This is a short essay on some surprises that I recently unravelled on Wikipedia that caused me considerable stress over the past two years until I figured out what was going on.
You see, my main exposure to Wikipedia during this period of time came from “articles for deletion” (AfD) discussions – specifically, from AfD discussions on obscure programming languages articles like Nemerle, Factor, Alice, Cat, and Pure, which I was referred to by programming.reddit.com, which is a popular news site in my field.
Two years ago, when I looked at my first AfD discussion, here’s what I saw:
I saw a forest of remarks, some short and some lengthy, most of which were prefixed with a bolded word of the form “Keep”, “Delete”, or “Comment”.
As I read further, I learned that these bolded words are commonly called “votes”.
Imagine my surprise, then, when I learned that these “votes” are not counted and, worse, that (potentially partisan) efforts to raise “voter turnout” are actively discouraged!
When I tried to learn more, I discovered that AfD discussions are closed “based on community consensus”.
However, this left me even more confused than before!
You see, I’m passingly familiar with (Quaker-style) consensus-based decision-making and, one of the things that I’ve learned about it is that it
Unfortunately, the AfD discussions seem to have none of these characteristics! Instead, AfD discussions routinely discard participant’s views when those views are deemed to be ill-formed, uninformative, illogical, or when the views are presented by users who are not editors-in-good-standing.
In any case, as a result of these contradictory pieces of evidence, I concluded, in the gentle words of a dear friend, that Wikipedia governance (at least so far as article deletion was concerned) “was some bizarre sham and that [I] clearly was not going to be able to figure out what these crazies were trying to achieve…”
Next, if you read further into the guidelines for deletion discussions, you may eventually learn that AfD discussions are closed by an IETF-style “rough consensus”.
Unfortunately, as will surprise none of you by now, this claim is also deeply misleading. You see, as it happens, I am passingly familiar with IETF-style rough consensus and, unlike AfD discussions:
the consensus decisions made by IETF WGs are generated over long periods of time – sometimes, months to years,
IETF Working Groups are led by an “area director” (AD) who decides when consensus is reached, and these ADs are fairly static and well-known people to the participants in the working group,
for the most part, everyone participating in an IETF WG knows everyone else and they further know that, mostly, they’re all going to be working together for years to come, and
finally, and perhaps most tellingly, the arguments that are advanced in IETF WGs are intended to be based on topical expertise!
So how should we explain AfD discussions?
As lawsuits.
Here’s why. As with the law:
good arguments are those which are made with reference to pre-established Guidelines (viz. pleadings, motions) but this is not apparent from the outside and
non-lawyers often hurt themselves when they argue their own suits and
the decisions are handed down by a presiding Admin who personally weighs the presented arguments, evidence, and votes in order to decides the outcome of the discussion.
Maybe someone actually involved in Wikipedia could help other people see these things. Dear reader, maybe that can be you?
Through my own exploration, there’s one fewer confused Wikipedia reader in the world; hopefully now that I’ve written this, there can be a few fewer.