily practice, they require a close following of human editors in order to understand how editors achieve both material and symbolic effects by their choice of classification technologies and systems. As Bowker and Star note, classification works beyond the individual level of the individual ‘mind’. ‘Classifications as technologies are powerful artifacts that may link thousands of communities and span highly complex boundaries’ (Bowker and Star, 2000: 287). Rediscovering the messiness of their origin and exposing the communal nature of knowledge building highlights the ways in which what looks to be merely technical decisions almost always have social and political effects.