[As ever, you can read this on the BBC News website]
For some time now the people behind Wikipedia, the online encyclopaedia assembled from reader contributions and edited and maintained by those who care to get involved, have been coping with the fallout from a widely-publicised failure of their quality control mechanism.
Last November US politician John Seigenthaler took Wikipedia to task in the columns of USA Today over a false and defamatory biography of him that had been posted on the site. The biography, it eventually emerged, had been written as a prank, but it had remained online for four months before it was noticed and removed.
Since its subject was neither controversial enough to merit consistent attention from others, nor interested enough in what happened online to bother to Google himself regularly, Seigenthaler’s biography simply sat there unremarked, although we have no way of knowing how many school essays mention his entirely fabricated involvement in the assassination of Robert and John Kennedy.
Those of us who had been using Wikipedia for some time were, of course, well aware that not everything on it was necessarily true or accurate, and the real surprise was more about the wider media interest in the site and its content that the Seigenthaler story triggered.
Wikipedia is, and will continue to be, a work in progress, a best effort by thousands of people to create an accurate, impartial and useful repository of human knowledge. As such it has succeeded in covering more topics, in more languages, than any other encyclopaedia.
But it necessarily contains errors, some placed there deliberately by writers with a specific agenda and others simply mistakes that have gone unnoticed. Sometimes the errors are entirely frivolous, of course, as happened earlier this month when fans of US comedian Stephen Colbert followed his joking suggestion and edited pages on elephants to say that their population had recently tripled.
The errors are not a reason to dismiss the site’s usefulness or importance. While Wikipedia should never be the last place one looks for information about a specific topic, I increasingly find that it is the best starting point for an exploration of a new subject.
However the nature of the ‘Wikipedia’ itself seems to be shifting, largely as a result of policy decisions made since the Seigenthaler case, and this may well affect its continued usefulness.
While it continues to advertise itself as ‘the free encyclopaedia that anyone can edit’, in practice there have always been limits on what some users can do, and an administrative and managerial team who have greater privileges than other users.
From relatively early in its existence it has been possible to block a page from editing except by an administrator, but recent changes make it harder for ordinary users to create and edit pages on the site. Over time this new layer of control could mean that its timeliness and breadth – which other encyclopaedia has a list of Muppet characters based on real celebrities? – suffer as those with something to share are deterred from doing so.
A big change at the end of 2005 was the introduction of ‘semi-protection’ for pages which were being vandalised. Once a page was marked in this way only registered users of at least four days standing could make changes.
Semi-protection seems like a sensible and moderate response to a major problem for the site, and it is clearly not being abused by administrators to limit debate unncessarily.
But now there are suggestions that a new architecture of control will be introduced for Wikipedia as a whole, if it proves successful when it is applied to the German-language site next month, and this could have far wider implications.
Under the new approach page edits will no longer be immediately applied to pages but will instead have to be approved by an administrator before they become visible. Vandalism or changes which are not approved will not appear.
This is a major shift, from a ‘publish and fix’ policy to one of prior restraint, where a cadre of privileged users will supervise what appears. It is still only a proposal, so it is not yet clear if the new checks would be applied to every page, but this is obviously being considered seriously by Wikipedia’s founder Jimmy Wales, and despite the democratic veneer which surrounds the site’s parent organisation, the Wikimedia Foundation, what Wales wants he seems to get.
The large number of control features that are being added to Wikipedia, raise an interesting question for all who care about the site and its content: when does the Wikipedia stop being a wiki and just become another website?
After all, if the special thing about a wiki is that pages can be edited by any user, then introducing layer upon layer of editorial control must mean that at some point Wikipedia becomes no different from any other online publication where content is approved before it is displayed.
And then the only special thing is that the editing tools allow in-page editing rather than requiring site visitors to use special software or go to an administration section of the site as most blogging sites do.
But that’s hardly the basis for a revolution in the way human knowledge is gathered and distributed, is it? It begins to look more and more like any other community website with a limited degree of user participation.
What makes Wikipedia special and encourages those of us who are registered with it to participate in the community is the sense that we can all make a contribution. Putting more and more steps between editing and publishing risks damaging that sense of engagement and, as a result, could rapidly diminish Wikipedia’s usefulness.
If Wikipedia can find a way to combine community participation with greater oversight, perhaps by encouraging every registered user to check changes and edits instead of leaving it largely to the central cabal of administrators, then they may be able to make the new approach work.
Perhaps we should all be asked to check one random page for every ten or twenty we look at, giving our time to make the site work in return for better content?