Jump to content

Wikipedia:Pending changes/Feedback/Archive 6

Add links
From Wikipedia, the free encyclopedia
Archive 1←Archive 4Archive 5Archive 6

Should make an excellent Anti-Sockpuppet tool

This tool should be a boon for pages constantly victimized by sockpuppets. Being able to institute Pending Changes protection for a page, instead of having to resort to semi-page protection is a great way to not bite the newcomers, while still protecting the integrity of the project.

I like what I've seen of Pending Changes so far. Good work with this fine addition to the code. BigK HeX (talk) 23:03, 12 August 2010 (UTC)

The tool will be worthless against a savvy puppeteer. —Jeremy (v^_^v Carl Johnson) 23:44, 12 August 2010 (UTC)
How so? Even the most savvy puppeteer eventually gets found out. Same IP registers another account, it gets watched quickly. CycloneGU (talk) 04:55, 13 August 2010 (UTC)
Yes, agreed. From what I have seen, the greater bulk of petty vandalism and stupidity comes from one-off flybys, throwaway accounts and anon IPs. Although a clever sockmaster may be able to defeat it, this tool helps defeat enough of the throwaways and not-so-savvy puppeteers. Any new tool that helps in any small way is a move in the right direction, say I. Kindzmarauli (talk) 08:06, 13 August 2010 (UTC)

Closure

Time to make a decision since the two months are over tomorrow, please see Wikipedia:Pending changes/Closure, we need to organize the discussions. Cenarium (talk) 11:28, 14 August 2010 (UTC)

Page-specific suggestion

Might I suggest creating an edit notice for this page specifying the existence of the closure discussion for comments? Tyrol5 [Talk] 01:06, 18 August 2010 (UTC)

A stopgap measure

This is better than nothing (except in the case where an articles protection level is downgraded), but it is a stopgap to the inevitable day when Wikipedia not only requires registration, but has a waiting period between registration and the ability to edit. Wikipedia will still be the encyclopedia that anyone can edit, but not capriciously. As others have pointed out, we are in the next phase of the project; we have an encyclopedia, and one that many others rely on. It is our collective duty to get it right. One one hand, there are many unregistered people who have contributed to Wikipedia, but how many would have not done so had registration been required? I don't know, but as Wikipedia matures my thought is that the project will lose credibility unless casual vandalism is reduced. Note that I am not recommending verification via email; the current process is sufficient. A short (24 hours) waiting period between registration and editing will do so much good and very little harm in my opinion, and we could dispense with this scheme. In the meantime, I do support he continuation of the Pending Changes program. SeaphotoTalk 00:41, 4 August 2010 (UTC)

WP:PEREN#Prohibit_anonymous_users_from_editing. ♫ Melodia Chaconne ♫ (talk) 01:45, 4 August 2010 (UTC)
I understand, and have read it before. Maybe it is the time I spend reverting vandalism, but it feels like it is getting worse. Recently, I have seen coordinated attacks on articles from multiple IP's that simply could not have happened if registration was required. The 24 hour waiting period addresses one of the main arguments in the perennial arguments section, that it wouldn't help, and founding principals, while very important, shouldn't be a suicide pact. I know there is no consensus for that now, but I feel we will get there someday. If increased vandalism wasn't an issue, why have the pending changes process? For now, we do what we can with the tools we have, such as this one. SeaphotoTalk 02:11, 4 August 2010 (UTC)
Pending changes, if successful, will allow anonymous editors to contribute where they can't currently. From WP:PC This is applied on articles where such problematic editing occurred, as an alternative to restricting who may edit the article via full- or semi-protection. Gerardw (talk) 10:54, 4 August 2010 (UTC)
My watchlist has grown enormously this year due to the unreferenced BLP project, yet I'm not seeing as much vandalism on my watchlisted articles as I used to. This is because the vandalfighting bots are cleaning up vandalism faster and the amount of vandalism on the project at any one time is actually reducing not increasing. Yes there are and long have been coordinated IP attacks on articles, but pending changes can resolve that problem. Recruiting editors is and remains an important priority and that means allowing IPs to edit, several rival projects have been launched that have made it more difficult to do your first edit - as far as I'm aware so far all have failed to take off. ϢereSpielChequers 13:02, 4 August 2010 (UTC)
PendingChanges is not a panacea for /b/ attack threads, chummer. —Jeremy (v^_^v Carl Johnson) 18:49, 9 August 2010 (UTC)
Exactly. See Larry Sanger's comment on Wikipedia:Wikipedia_Signpost/2010-08-02/In_the_news. Now that WP is large and successful, it should change the thing that made it successful? That's insane. Gerardw (talk) 13:31, 4 August 2010 (UTC)
These are good points, and the policy of access versus vandalism is not going to change anytime soon. WereSpielChequers makes a good point about other online encyclopedias and their relative lack of growth. Institutions do evolve however, and I wouldn't suggest it is insane to challenge fundamental assumptions now and then LOL. One thing we do need if this is going to work is more reviewers able to check references in foreign languages. particularly in India. Looking the recent changes stream with Huggle, there are many edits about locations, organizations and educational institutions that seem suspect (to be fair, sometimes because of the quality of the language), but do not fall under the guidelines of unquestionable vandalism that should be reverted on the spot. SeaphotoTalk 17:03, 4 August 2010 (UTC)
I think Huggle is wonderful and I greatly appreciate the work that those who use it do. However I'm pretty sure that it works by bringing the most likely vandalism to the attention of the hugglers, and that as a result hugglers often get a jaundiced view of the contributions of IPs and Newbies. Occasionally I go to recent changes and welcome or warn a bunch of editors with redlinked talkpages. Usually I find myself dishing out sufficiently more welcomes than warnings to reassure my confidence in the general public. The reason I like pending changes is that it can work on the basis that suspect edits get marked as checked when one person has checked them, rather than as at present some edits being cross checked by multiple people and others by none. ϢereSpielChequers 11:44, 5 August 2010 (UTC)
Regarding foreign language references, it might be useful to permit users with reviewer permissions on any language Wikipedia to review articles on all other language Wikipedias as well. Not so much of an issue in the case of India, but with European languages for instance, it would be useful for French, German, etc. users browsing English language content to be able to approve it, and for us to approve theirs. This was briefly discussed elsewhere.
Essentially, once someone is trusted with reviewer permissions on one Wikipedia, that trust should extend to the entire project. --pmj (talk) 01:21, 25 August 2010 (UTC)

Usability: Rename Accept/Unaccept actions to state names

The usability of the "pending changes" is a nightmare, not just for the interface but because the model of how it affects the edit history is too difficult to understand. In order to clarify it, the model should be based on the possible states of a pending edit, not on the actions the reviewers take on it. Instead of accept / unaccept, allow the edits to be in one of these:

  • Pending
  • Accepted
  • Rejected

And provide buttons so that anyone can put it in the desired state. For example - if I want to undo the action of an Accepted edit, I will not "unaccept" it (ugh!), I will put it again as Pending. States that cannot be achieved given the current edit flow would be disabled.

I'd also recommend change the word "Accepted" for "Visible" or "Included", to discard the impression that the reviewer endorses the change while showing that it's available in the main article. Diego Moya (talk) 13:11, 27 August 2010 (UTC)

Initial list for "Pending Changes v2"

We're currently working out a timetable for improvements to Pending Changes as requested on User_talk:Jimbo Wales. The list we're working off of is on mediawiki.org here: mw:Pending Changes enwiki trial/v2 Features. Still a work in progress as we gather the list of features and figure out what is reasonable to implement in the near term. -- RobLa-WMF (talk) 00:26, 14 September 2010 (UTC)

I Like It

I like being a Reviewer or Review editor and author more than look for articles that need help. I have created 400+ articles and edited countless times and have more than basic knowledge. I like a ranking between "anyone can edit" and "Administrator" as it takes some work off administrators' shoulders. Noles1984 (talk) 16:49, 3 September 2010 (UTC)

That's the problem with this whole thing. It doesn't take work off of anyone's shoulders. It just creates more work unnecessarily. If an IP vandalizes a pending changes article, we still have to revert it and then warn or block the miscreant. SchuminWeb (Talk) 19:37, 6 September 2010 (UTC)
But when its a good edit then the first reviewer to approve it can mark it as reviewed, instead of articles being looked at by all their watchlisters, and some being looked at many times and others not at all. So more efficient at finding vandalism and more efficient at using editors time. ϢereSpielChequers 17:04, 20 September 2010 (UTC)

Is it too late?

Is it too late to leave an opinion?

I would just like to say hat i found the process tedious and slow but especially slow because of the loading times. Has this been said before by anyone?

Simply south (talk) 13:58, 20 September 2010 (UTC)

Yes. Gerardw (talk) 15:06, 20 September 2010 (UTC)

Wreckless acceptance of changes

The concept as a whole seems okay, but I'd like to point out that I've seen at least one vandalism incident approved by someone that didn't bother to verify that an unsourced claim was true (see this link). My point is that if this is implemented on a large scale, it is important that people do not wrecklessly accept or reject changes, or the system will fail. PCHS-NJROTC (Messages) 02:15, 23 September 2010 (UTC)

This has been discussed. Per the guidelines, reviewing is a vandalism screen, not article quality checking: please see Wikipedia:Pending_changes/Feedback/Archive_4#Users_accepting_false_additions in the archive. Gerardw (talk) 03:28, 23 September 2010 (UTC)

Implementation of a "Pass" Button

I have noticed that the yellow "under review" highlighting in is generated when someone with reviewer privileges looks at the diff of a pending change in any form (it does not matter if you follow the review link). I have noticed sometimes people look at the diff and then move on and decide to do nothing, while the under review tag does go off eventually I'm pretty sure the time out rate is over 10 minutes. During this time the banner prevents other editors from reviewing and accepting/reverting the change. If a button to "pass" could be added to remove the yellow under review tag before the timeout that could help alleviate this problem. --nn123645 (talk) 15:05, 24 September 2010 (UTC) Nevermind, already suggested... --nn123645 (talk) 15:54, 24 September 2010 (UTC)

Naming of special pages

Hi all, I spoke with Chad and Aaron today about the proposed renaming of these special pages:

It turns out that this is more difficult for the November 9 release than previously realized. The problem is that we have one underlying technology ("FlaggedRevs") supporting two different configurations ("classic" FlaggedRevs and "Pending Changes"), but making two different names is difficult. The long term solution is to fork the FlaggedRevs extension, and actually create a "Pending Changes" extension different from the FlaggedRevs extension that uses different database tables and different language strings (MediaWiki messages); this is something we can do with more time, but not by November 9. In the meantime, we need to choose one of these two options:

  • Come up with names that work for both FlaggedRevs and Pending Changes configurations. This means coming up with names that are acceptable to the large community of users that are already using the classic FlaggedRevs configuration in addition to everyone here
  • Punt on this one until after we're able to fork

Which of these options should we consider? - RobLa-WMF (talk) 22:56, 20 September 2010 (UTC)

Punt now. Later, get a good consensus on the names English WP wants for PC, make code changes once correctly and methodically. Gerardw (talk) 01:48, 21 September 2010 (UTC)
Punting seems ideal, but I'm curious if there are easy solutions. The word Review is shared by both configurations. What about:
Those seem generally flexible and more intuitive. Thoughts? Ocaasi (talk) 07:59, 21 September 2010 (UTC)
I was investigating the different mechanisms we could use when I noticed that Aaron had made one of the changes (before we were discussing this in earnest). Here's what is there now:
Getting rid of "OldReviewedPages" was the biggest problem. Most of the others I think we can probably live with until we fork, though "StablePages" is one I could see being a problem. Thoughts? -- RobLa-WMF (talk) 20:39, 24 September 2010 (UTC)
That was the big one. The rest can wait, but would still ideally like them to be a bit more consistent and 'obvious'. I don't like Advanced Review, since I still forget what it means and can't tell just by looking. That's my metric. Would a new reviewer know which page does what just from the title. Ocaasi (talk) 18:19, 28 September 2010 (UTC)

Closing the Flagged Revision test wiki

The test wiki is going to be closed, and no upgrades will be loaded there: http://flaggedrevs.labs.wikimedia.org/wiki/Wikimedia:Pending_Changes_updates

A new environment is being set up here: http://prototype.wikimedia.org/flaggedrevs/

So far, it seems that there are only a dozen or so articles to play around with, and the most one can do is create an account and get reviewer rights. Not sure this is particularly useful in this configuration (no opportunity to test the admin interface or see how other advanced permission tools work with it), but let's give it a week for them to make it worthwhile. I figure it needs at least 10K articles, a resident couple of bureaucrats who aren't developers (I recognise they're too busy), and access to the full range of advanced permissions. Risker (talk) 21:28, 30 September 2010 (UTC)

There are no intentional policy differences between flaggedrevs.labs.wikimedia.org and prototype.wikimedia.org/flaggedrevs. If someone was an admin/bureaucrat/whatever on flaggedrevs.labs, they can be one on this site. Since it's not using production CentralAuth, it might be good to put the requests for admin page here on enwiki as a low-tech way of vetting that someone isn't spoofing someone elses account. -- RobLa-WMF (talk) 22:23, 30 September 2010 (UTC)
Thanks, RobLa-WMF. I don't know whether the 'crats from the lab wiki will be interested in continuing on; I held 'crat and oversight on the other wiki (am fully identified to WMF), and I did use suppression for some personal attacks that happened there, but I'll see if I can rouse another person to pair up. Most of the 'crat work the last time was done by MZMcBride, as I recall, but I have no idea whether he will be coming over. Risker (talk) 22:54, 30 September 2010 (UTC)

Bug?

I just reverted some pending changes. Now, as you can see from the history of that article (scroll to October 4), I am listed as "accepted by Lilac Soul". In other words, my initial reversal was flagged as "pending" (and I am auto-accepted, mind you), and I then had to accept my own revision. This must be a bug. I have noticed at least one instance in the past where a different reviewer had accepted his own revision rather than being auto-accepted, so this isn't a once-in-a-lifetime oddity. Perhaps the issue has to do with reverting multiple changes at once (there were three)? -Lilac Soul (Talk • Contribs) 09:25, 4 October 2010 (UTC)

If you make an edit when there are existing pending changes you have to explicitly accept it. Gerardw (talk) 02:52, 6 October 2010 (UTC)
Thanks. I noticed this to be the case yesterday. Still seems a little odd that the reversal wasn't auto-accepted as it usually is. Still, probably something on my part rather than a Wikipedia bug. I'll consider this closed for now and only report back if I again become convinced that something is awry. -Lilac Soul (Talk • Contribs) 04:48, 6 October 2010 (UTC)

Call for specific feedback on UI elements

There are a number of user interface issues we'd like to nail down for the November release of Pending Changes. Among those:

  • Bugzilla:25295 - Improve reviewer experience when multiple simultaneous users review Pending Changes
    • There are many complicated fixes we could implement that would address this problem, as well as simple fixes that would probably be of more superficial benefit. If there's one or two things that we really should implement by November, please weigh in on Bugzilla or here.
  • Bugzilla:25296 - "History style cleanup - investigate possible fixes and detail the fixes"
    • Our usability team has some ideas about changing the use of highlighting, but any further help would be greatly appreciated. What specifically is wrong with the way the history looks?
  • Bugzilla:25298 - "Figure out what (if any) new Pending Changes links there should be in the side bar"
    • Should there be a link to "Pending Changes" from the sidebar for those logged in with the proper access?
  • Bugzilla:25299 - "Make pending revision status clearer when viewing page"
    • Should we change the background color when viewing the pending revision of an article? Any other visual indicators?
  • Bugzilla:25301 - "Firm up the list of minor UI improvements for the November 2010 Pending Changes release"
    • Is there anything not on our current list that is "must do" for November?

Please let us know, either here or on Bugzilla, if you have any thoughts on any of these. Thanks! -- RobLa-WMF (talk) 22:22, 6 October 2010 (UTC)

Hey Rob, sorry about not responding to your comment on my talkpage. The issue that I had was that I was looking at an article and it seemed like most of the edits were getting broken into another row, which was just not aesthetically pleasing. However, it's not a big deal. You could shorten the "automatically accepted" message but I wouldn't worry about it too much. The clutter in general is due to, for whatever reason, these pages seeming to receive a high volume of edits. Maybe it's just that they are usually semiprotected. The fix, which is long overdue, is to allow for more screening/filtering of the history. Also, if we want to implement real quality-control, we should allow more than just whichever reviewer happens to hit the "accept" button to sign off on edits, and we should implement another option indicating a higher level of vetting. But that's a long-term idea. I don't really have much comment on your other questions. II | (t - c) 23:03, 6 October 2010 (UTC)
Thanks for commenting. In response to your request to allow for more screening/filtering of the history, please take a look at reversion collapsing. That's not on our short term roadmap, but is something we're considering. - RobLa-WMF (talk) 20:52, 7 October 2010 (UTC)

Announcment on German wiki

I don't think that announcing this change via the Cafè on German Wikipedia is useful. Hardly anyone looks there and generally it's occupied by fluff or non-serious side stories. I think it should be better placed on Fragen zur Wikipedia so people actually see it and won't go "you didn't inform us, you are doing things behind closed doors" afterwards. --94.134.199.250 (talk) 09:49, 16 November 2010 (UTC)

Sure, I guess I can do that. -- RobLa-WMF (talk) 17:58, 16 November 2010 (UTC)

Second: How can I test the new features when I can't get editor? – Giftpflanze (talk) 11:35, 16 November 2010 (UTC)

Good point. I've added those to all the test wikis in my original anouncement. -- RobLa-WMF (talk) 17:58, 16 November 2010 (UTC)
Erm, sorry, but the Polish test wiki produces PHP errors now (Cannot modify header information - headers already sent by) - http://prototype.wikimedia.org/pl.wikipedia.org/. Lampak (talk) 19:51, 16 November 2010 (UTC)
Gah, sorry about that! I did that in a rush and clearly forgot to check my work. -- RobLa-WMF (talk) 23:47, 16 November 2010 (UTC)

Third: There's no possibility to configure pages so that they can actually be flagged. – Giftpflanze (talk) 15:52, 17 November 2010 (UTC) Giftpflanze (talk) Link added --PhChAK (talk) 14:38, 25 November 2010 (UTC)

Hi Rob, considering your announcment on German wiki. Since yesterday we have some problems by looking for recent changes. We only can see the red or green marked differences - but not the preview version included the category, weblinks or something like that. --Laibwächter (talk) 13:01, 24 November 2010 (UTC)

The interwikis are also missing in the diffview. It's really difficult to work with this situation. This bug should be corrected as soon as possible. --Mazbln (talk) 19:22, 24 November 2010 (UTC)
We absolutely need a possibility to be able to view the differences AND the newer version of an article in the same window (as it was until last week). It is not possible to flag revisions quickly and correctly when one is obliged to click back and forward between the actual (or newer) text and the differences display - Please - please fix that SEVERE BUG ASAP !! - Thank you.
If this is not the right place for this comment please transfer it to a better place - thank you. --PhChAK (talk) 14:38, 25 November 2010 (UTC)

Problems using the API

Hi, since the change I have problems doing reviews via the API. Everything else like changing pages still works fine, so I suspect that it has something to do with this change. Did you change anything in the API / or the policy? I am using the pywikibot to review changes directly over the API but I suspect that this behavior will also occur with other programming languages/frameworks. I also posted the problem here: https://bugzilla.wikimedia.org/show_bug.cgi?id=26110

In practice, I get an 500 Internal Server Error when I try to review an oldid, whereas trying to edit with a bad token or no token produced a correct "badtoken" or "notoken" response.

to reproduce, try this: 

$ svn checkout http://svn.wikimedia.org/svnroot/pywikipedia/trunk/pywikipedia/ pywikipedia

create a user-config.py, e.g. like this 

# -*- coding: utf-8  -*-
family = 'wikipedia'
mylang = 'de'
usernames['wikipedia']['de'] = u'Hannes Röst'

login by 

$ python login.py

run the following test script, trying to review the following revision change: http://de.wikipedia.org/w/index.php?oldid=81798619 

import wikipedia
predata = { 
            'action'        : 'review',
            'format'        : 'xml',
            'revid'         : '81798619',
            'token'         : wikipedia.getSite().getToken(getagain = True)

}   
address =wikipedia.getSite().family.api_address(wikipedia.getSite().lang)
response, data = wikipedia.getSite().postForm(address, predata=predata)
print response, data

my response: 

HTTPError: 500 Internal Server Error
WARNING: Could not open 'http://de.wikipedia.org/w/api.php'.
Maybe the server is down. Retrying in 1 minutes...


when entering bogus values for the token, this happens: 

$ python test.py 
<addinfourl at 148335948 whose fp = <socket._fileobject object at 0x8d3b86c>> <?xml version="1.0"?><api><error code="badtoken" info="Invalid token" /></api>
Thanks for the report! (and sorry for the problem) I replied in Bugzilla already, but please join us on IRC (#mediawiki) for real-time debugging of this problem if you get a chance. Thanks! -- RobLa-WMF (talk) 23:05, 24 November 2010 (UTC)
Great, that was quick. I had to reopen the bugreport even though the initial problem vanished, it is still not possible to review a revision over the API. The error is different and maybe even more difficult to solve. Maybe we need some testing framework for this to validate the functionality / or add it to the existing testing framework? Greetings --hroest 10:20, 25 November 2010 (UTC)


Revision History not readable

http://en.wikipedia.org/w/index.php?title=Mohammad_Amir&action=history is not readable for users like me that have turned on the light-on-dark gadget in preferences. ("Use a black background with green text on the Monobook skin"). Yellow text on a white background appears. The tweaks that display pending changes need to be improved so that they don't assume a light-colored background.--Elvey (talk) 06:07, 27 December 2010 (UTC)

edit summary

I would like to be able to see the reason / the edit summary a reviewer has accepted an edit for, that would be beneficial imo - as following up you would be able to assess the level of understanding the user had, as in - seems ok, and checked cite content is supported , these two comments are both perhaps good reasons to accept a non vandal type addition but for feedback the edit summary show a clear difference in assertion and verifiability. Off2riorob (talk) 15:17, 9 January 2011 (UTC)

Can pending revisions be applied to new articles?

Is it possible for admins to apply pending revisions to an article now that the trial is over, or is this forbidden? —Charles Edward (Talk | Contribs) 18:52, 11 May 2011 (UTC)

It's my understanding pending changes is currently being phased out. If there are issues with a particular article WP:RPP would be the way to go. Gerardw (talk) 19:36, 11 May 2011 (UTC)
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia:Pending_changes/Feedback/Archive_6&oldid=1144754305"