Wikipedia talk:Reviewing guideline

Initial comments
I think we should cut down a bit on what reviewers should do. Wikiproject banners? Let those who care do that. I also disagree with "there is no urgency to patrol edits". Editors expect their changes to go live as soon as possible. We do not want to change the BOLD, revert, discuss cycle into "BOLD, wait, discuss, wait, maybe flag". Flagging is supposed to be a quick check for the worst problems, not a featured article review. --Apoc2400 (talk) 11:12, 2 April 2009 (UTC)
 * As it's passive, patrol won't have any effect on edits going live, while on the contrary there is a certain urgency to review edits on semi flag protected pages. As for project banners, it's just to make the article within the scope of at least one wikiproject. If an article has no banner, then those who care cannot notice the article. They don't have to make the assessment, but low-watched, or orphan, articles suffer from lack of monitoring, being within the scope of a wikiproject will mean that some people will be aware of the article and take care of it more that it would be otherwise. The first patrol is special and needs a particular attention, I think it should make sure that the article is going to be monitored the best it can. Cenarium (talk) 12:11, 2 April 2009 (UTC)
 * It should be clear what applied to passive patrol and what is for flagged protection. In my experience, the WikiProject tags do not attract extra attention to an article at all. An other thing: Should "A revision should not be patrolled if it" be changed to "An edit should not be patrolled if it"? Surely we do not expect reviewers to check the whole article every time an edit is patrolled? The exception would be the first time of course. --Apoc2400 (talk) 13:14, 2 April 2009 (UTC)
 * Generally, it'll be based on the diff to the latest patrolled version, most of the time, several edits will have been made in between. Cenarium (talk) 20:13, 3 April 2009 (UTC)

Non-approval
If a reviewer comes across an edit which, e.g. contains vandalism or contradicts BLP, what should they do? Revert it? AndrewRT(Talk) 22:06, 9 April 2009 (UTC)

Seems a little complicated
I'll repeat a bit of what I said about the related earlier proposal:

How about, instead of a detailed guideline, just "edits should only be approved if they improve the article". It should be as simple as that, for semi-flagged protection. Any edit that an established editor thinks should be left as-is without reverting or modifying should be approved. Edits that have value but are flawed don't need to be approved (although they can be), because instead editors can simply adjust them (creating a new auto-approved version). Bad edits can (and should) be reverted. With full flagged protection, it should be, "edits should only be approved if they reflect consensus or are uncontroversial improvements" (i.e., edits that it would be acceptable for an administrator to make on a fully protected article).

I particularly object to "Edits should not generally be left unconfirmed on other grounds, such as containing Original Research, not being neutral, or not being adequately verified". Apoc2400 had it better at Wikipedia talk:Flagged protection approval with "Reviewers are not required to check for Original Research, neutrality, and verifiability and should not be held responsible for approving such edits." The implication there is that none of those problems should be approved if editors are aware of them, but that they can and will slip through because approval is only a quick-and-dirty screening to prevent even worse problems. No one should feel compelled to approve an edit they find problematic, just because its problems are not as blatant as they might be.--ragesoss (talk) 17:06, 11 April 2009 (UTC)


 * I think of it as being a bit like the difference between the Deletion Policy and the Criteria for Speedy Deletion. Rather like CSD, this is a way of quickly catching the clear cut cases of vandalism and unsourced negative claims about individuals. It's not about catching more grey areas or making difficult (and lengthy) judgements about "good" or "bad" edits. I favour a very strict and narrow criteria on non-approvals: only clear violations should be blocked through this process. Editors can add, correct and improve as normal and of course the person reviewing the article can do this at the same time if they want. But that should be done outside of this process. AndrewRT(Talk) 17:17, 16 April 2009 (UTC)

Trial guideline?
I take it this is a working draft of a proposed trial guideline. If it ever is made a guideline for the trial period, I assume it will cease to be a guideline when the trial period is over. Is that correct? --Pixelface (talk) 21:52, 13 April 2009 (UTC)
 * Yes. More precisely, it will depend on the discussion we'll have on continuing the implementation or not, default being deactivation. If we stop, then's it's marked historical, if we continue, we can make this into a guideline. Cenarium (talk) 07:13, 16 April 2009 (UTC)
 * Yes and no. It's a trial of course and it will only be immediately relevant for the trial. However, after the trial, I would presume it would be the basis for any future implementation that was decided - although after an opportunity for evaluation and further discussion. AndrewRT(Talk) 17:26, 16 April 2009 (UTC)
 * My concern is that if this page is made a guideline for the length of the trial, that after the trial is over, this page's status as a guideline will be cited as a reason to continue Flagged protection and patrolled revisions (or any other implementation of Flagged revisions) "since we already have a Reviewing guideline." --Pixelface (talk) 05:59, 17 April 2009 (UTC)
 * That would be a silly argument, and would have no weight whatsoever. I'm also thinking to merge this in Flagged protection and Patrolled revisions, and turn those too in 'trial guidelines'. Cenarium (talk) 18:07, 19 April 2009 (UTC)

Merging in Flagged protection and Patrolled revisions
To keep the number of project pages directly related to WP:FLP/PR under control and also because, as pointed above, it may be unecessary to have a standalone guideline for reviewing (in the sense patrolling, confirming, and validating), I think it should be merged in those. They could then be turned into guidelines, and we could also create /Help subpages for both of them. Flagged protection and patrolled revisions would be the general documentation page. Reviewing guideline page could redirect to Reviewing that could be made into a disambiguation page. What do you think ? Cenarium (talk) 00:19, 20 April 2009 (UTC)
 * Well, I created this page so it's probably no surprise I disagree! I think the success of this process will lie in details like this - we need an effective way to reassure people that over-zealous reviewers wont go round rejecting edits without good cause. It's easier to keep the discussions separate for now - given that we have a navigation template now. AndrewRT(Talk) 00:34, 20 April 2009 (UTC)
 * The guideline will still exist, but merged in Flagged protection and Patrolled revisions. I've copy/pasted the content in sections there, those page are still in draft state. I think it's better to centralize the content, as it gives context and also highlights the differences between patrolling and confirming/validating. Cenarium (talk) 00:42, 20 April 2009 (UTC)

Too vague to mean anything
Edits should not generally be left unconfirmed on other grounds, such as containing Original Research, not being neutral, or not being adequately verified, except where this contradicts the policy on the biography of living persons, if the edit introduced many such problems, or the violations are egregious. This essentially translates to "the reviewer can do whatever (s)he wants". So what's the point of this guideline? It says nothing. Better focus on a credible election process for the reviewers. Pcap ping  09:24, 31 August 2009 (UTC)