User:Cenarium/Flagged revisions/Examples

This is a list of examples of use of sighted and confirmed revisions, the problems they may cause and how to solve them.

Example 1
Let Foo be an article on a football club. Suppose Foo has sighted revisions enabled for some reason, and that User:A, who happens to be a surveyor, has edited this article on a regular basis. Now suppose IP:B (or any non-surveyor) makes an edit on Foo, and that this edit is sighted, because it is not vandalism, and not a flagrant violation of our policies. User:A reverts this edit. The revert is then automatically sighted (A being a surveyor). IP:B reverts this revert. User:C, a surveyor, finds this last revert on Special:OldReviewedpages and is puzzled as to whether it should be sighted, because it may well be the beginning of an edit war... User:D, following the sighting guidelines, sights this last revert. User:A then de-sights this edit, with the rationale "this player was not hired for so much money, it's false information" causing it to come back on OldReviewedpages. However, it doesn't look obviously false and User:A has not presented a source to backup the claims. User:D then warns User:A that per the guideline, surveyor rights should not be used to gain advantage in a content dispute, and should instead revert or, better, bring sources and discuss. User:A keeps de-sighting despite warnings, or has already a past of such behavior. Admin:E is made aware of the situation and removes surveyor rights of User:A.

Analysis
We are in a classic case of limited content dispute. User:A has repeatedly abused surveyor rights, and had them removed. The situation has now to be resolved by usual dispute resolution channels. However, removal of rights may not always be a viable solution, as seen in Ex. 2.

Alternative
Admin:E may think that sighted revisions are no longer needed on this article (not related to a current event, weak vandalism levels, etc), or was set to expire soon anyway, and that User:A is a prolific editor, and so removing surveyor rights would cause a rise of edits to review. Admin:E decides to disable sighted revisions on this article.

In other circumstances, Admin:E may decide both to disable sighted revisions and to remove surveyor rights to User:A.

Second analysis
In any case, the situation is drawn back to a classic content dispute, with classic remedies. Note that this case was problematic because editors with different access levels were involved. If only non-surveyors are involved, then having sighted revisions or not won't change the deal. Same if only surveyors are involved.

Example 2
Let A be an article on an active US politician. Suppose A is semi-protected, and sighted revisions enabled. The article is highly contentious and subject to edit wars on a regular basis. Edits again consensus are frequent, and some are damageable to the subject (blp violations), even if not always flagrant for a random surveyor. A great number of users edit this article, among them many are surveyors, and they edit a large quantity of articles together. Several users have pointed out that the surveyor status gave a substantial advantage in such a controversial article (automatic sighting especially), that it allows breaches of WP:OWN. Others complain that some edits are inappropriately sighted, and some are not while they should. Sighting wars happen frequently, this led to removal of rights on several occasions. Some users whose rights have been removed are productive or recognized editors, this led to a surge of edits to be reviewed. Some users claim that sighted revisions cause more trouble than good, but others argue that having this article without sighted revisions in the middle of such a high media attention would be terrible, that all these blp violations and occasional vandalism (that could only rise, attracting various 'predators') would have a disastrous effect, etc, etc.

Finally, it is proposed that confirmed revisions be enabled on the article, and it is accepted by consensus on ANI. Thus, only the latest 'confirmed' revision will be showed to IPs. Admins or certain trusted, neutral users ('moderators'), will confirm revisions as long as they are non-controversial or consensual. The users with otherwise good contributions get their surveyor rights back. Article development continues, and in the same time, non-consensual or disruptive edits are not showed to IPs. Edits are confirmed in a timely manner, since the number of moderators is high enough and this article attracts a lot of interest.

Analysis
In this case, we see the limits of sighted revisions. The recourse to confirmed revisions makes sense and is unlikely to cause problems.

RE: Example 1
Suppose that in this case Admin:E has removed surveyor rights to User:A (and possibly issued some blocks), and that the situation is still unresolved. A User:F then proposes that we use sighted revisions to control the dispute: reverts won't be 'sighted' any more. User:F is unanswered and is bold, making a null edit "don't sight any more, see talk". Depending on the last sighted revision, User:A or IP:B claims that it's unfair their edit isn't sighted, etc, etc. Some surveyors complain that the edits remain at Special:OldReviewedpages and they waste their time hitting it. Some others remark that it didn't do anything to help the situation.

In most cases though, we won't have so many commentators around this kind of, nonetheless frequent, situations. The above situation could evolve in different scenario:
 * 1) A consensus is reached on the talk page. Then the edit is sighed and life goes on. (easy case)
 * 2) Edit war continues, blocks are issued, other editors become involved... Now suppose one of them, user:G, is a surveyor and the latest edit hasn't be sighted. So the edit is not automatically sighted, suppose User:G sights it. Is it a violation of sighting guidelines ? It's not vandalism and not a flagrant violation of our core policies. What will do the users watching the article, when the article is edited, will sighting resumes, or not ? That would be an advantage for the surveyor. This is one of the problems disputes will create on articles with sighted revisions. It is not clear that the sighting guidelines have been violated, a removal is not justified.

Analysis
Thus, we have two choices:
 * Disabling sighted revisions
 * Enabling confirmed revisions

It will happen more generally when sighted revisions on a certain article is a source of problems that cannot be handled by removing surveyor rights. We'll have to determine whether it is 'safe enough' to remove sighted revisions, and if not, we have to use confirmed revisions.

Example 3
A series of articles, subject to an arbitration case, have for some reason sighted revisions enabled. Suppose we repeatedly have sockpuppets edit warring on those articles. Suppose they are determined. Some of them have managed to get surveyor rights. There are three possible ways to do that: by automatic grant, but the requirements are, hopefully, relatively high (say, 700 edits, no blocks, at least 50 articles edited, etc), by request on review by an admin, and as a subsidiary of rollback rights (also on request) (surveyor rights come along rollback for technical compatibility). Thus, a sock could choose to become a vandal fighter to be granted rollback rights, or make good impression overall to be granted surveyor rights. It may be easier than automatic assignment, but a positive is that an admin will look after the user's edits, and generally a few times after the grant also. Now, you see that the surveyor access can be abused this way, though users have to be determined because this kind of behavior is generally rapidly spotted and the work needed is quite high (thus, it'll be overall positive for Wikipedia).

However, incidents will still occur regularly. In such cases, it may be appropriate to enable confirmed revisions temporarily, as an enforcement to the arbitration case.

Temporary conclusion
Sighted revisions are a useful tool to deal with vandalism and blatant violations, and their usage should be limited to this. As a consequence, any edit that is not vandalism and not a blatant violation must be sighted. Indeed, attempting to use them for other goals will inevitably lead to problems and reduce the controllability of sighted revisions and surveyor's behavior. They are cases where sighted revisions will be problematic, essentially during disputes, in those case, confirmed revisions will be used. It'll potentially avoid full protection, just like sighted revisions will avoid semi-protection on certain articles.