Self-reversion

(English) This is an essay. It expresses the opinions and ideas of some Wikimedians but may not have wide support. This is not policy on Meta, but it may be a policy or guideline on other Wikimedia projects. Feel free to update this page as needed, or use the discussion page to propose major changes.

Self-reversion is the practice of making an edit, so as to show editors what revision one proposes, and then reverting oneself immediately afterward. It is rather pointless. If the editor is not sure whether his revision should stick, he may as well create a fork (e.g. on a subpage) so that he and/or other editors can make further revisions to perfect the fork. Unless a fork is created, any further changes to the draft will require making two more revisions (one to make the revision and one to revert it).

If the editor does believe his edit should stick, then he may as well leave it in place and see if anyone reverts it, per BRD. If it is a page to which edits are not permitted without prior approval, then one may as well refrain from editing it since even if self-reversion is done, it can sometimes be construed as flouting the rule. It would be better to make an edit to the talk page in such cases, and create a fork if illustration of what one has in mind is needed. Leucosticte (talk) 01:02, 27 January 2014 (UTC)[reply]

Thanks for sharing, Leucosticte. Self-reversion is useful to make it very easy for other editors to see an edit in context, they may look at it in the diff or in a permanent version. It is then one-button to restore the edit. Self-reversion has a number of places where it has been shown to be useful.

Edits to policy pages where it is anticipated that there may be a process objection. I did, once, see an administrator complain about a self-reverted edit to a policy page. That administrator was shortly desysopped, being unclear on many concepts.
Edits under a topic ban or even a site ban. This is, of course, controversial. However, it has been used by a number of topic banned users to make constructive edits without complicating ban enforcement. It is far more efficient than the normal advice to suggest the edit on talk, and the result of self-reversion, when properly practiced, was (1) constructive edits, accepted and representing cooperation between the community and the banned user, I saw a user reverting an edit back in who was the user who had requested the topic ban! (2) demonstration of cooperation both with a ban and in making positive edits. It's totally useless to make disruptive edits and self-revert, unless an editor just wants to create a fuss by, say, making a grossly uncivil statement and self-reverting. That doesn't work. I've only seen it tried once. Self-reversion under ban doesn't necessarily work if the active community follows "a ban is a ban is a ban." But it doesn't do any harm, either. Where I saw it fail to work is where an administrator used revision deletion to hide the edits so that ordinary users couldn't then see and approve the edit. And that's obviously a case where the "ban" has become more important than improving the project.

Leucosticte has been a wiki editor for longer than I, however, his experience has been rather pointless, itself, and I wonder why he bothered creating an essay to claim that self-reversion was "rather pointless." I've seen it work many times, that is, accomplish improvement of a page while avoiding disruption. Wikis allow any user to edit, straight. Self-reversion is a device for efficiently *proposing an edit* while leaving the page at status quo ante. It evinces a respect for seeking consensus. It's been claimed that if the edit is good, why self-revert? Following the above, two reasons:

I may think that a change is desirable, but I might also think I could be wrong. I just self-reverted the insertion of a link on a policy page to a standard notice I created. It turned out that a standard notice already existed, merely linked from an unexpected subpage. So my self-reversion left nothing that needed to be fixed. Nobody had to revert me. In another example, I created a list of globally banned users, linked it from the ban policy page, and self-reverted. That led to a different solution, linking to the RfC of the only globally banned user. The list, on a subpage, might or might not be useful someday, it lists all RfCs that requested global bans, and gave the outcomes. Again, the self-reverted edit triggered examination and page improvement.
The second reason would be if the edit would violate policy if not self-reverted. One might think that one shouldn't make the edit in the first place, if that's the case, but situations arise. One might simply have a second thought and decide that the edit was a Bad Idea. A quick self-revert might avoid sanctions. The other example is that the user is blocked or banned. Some administrators will ignore the self reversion and apply sanctions anyway, but others won't. I suggested self-reversion in a case where I was an administrator, and I monitored the user's contributions carefully. He was locally blocked, but not actually banned, but he was globally locked at the time. This was a rather well-known user. The series of self-reverted edits, reviewed and accepted by other users, demonstrated cooperation. (Self-reverted edits in this case will have an edit summary that says "self-revert per block" or "self-revert per ban." The self-reversion leaves no disruption in place, i.e., enforcing administrators do not need to revert the edit, they can actually follow the last part of RBI. Why block an editor who leaves no mess? Obviously, there are different opinions about this, but, bottom line, the editor in question was unblocked, the global lock removed through renaming (that may not work any more), and the existence of positive contributions helped convince the community that there was value in acting to allow the user to edit.

It was proposed on en.wiki, by me, on the ban policy page, that self-reverted edits, not otherwise disruptive, not be considered ban violations. That suggestion had been approved by an arbitrator. There was no comment on the proposal on the policy talk page, and there it sat for some time. However, then, I was topic banned by an administrator, in a situation that eventually led to the desysopping of that administrator, it was really improper. While under a strict ban from him, I saw an error in wikitext that was causing page display to break. So I tried to fix it, and self-reverted. I was blocked immediately. An adverse editor who normally was opposed to my working on that article, reverted my edit back in, but the page was still broken (as could easily have been seen by looking at the permanent version before my reversion). However, then, the error became really obvious and was fixed, promptly. It had been there for quite some time.
Basically, self-reversion worked for improving that page. That same administrator had opined, before, that banned editors could make simple corrections and it was stupid to block them for it.....
What had not been understood by some was that "harmless edits" complicate ban enforcement. Self-reversion addresses that, if it were normal to essentially ignore self-reverted edits -- unless one sees the value of the edit and pulls it back in. I.e., normally, ban enforcement cannot improve the wiki, but ban enforcement means watching the edits of a banned editor. If they are all self-reverted, done, quickly, no action needed. But if the enforcing admin sees the edit, and it's a harmless spelling correction, for example, the admin can undo the self-reversion, thus converting some part of enforcement labor into actual improvement, and ... mirable dieux, the banned editor and the enforcing administrator are cooperating. That is no small benefit! It actually does build good will. But if the enforcing administrator thinks "a ban is a ban is a ban," and blocks and starts using range blocks, revision deletion, or the edit filter to prevent an banned editor from self-identifying in the edit, i.e., an edit filter is written to block and report edits with "self-revert per ban of XXX," the IP is then blocked, while the edit has been prevented. Essentially, in the name of ban enforcement, ostensibly to protect the wiki, the admin has prohibited the banned user from identifying himself. What happens next is totally predictable. The editor stops identifying himself, and stops self-reversion, perhaps then registers an account and starts socking, and cannot identify himself, again, because he'd be blocked. Then much effort must be expended to enforce the ban. Somehow the logic of this has escaped many. Failure to maintain an open channel for cooperation leads to conflict, disruption, and wasted labor. It's really that way all through life.
Anyway, once I actually had used self-reversion once, and this was, of course, taken to AN/I, comment poured into the ban policy page that self-reversion was Bad. "A ban is a ban is a ban." An editor who had been using self-reversion successfully, to improve the article he was banned from, with the collaboration, as I mention above, of the editor who had requested the ban, was told to not follow the bad advice of that fellow Abd. And he stopped. And apparently he then turned to sock puppetry and was site banned. Cut off the means for positive cooperation, don't be surprised if something else happens. The guy was a subject matter expert, actually famous on the topic he was banned from.
In my view, Wikipedia shot itself in the foot. It's a minor example, though, much worse happens every day. Not my problem. --Abd (talk) 06:09, 28 January 2014 (UTC)[reply]