Project:WMF Steward study
Draft by Abd
-  is an announcement on the Wikimedia Forum of a study of antispam practices which was compiled on meta. The study was never completed, because the pages were, after weeks of work, abruptly deleted and suppressed (oversighted), as shown in the Forum archive linked. I was told that if I even requested undeletion, I'd be blocked.
Why is this of interest? Yesterday, I noticed that the new global account policy is eliminating the ability of users to register an account at a local wiki, without creating a global account. So what?
Well, local accounts cannot be globally locked, SUL accounts can. A global lock is a highly intrusive tool. A user whose account is globally locked cannot access their account at all. They cannot change watchlist and email notification settings. Log-in fails. Local administrators can do nothing about a global lock. Unlike the case with global blocks (which are only for IP addresses), there is no local whitelist allowing local administrators to make an exception.
So who cares? Aren't global locks only used for cross-wiki spammers and vandals?
Almost entirely, but not entirely. Global locks have been used for critics of the Wikimedia Foundation. They have been used for one globally banned user, and the closer of that ban did think that local wikis could defeat the global lock he imposed by renaming the account. The new SUL system removes the right of local bureaucrats to rename the account locally.
Because I found that antispam tools were occasionally used for non-spam purposes, I started a study of antispam practices. Initially, I had certain problematic cases in mind, such as the case of Denise Santarelli. However, how much of an exception was that case?
Very much of an exception. I studied the last 5000 account locks and unlocks from 2013, and found only a handful that were questionable. (There may be more, but they did not stand out, most lock actions, examined, seemed likely to be reasonable.) What did stand out was a very few account locks of users with no apparent history of spam or vandalism, and certainly no global ban (which is a difficult process, deliberately so).
I asked whether the impact of the new account rules, to be implemented September 15, had been discussed. I got the kind of hostile response from a steward that has become disturbingly common. Permanent link.
The steward is, essentially, trolling. If I respond to the gross assumption of bad faith, with just about anything, I will be blocked, it's a near certainty. I will take the matter to the Forum, which is mostly neglected. I may, instead, take it to RfC, which is almost as bad.
The steward's comment assumes that there is a usable process for appealing problem locks. There is. There have been accidental locks, a few, and the appeal process works. But if a lock is a deliberate action by a steward, based on a personal agenda, and effectively a global ban, and this shows in the evidence, appeal is useless and even dangerous.
The antispam study was designed to be a neutral study of antispam practices, so that practice and policy could be compared, and without an assumption that policy was correct and practice was not. In fact, actual practice routinely differs from policy, and it is likely that policy should be fixed.
However, there are a few exceptions where an attempt to fix policy in a way that would legitimate certain actions and practices would cause a firestorm.
It's very clear: stewards have become insular and above community restraint. Before the community could even think of restraining steward behavior, it would need to know what is actually happening, and it was study of that which was not just deleted but suppressed, a gross abuse of the oversight right.
The way many Wikimedians, especially Wikipedians, think is that if someone goes to as much work as I did to study this matter, I must be out to harm them.
I had made a couple of unlock requests, and I saw what happened when a steward's target requested unlock. I will compile evidence on those cases here. There are a few examples where the violation of policy was noticed, not just by me. The stewards ignore it. One mentioned that something was questionable. He is new, and dropped the subject.
It used to be policy that stewards would communicate on-wiki except where privacy was necessary. They dropped that long ago.
I reported the deletion of the antispam study on the Forum, as linked above. Silence.
If the community doesn't care, this is guaranteed to get worse. And the community didn't care with that deletion/oversight, not enough to do anything about it. So I'm doing this here mostly to wash my hands of it, and because someone -- who may no longer be interested -- asked me to provide it.
It's entirely up to the community if it wants fair, safe, sane governance, or not. It will get what it stands for.
The studies were working notes, not finished documents, they were largely raw evidence, not processed and boiled down to show conclusions. Study of steward activity is not allowed.