The Longest Now

Wikipedia can and is selective about what subjects are ‘encyclopedic’. The weeding process is nevertheless slow in comparison to the influx of new content; in contrast to a traditional encyclopedia, where both proceed at roughly the same rate.

Every now and then there are projects to merge a certain class of stubs, giving them prominence more suitable to their importance; or projects to rename or remove other classes of articles. This was in fact part of the genesis of Wiktionary; dictinoary definitions were considered unencyclopedic and Had To Go.

Clay Shirky recently wrote about a Wikipedia article about a new Linux distribution (now in alpha) which was referenced by a Slashdot post; and the Slashdotting of the site then appended as a footnote to the article. He suggested that this was an example of someone shilling for their work; a form of astroturfing to create an “instant trend”.

However, the Wikipedia author in question didn’t know much about the OS being written about; and has been an active contributor to WP for many months. This wasn’t a reputation hack, although from glancing at blog posts about Shirky’s writing, you might guess that it was.

I wish that Shirky would assume good faith… in this case, assuming that both the /. poster and the WP editor were enthusiastic about the tool, and not out to scam up some quick advertising. The Wikipedia editor who started the article got into a flame war with Shirky on its talk page, and came out looking like an ass. But Shirky’s post, in assuming bad faith, was offensive. That’s how any number of ugly flame wars start. There’s nothing like someone baselessly accusing you of conspiracy or lying to raise your hackels.

Hints that this wasn’t a reputation hack: as Shirky noted himself, the article was edited 20 times by this person during its first three weeks, on many different days. That’s not a sign of someone trying to drop in an ad for a favored tool. Checking the article’s revision history, one can see this early edit summary: “added info box that I pasted from the Fedora Project. This article needs cleanup and expansion. I will ask for help from the Symphony OS forum.” Not the comment of a long-time project devotee.

The world of amateurized publishing, particularly on Wikikpedia, is not really all that anonymous. A glance at the forum on the Symphony OS website turns up the aforementioned request for help, and shows how [un]familiar the author was with the OS when creating the initial article.

Finally, Shirky clncludes with “the threshold for exclusion from the Wikipedia is so low… it’s hard to see how to defend against the creation of pages where so little is at stake for anyone but the advertiser” — little may be at stake for him; not so for those editors who obsess about the quality of the encyclopedia. Here’s a little statistical TMI:

For love of Shirky, I trolled thorugh the latest 50 new pages (going back 45 minutes), to see how many had been clarified or NPOV’ed since their creation. Most new pages cover minor subjects, about which a passing observer might think there was “little at stake” (e.g., the List of fish on stamps of Cote d’Ivoire). About a dozen of them had been tweaked in some way; often with a “cleanup”, “explain significance” or “deletion” notice (see for instance the first hour in the life of “The Singapore Stone“). Another five had been marked as stubs or categorized, both of which help attract the notice of subject-specific readers who would notice minor embellishments or mistakes. So within an hour, 25% of new articles had been looked at critically and been changed to conform to standards of suitability or style. Two or three looked as though noone had yet looked at them or cleaned them up.

I would guess that 95% of new articles are checked out for suitability and appropriateness of content within their first day of existence. Perhaps one in a thousand doesn’t get looked at for weeks, most likely an article on a niche subject, or a stub with few links that only got edited once. To be an effective reputation hack, you have to TELL someone else about it, feeding a negative feedback loop.

Now of course, if Clay’s blog were part of Wikipedia, those subtle POV errors in his post would eventually be corrected; wikis are good with such subtletlies. As it is, people will refer to that post as an example of astroturfing and reputation hacking for months.