Archive for the 'Scholarpedia' Category

YAWA: Yet Another Wikipedia Alternative

December 21, 2006

Tech News World, a site I’ve never really paid attention to, caught my attention with their rave review of Scholarpedia, a Wikipedia alternative that claims to be the first “free peer-reviewed encyclopedia”. By this, it means that the site uses the same open-to-mostly-anyone model as Wikipedia, but uses a very different model to produce article content.

Scholarpedia can be seen as a cross between Wikipedia and Citizendium: while the former gives equal access to everybody, and the latter uses expert “editors” to moderate and guide the contributions of the rest of us, Scholarpedia attaches one or more experts to each article. Each article is written by an author that has been elected or invited by the community, while the curator for that article is responsible for maintaining it using feedback provided through an anonymous feedback forum.

In some sense, Scholarpedia already has a type of stable versions system implemented. Edits can be made by anyone with an easily-registered account, but curators are then employed to moderate those changes and ensure that the chaff is filtered out. However, those changes are shown immediately, prompting the following message at the top of articles with unmoderated changes:

This article is undergoing 2 initial reviews; It may contain inaccuracies and unapproved changes made by anonymous reviewers.

Tech News World is enthusiastic about the chances for Scholarpedia’s success, but I don’t believe that’s really justified. They end on the funny note that:

Scholarpedia may be the second-generation wiki that makes the grade.

Frankly, that’s nonsense. As impressive as the project appears — and it really does seem to be working quite well at the moment, for such a small website — the simple fact is that Scholarpedia can not scale.

The English Wikipedia, with its 1.5 million articles, has an editor base of approximately 160,000 users — only one quarter of whom have ever made more than a handful of edits. That means that the core of Wikipedia has largely been developed by a small number of extremely dedicated users, while the rest are casual users who add changes to the site as they see fit.

Tech News World asserts that Scholarpedia could be the “second-generation wiki” we’ve been waiting for, but even for a website one-tenth the size of Wikipedia, the workload that would be placed on the site’s curators would be obscene. Even a base of just 100,000 articles would require many thousands of active, dedicated and experienced editors to maintain, and not even Wikipedia can boast such a userbase. It simply won’t happen, not as it has been implemented in Scholarpedia.

Secondly is the issue of licensing; that is, it isn’t free like Wikipedia. While the world famous site can be copied verbatim by anybody that likes (provided they follow the fairly open terms of the GNU Free Documentation License), but Scholarpedia will have none of that:

The content on Scholarpedia is protected by international copyright, patent, and trademark laws. You may display, print or download content on Scholarpedia only for academic, non-commercial use, provided that you cite Scholarpedia. You may not publish, distribute, retransmit, sell or provide access to the content of Scholarpedia, except as permitted under applicable law and as described here.

The use of the GFDL has been a crucial point in Wikipedia’s success, as sites like has latched on to its content and provided a free source of credibility and mindshare for the encyclopedia. Its free license allows anybody to use it for nearly any purpose, and has spurred an enormous loyalty from its followers, especially in the form of the donations that keep the website alive. Scholarpedia, with a proprietary license like the above, can’t hope to benefit in the same way.

However, there could be some things to learn from the site. Like Citizendium, Scholarpedia is already using a stable versions system (one that attempts to display only “approved”, “mature” or “stable” versions of an article to the user, requiring that anonymous edits be approved prior to being displayed) that Wikipedia has been waiting on for years. The long-awaited feature was announced some months ago, but has never been added to the MediaWiki software that runs Wikipedia. Even from this example, the value of stable versions for reducing the workload strain on the website’s small userbase is fairly easy to see.