Rich Morin :: tchotchkes

Mechanical augmentation of Wikipedia

I'm a big fan of Wikipedia. I use it both as a personal reference tool and as an easy way to add depth to web-based documents. However, I think that its utility might be improved by a bit more mechanical augmentation. This augmentation could take (at least :-) three forms: generated pages, automatic content, and requested content.

Generated Pages

Wikipedia (or, more generally, MediaWiki) provides a variety of mechanically-generated pages. Several dozen special pages, for example, are used for administrative content (e.g., indexes, searches, reports). There are also some "magic" pages that support administration of regular pages, documentation of uploaded images, etc.

This idea could easily be extended to cover general content. For example, astronomers might enjoy having a wiki page for each cataloged celestial object. Aside from giving basic information about the object, each page would be cross-referenced to pages that describe its discoverer(s), neighbors (both apparent and actual), relevant literature, spectral class, etc.

Although such a collection might look quite a bit like a web-based catalog of astronomical objects, the fact that it was embedded in a wiki would give it some extra features. Users could post comments and/or questions on the "discussion" page, engaging other users in dialogs about the object. Suggestions for additional cross-references (etc) could also be made in a convenient, interactive manner.

Some types of material (e.g., biased or commercial content) are not appropriate for Wikipedia. However, there is no reason that other wiki-based venues could not publish any material they wished. Some possible mechanized wikis might cover:

• politicians, impending laws, etc. (see TheyWorkForYou)

• passages of the Bible, Koran, Talmud, etc.

• a software package's application programming interface, etc.

• a class of standardized products (e.g., fittings)

• every file, interface, and other object of interest in the Unix OS.

Of course, this is not a "one size fits all" solution. Before creating a set of mechanically-generated wiki pages, one should consider whether the content:

• can be generated in an automated (or at least semi-automated) manner

• would benefit from the features (e.g., discussion pages) provided by wikis

• would benefit from the social interaction provided by wikis

Automatic Content

MediaWiki generates some page content automagically. For example, it generates pop-up messages for links, links to versions of the page in other languages, a "table of contents" for larger pages, and a variety of "toolbox" links. However, there are many other possibilities.

Existing category and link information could be used to create concept maps or other forms of diagrams that could provide context and ease navigation. Mechanized analysis (e.g., a naive Bayes classifier) can also be used to detect relationships between articles. However, the most interesting source of information may come from enhanced links.

The Semantic MediaWiki folks are working on adding Semantic Wiki extensions to MediaWiki. Once these are in place, editors will be able to add "type" information to wiki links. By harvesting and analyzing this information, the wiki software can provide context and navigation diagrams, "intelligent" search, and other useful features.

Semantic Wikis can take advantage of all of the tooling that is being developed for the Semantic Web, so it would be foolish to try to predict all of the ways in which this information could be used. Expect the unexpected...

Requested Content

The most interesting possibilities, however, may be in the area of "requested content". Using a facility such as Getlets, a wiki page can request arbitrary content from arbitrary servers. The content may be cached (e.g., static, with occasional updates) or generated and included on a dynamic basis.

Let's say that a user is editing an article on a favorite sports team. By creating a link to the appropriate server, s/he could include the current standings of the team in the middle of the current web page. Something like:

  ==Current Standings==

  {{sports_team_standings:Wombats | current}}

The resulting content might be a formatted table, a generated image, an interactive (e.g., AJAX) region, or even a whiz-bang, multimedia presentation. After all, (most) wiki software is based on the web, so any web technology can be used on a wiki!

To be sure, there are many issues to consider in allowing this sort of activity. Is the server reliable and trustworthy? Does the facility open up security holes? Will it be a resource sink or degrade the user experience? Can users be trusted to honor copyrights and other legal provisions?

However, as Wikipedia's Jimmy Wales notes, we don't put restaurant patrons in cages, for fear that they will attack each other with their steak knives. So, let's start looking into ways that this technology can serve the world!

Technorati Tags: Getlet, MediaWiki, Model-based Documentation, semantic wiki, wiki, Wikipedia

Technorati Tags: Getlet, MediaWiki, model-based documentation, semantic wiki, wiki, Wikipedia

Mechanical augmentation of Wikipedia in Computers , Technology - posted at Mon, 17 Apr, 15:13 Pacific | «e» | TrackBack