The Work of Sustaining Order in Wikipedia

From Driscollwiki

Jump to: navigation, search


Geiger, R. S. and Ribes, D. (2010) The Work of Sustaining Order in Wikipedia: The Banning of a Vandal. In Proceedings of the 2010 ACM Conference on Computer Supported Cooperative Work (CSCW), ACM, New York (2010).

Social roles of software tools in Wikipedia (1)

  • Autonomous editing programs
  • Assisted editing tools
  • "Often-unofficial" technologies
  • Role of non-human actors in enabling a decentralized activity of collective intelligence
    • Independent of specific norms currently in force
  • Previous research often focuses exclusively on human actors (2)

"Vandal fighting" (1)

  • Epistemic process of distributed cognition

Contents

Case overview

Anon vandalism

  • Caught by detection algorithms
  • "Appears" to large number of human and non-human editors
    • Some of whom monitor WP in real-time
  • Huggle enables quick assessment of scenario, easy revert

Who keeps vandalism from taking over?

  • Heterogeneous assemblage of human and non-human actors
  • Identifying and temporary blocking of malicious contributors (1)
  • Routine activity
  • 100s of temp blocks a day

Vandal fighting as "organizationally important"

  • Distributed cognition
  • "First line of defense"
  • Newcomers' first intro to WP policies, standards, procedures
  • Taken together, individual actors enforce an epistemology without knowing specifically about one another
    • Or the big picture of a given user
    • Not one editor enforcing idiosyncratic worldviews on others (2)
  • Actively reshape the way in which edotrs engage w/ WP and its content

Why are non-human actors missing from previous research?

  • Previous research focuses exclusively on social explanations (2)
    • Behavioral, informational, computational researches
    • Some attention to the platform of WP (e.g. History, Edit, Revert)
    • Very little attention to the "largely-unofficial" software ecology
  • One reason for the oversight is that the data is often from 2006
    • In that data, tools appear as "force-multipliers"
      • Enabling human actors to be more efficient
      • Not shaping the processes themselves
  • Research on tools tends to emphasize their efficacy
    • Efficiency, accuracy
  • Need for updated analysis of the "sociality" of the technologies (2-3)

Bots

  • Fully-automated software agents
  • Algorithmically-defined tasks
    • Editing, maintenance, administration
  • 16.33% of all edits are made by bots (3)
  • Written and maintained by Wikipedians

RamBot

  • First notable bot in WP
  • Imported public domain census data into articles about cities and towns

Assisted-editing programs

  • Automating common tasks
  • Richer interface to WP data
  • Real-time views, filtering
  • Over 12% of all edits are made with such tools
  • Written and maintained by Wikipedians

Method: "Trace ethnography"

  • Method for studying distributed cognition in sociotechnical networks (3)
  • Generating "rich accounts of interaction" by combining...
    • Fine-grained analysis of the various 'traces' that are automatically recorded (by Wikipedia software)
    • Ethnographically-derived understanding of the tools, techniques, practices, and procedures that generate such traces
  • Systematic "reassembly" of markers, logs, traces of various activities by human and non-human actors
  • Made possible by publicly-available data, API
    • "Edit summaries" comprise the bulk of the traces used here
    • Bots, AEPs leave behind footprints of a kind in the edit summary
      • e.g., prefix (TW) for Twinkle, (HG) for Huggle (3)
  • Ethnographic knowledge of certain tools allows researcher to reconstruct events based on these traces
    • "Software they used, evidence they were presented with, buttons they clicked"
  • Empirical work of trace ethno can provide "rich theoretical accounts of action and practice" (4)

Advantages over single or multi-site ethnography

  • Capture network-level phenomena (4)
  • Collective, distributed work rendered "directly observable" via reconstituted traces (4)

Weaknesses

  • No larger sociocultural significance (9)
  • No history (9)
  • Partially addressable via mixed method studies
    • Archival research
    • Traditional ethnography
    • Survey
    • Interview
    • Statistical

Vandal fighting as distributive cognition

Hutchins' work on a Navy ship, navigation

  • Because cognition is distributed and must be communicated, it is directly observable (4)
  • Charts enable generalizable skills to be applied to specific tasks
  • Charts may be used, extended by many different people
    • Not so with internal mental practices

Distributive cognition is not information sharing

  • Collective analysis of data
  • Powerful, functional even when members are seemingly in isolation (4)

Case study: banning a single vandal

  • Detailed description of Huggle interface, functionality, user experience (5)
  • Talk page warnings "operationalize each offending edit into the social structure through which admins come to know users as vandals" (6)
    • Work of many independent vandal fighters "collated"/"compressed" into a single number
    • Both a warning as well as a channel for communication, coordination among vandal fighters (8)
    • Re: actor-network theory: immutable mobile inscription
  • Various tools read and write canned text to Talk Page
  • Some edits reverted automatically by ClueBot
    • because of contribution history + content
  • Other edits flagged and presented to a human by Twinkle or Huggle
  • User eventually queued on a "meta page" about potential vandals
    • Meta page is a "passage point" viz actor-network theory
  • Eventually, bots, editors, and admins converge and user is banned (7)

Transformation to the moral order of Wikipedia (7)

  • Bots and AEPs change the methods by which
    • Edits are evaluated
    • Content is reverted
    • Users are banned
  • Redistribution of moral agency to automated, semi-automated tools (7)
    • Valid, invalid content
    • Legit, illegitimate contributors, contributions
    • "Inherently moral in quality" (9)
    • "Who is left out and what knowledge is erased" (9)
    • ClueBots make automated reverts
  • Tools take users through "standardized scripts"
    • It's always possible to act differently
    • But it requires "inventiveness and time" (9)

Editors enabled by AEP

  • Domain expertise is no longer the content/subject of the article
    • But the operation of AEPs and WP (8)
  • Tools "re-present" edits in a way that make them legible to anyone
    • Responses are rendered "common sensical" (8)
  • Users not familiar with subject matter are still able to maintain order
    • Difference between "interactional and contributory expertise"
    • See Collins, Evans, 2002

Delegated cognition

  • Huggle ranks edits according to pre-established set of vandalism-identification algorithms (8)

Jump offs

Wikipedia jump offs

  • Beschastnikh, I., Kriplean, T., and McDonald, D. (2008) Wikipedian self-governance in action: Motivating the policy lens. Proceedings of the Second International Conference on Weblogs and Social Media.
    • Policies
  • Butler, B., Joyce, E. and Pike, J. (2008) Don't look now, but we've created a bureaucracy: The nature and roles of policies and rules in wikipedia. Proceeding of the 2008 SIGCHI conference on human factors in computing systems, ACM, 1101-1110.
    • Bureaucracy
  • Viegas, F., Wattenberg, M., and McKeon, M. (2007) The hidden order of Wikipedia. In Online Communities and Social Computing, pp. 445-454.
    • Formal review processes
  • Forte, A. and Bruckman, A. (2008) Scaling consensus: Increasing decentralization in wikipedia governance. Proceedings of the 41st Annual Hawaii International Conference on System Sciences, IEEE, pp. 157.
  • Kittus, A., Suh, B., Pendelton, B. A., and Chi, E. H. (2007) He says, she says: Conflict and coordination in wikipedia. Proceedings of CHI 2007, ACM, pp. 453-462.

Actor-network theory

  • Callon, M. (1986) Some elements of a sociology of translation: Domestication of the scallops and the fisherman of StBrieuc Bay. In Power, Action and Belief: A New Sociology of Knowledge, pp. 196. London: Routledge & Kegan.

Method jump offs

Retrieved from "http://www.kevindriscoll.org/wiki/The_Work_of_Sustaining_Order_in_Wikipedia"
Personal tools