Topic on User talk:Ostrzyciel/Flow

Naleksuh (talkcontribs)

Hi there! I heard you are looking for feedback on the "manual revert" tag? If so, let me know! Interested in some thoughts on the tag.

Ostrzyciel (talkcontribs)

Yes, I am :) Any thoughts are welcome.

Naleksuh (talkcontribs)

I think that the the usefulness of the tag in general is good, but the name should be something other than "Manual revert". It isnt checking that steps were taken to revert edits, just that the edit is exactly the same as a previous version. There have been cases of "false" uses of this tag such as en:Special:Diff/971020476 in which I didn't even know the previous edit was adding this tag, I was just going into the page and removing it. So this edit shouldn't have been a manual revert as I had no idea it was even a direct revert of the previous edit. There should be a new name which clarifies that its the same as a previous version, but doesnt use the word "revert" (hard to come up with one short enough). However, a new revert tag, which, checks that edits are being reverted via the old id method, would be useful. Also, why does it 'time out' after some revisions? Surely with page hashing it would be feasible to check all of them.


Hope this feedback is helpful to you. Not trying to be destructive of efforts but improve mw :)

Ostrzyciel (talkcontribs)

Yeah, when discussing this we realized that the name may be confusing to some communities, especially Wikipedia, we didn't have a better candidate, though. It would also be close to impossible to guess what is the intent behind an edit, that is a good topic for a psychological study :D I'm still looking for proposals on how the name could be changed, so let me know if you come up with something. We will see, maybe wikipedians will just get used to it.


The limit is there because the revision table lacks an index over rev_sha1 field and introducing it would be a lot of hassle. Also, it's good to limit the search, as with very deep reverts you may get really weird and unexpected results. 15 is actually a number that was suggested to be "optimal" by this research paper: https://md.ekstrandom.net/pubs/isw09-wiki-viz.pdf

Reply to "manual revert"