And yet, still so far to go

February 3, 2007

I’ve been quite the Citizendium fanboy of late, so now allow me to vent my anger at some of the site’s failures. From the ever-interesting Citizendium Blog:

After considerable deliberation, indicating broad support, we have decided to delete all inactive Wikipedia articles from the Citizendium pilot project wiki. This will leave us with only those articles that we’ve been working on. The deletion will take place on Saturday at noon, Eastern time.

This is an experiment. In other words, we’re quite seriously thinking of not forking Wikipedia after all. If we see more activity on the wiki, which is what I expect, then the Wikipedia articles will stay deleted.

I’m glad to see that Larry Sanger has finally decided to do something about one of my earliest criticisms of the his project, that it would be better to create original content than to get bogged down in simply differentiating itself from Wikipedia using the same source content. As of now, there are just a few hundred articles remaining in existence at the website — the vast majority having been baleeted with a certain amount of prejudice — but the problems are still legion.


Approval

The first and most egregious mistake that has been made with Citiendium is in its Approval process, which is flawed beyond belief. Article Approval (otherwise known as stable versioning) would allow the public versions of articles to remain stable while work is performed on a draft, preventing vandalism from affecting the most viewed pages on a wiki. Approval has been in the works at Wikipedia and other projects for a long time, but Citizendium already has a half-baked implementation.

The process itself currently goes something like this:

  1. An editor decides that an article at Citizendium is a fine piece of work. They then approve it or, if they’ve edited it themselves, ask another editor to approve it for them. If they are not the only one to have done significant work on the article, a group of three or more editors can agree to approve it.
  2. A tag is placed at the top of the page, alerting all readers that the article is under consideration for approval. If no other editor has contested the proposal after a set time period — “usually several days to a week”, according to the explanation at Citizendium — it is automatically approved.
  3. The article is then approved by protecting the existing version, and creating a draft version to which edits can be made as usual.

To begin with, the Approval process throws away any hint of peer review, and replaces it with its own mangled alternative. At the present time, the vast majority of contributions at Citizendium are made by editors, not ordinary authors. The conflict of interest should be obvious: articles are written and Approved by editors, not authors.
This state of affairs was demonstrated nicely in the case of Citizendium’s very first Approved article, Biology, which was collaboratively written by a tiny group of users. The breakdown of the last 500 edits to the article looks like this:

  1. Nancy Sculerati (301 edits, 60% of total)
  2. Gareth Leng (80 edits)
  3. Chris Day (38 edits)
  4. David Tribe (38 edits)
  5. D. Matt Innis (10 edits)

These five users accounted for 467 out of the last 500 edits — 93% of them!

How many participated in the Approval process itself? Four did: Nancy, David and Chris, later joined by Larry Sanger once he’d been emailed. Three of the primary editors behind the article — who together made roughly three quarters of the contributions to the article — plus the project’s founder. In total, the article was up for review for just one day before it was approved, and no outside users were informed beforehand. Approval at Citizendium got off to a great start. Or, to put it more bluntly:

There is no separation of power at Citizendium, and it hints at a future circle-jerk of colossal proportions.

The joke of a review that Biology went through brings me to the next major problem with Approval; namely, that there is no attempt to invite outside readers to moderate the process. As an article enters the second step of the Approval process — and assuming that all goes well, and the nomination was legitimate — it is tagged with the “ToApprove” template, placing it in the “Articles to Approve” category. Right now, there are three pages there: the current draft of “Biology” (heh), the Citizendium page on the Approval process itself, and the article “Vertebral subluxation”.

Having a category that allows users to review current nominations is a great idea, but Citizendium even managed to implemented this simple feature badly. The real problem here is that not one single page at Citizendium links to the category. It is impossible to get there without either clicking the category link at the bottom of a currently-nominated page, or by checking out the source code of the “ToApprove” template.

Take a look for yourself:

Picture 5

Update: Mike Johnson added links to the category from two pages on the day this blog post was written.

This leaves us with the following situation:

  1. An article is tagged as awaiting Approval by an editor, who was very probably a major contributor to that very article.
  2. Nobody is told about it, unless they know exactly where to look.
  3. The article is then Approved automatically after an uncertain period of time.


Frankly, I cannot believe that such a situation came about. Didn’t anybody ever think to look at how people did it at another open source encyclopedia like, say, Wikipedia? For all the media’s concerns about that site’s accuracy, it has a damn good process for promoting articles already in place.


Appropriation

Let’s move to another obvious one, described nicely in Sanger’s words:

Yep, some unsigned-in person over on Wikipedia went to Wikipedia’s “Biology” article and basically replaced it with the first paragraph of the Citizendium’s “Biology” article (our first approved article), posted here (PDF, 735K). And they didn’t give us credit, as they are required to do by the GFDL!

Indeed, that’s one problem that needs to be dealt with early on: the appropriation and potential misattribution of Citizendium content around the web. This example (which is no longer visible to the public or even Wikipedia’s administrators, having been Oversighted to respect the GFDL) highlights a major problem for the site: Citizendium’s success relies on its unique content, and that’s threatened by any significant use at other websites — including Wikipedia.

Each site clearly needs some way of dealing with the other’s licensing terms, and a MediaWiki extension currently under development by the CZ crew seems like a step in the right direction

In the meantime, however, Citizendium needs to live by its own ideals. Numerous new pages at the site, such as Language Attrition, note that the page was imported from Wikipedia only in the first edit summary, making no attempt to inform the user when simply reading the article. This is despite a recent call for caution, which — like so many other attempts at communication there — has failed to make its way to actual contributors:

January 24, 2007

Please make sure that if your article borrows from Wikipedia, you make a note of that down by the bottom if the little Wikipedia notice isn’t there. We don’t want to get caught accidentally breaking the GFDL. —ZachPruckowski 11:47, 24 January 2007 (CST)

The Jack-of-all-trades complex

Whereas Wikipedia has well over one million articles, Citizendium boasts a million irrelevant features that promise to improve nearly every aspect of editing a simple wiki. Features like Topic Informants, a ridiculous Approval process and a pointless logo contest all draw attention from Citizendium’s only current problem: producing content.

Citizendium boasts an Approval process — however ludicrous it may be — yet has just a few hundred articles to even consider approving. Topic Informants (a laughable concept all by itself, but I’ll leave that for another post) promise to make libel at Citizendium impossible, without having any content with which to commit libel. And so on, and so forth.

The website needs to step back and work on just one thing: producing content to rival Wikipedia. All these side projects need to wait until the encyclopedia itself has reached at least a minimal level of maturity, because they’re ineffective and pointless until that time. The effort spent on them is wasted for now, and ultimately does nothing but hurt the fledgling website.

Technorati Tags: , , , ,

Advertisements

4 Responses to “And yet, still so far to go”

  1. Mike Johnson Says:

    Davey,

    The phrases you throw around with such a cavalier attitude–
    “the problems are still legion”
    “flawed beyond belief”
    “half-baked”
    “mangled”
    “The joke of a review”
    “Citizendium even managed to implemented this simple feature badly”
    –all take away from your message. Yes, you can use clever turns of words to insult things. But other than allowing you to feel clever, this merely obscures the complex issues you’re speaking of and hinders your readers’ nuanced understanding of the subject.

    You’re a smart fellow with some good points, worthwhile suggestions, and interesting statistics. Don’t throw away the credibility that comes from that by writing like a partisan hack. You’re better than that.

    Mike Johnson

  2. daveydweeb Says:

    So, I understand that you disapprove of the language I used.

    What about the content?

  3. Mike Johnson Says:

    The content is good. I just saw your thread on the Citizendium forums– I’ll respond in depth there.

  4. David Tribe Says:

    Re content:

    For the record many things you’ve said about the approval process and CZ are arguably wrong or misjudged, but your statement about no page linking to “To Be
    Approved ” articles is factually wrong, and your link goes to the talk page rather than the article page for this function. In short you’ve messed up internet documentation of an issue that you are factually wrong about

    At the page you apparently don’t understand, which provides display of To Approve tagged articles at the moment I see the following:

    Pages in category “Articles to Approve”

    There are 3 articles in this category.
    C

    * Citizendium Pilot:Approval Process

    T

    * Talk:Biology/Draft
    * Talk:RNA interference

    I suggest you post a correction
    David Tribe, Biology editor Citizendium


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: