Readings - Week 8

Studying cooperation and conflict between authors with history flow visualizations

What are the main problems of the Wikipedia as described by the authors, and how does the community deal with them?

Wikipedia is an online encyclopedia which detours from the normal “conventional” way of online encyclopedia’s. Unlike conventional ones, where the articles are written by a single author and then submitted for review and possible editing before being posted, Wikipedia followed a different route. This route allows for anyone to create, edit and post an article on the site, making it an open community with free access. This radical new view had a lot of pessimism at first, but it soon turned out to be a huge success and set the tone for other online community sites. The authors and community also use certain “tools” to their disposal to track their or other articles in which they have interest. These tools allow them to quickly respond to new articles, edits of their articles or general discussions concerning an article or new post.

• Page history where a user can see the history of each article.
• Recent changes where the list of recently updated or added articles can be found
• Watch list which allows users to keep a watch list of a pages they wish to monitor closely.
• Talk pages where the Wikipedia community discuss topics on “real” pages.

Due to the nature and design of Wikipedia, it is vulnerable to a few snags and problems, which will be described next:

Vandalism and Repair

Wiki’s are subject to malicious edits or vandalism which can take several forms:

1. Mass deletion: deletion of all contents of a page
2. Offensive copy: insertion of vulgarities or slurs
3. Phony copy: insertion of text unrelated to the page topic
4. Phony redirection: linking a page to unrelated or offensive terms
5. Idiosyncratic copy: adding text that is related to the topic of the page but which is clearly one-sided, not of general interest, or inflammatory (biased)

Within the community, authors quickly ascertain when an act of vandalism has occurred using the tools described above. The vandalism is quickly repaired and the articles are again posted in the state it was before the act. Sometimes, authors cannot agree whether an edit constitutes an act of vandalism and therefore, a vandalism-tracking page has been set up where authors can discuss this in more depth and natural basis.

Edit Wars

A second pattern that often emerges in Wikipedia is that of “edit wars” where people or groups alternate between versions of a document. This is often found when two or more people are contrasting each other on certain information and/or content, since they both have different views. These issues are resolved in different ways. One such is where the authors leave comments on why they have edited a specific page. Often these comments are addressed to the previous editor and explains the reason for the edit directly to that author, making it a more personal form of communication. Another method is the use of “talk pages” which have been designed specifically for this situation. Here, authors and editors can discuss the article at length and come to a formal conclusion and solution, which is then incorporated into the wiki article.

Authorship

Another contentious subject is that of article authorship. Whilst many feel the need to make it visible that “they” were the ones to submit an article, the larger community feels that it violates the open policy of Wikipedia and also infringes on the “Neutral Point of View” (NPOV) which Wikipedia strives for. Therefore, whilst the authorship of a specific article is not present on the article page itself, the talk-pages, where authors leave comments or discuss the topic are often signed with the authors user name, and in the case of anonymous users, an IP address is appended. This creates a sense of personal community where people can get to know the authors of certain articles and build a personal preference towards some.

A Content-Driven reputation system for the Wikipedia

What different notions of reputation in Wikipedia do the authors discuss?

One of the most predominant reputation measuring systems used by online systems today is that of user-driven reputation. This is where user A rates user B on what they perceive to be good submitted content or information. This however is based on personal preference and often lacks the critical eye which distances itself from personal preference. Subjective, rather than objective…

In this paper, the authors propose a content-driven reputation system, which could be used by Wikipedia to rate authors based on the content where they gain reputation when the edits they perform to Wikipedia articles are preserved by subsequent authors, and they lose reputation when the edits are undone in short order. The lifespan of an edit to a Wikipedia article is inferred from an analysis of the subsequent versions of the article. We call such reputation content-driven, since it is computed on the basis of how content evolves, rather than on the basis of user comments or ratings. This form of reputation rating will basically portray three forms of reputation or value:

• Prescriptive value: reputation systems specify the way in which users can gain reputation, and thus define, and prescribe, what constitutes “good behavior” on the users’ part. Users are strongly encouraged to follow the prescribed behavior, lest their reputation —and their ability to use the system — suffer.

• Descriptive value: reputation systems can be used to classify users, and their contributions, on the basis of their reputation, making it easier to spot high-quality users or contributions, and flagging contributions or users with low reputation.

• Predictive value: an author’s reputation should be statistically related to the quality of the author’s future contributions. This value could also lead to author-trust where articles by that author can be trusted to be accurate in the future by other users.