Wikipedia talk:Pending changes/Closure/Archive 1

to return to the discussion please click on the closure section of the link.

Archive 1

Approach to the work summary

Working summary section inappropriate. Discussions should be per Wikipedia:TPG#YES -- posts should be signed. See [1]Gerardw (talk) 15:43, 15 August 2010 (UTC)

Additionally, it now does not represent consensus as very POV statements are being added.Gerardw (talk) 16:18, 15 August 2010 (UTC)

Added disclaimer. Clearly not representing consensus, just a collection of ideas. Ocaasi (talk) 16:30, 15 August 2010 (UTC)

Wikipedia:Pending changes/Closure is not a talk page. Some bits are being treated as such, but I think the work summary should be seen more like a non-talk page. i.e. we try to come up with a work summary that we all agree with. That was how it was being done, but it seems to have changed recently. N.B. I have received this message on my talk page. If I am wrong about the work summary then I shouldn't have changed the comments added by others. However, I think a better approach would be if editors who disagree on the work summary take the discussion to the talk page, like with similar issues on other pages. Yaris678 (talk) 10:42, 16 August 2010 (UTC)

I added a few comments to the bullet points and they were changes a couple of times, I have now signed them.. If I am not to comment there or if my comments there can be altered by other users then I would rather remove them, please feel free to let me know if this is the case, thanks..It is so annoying when confusion is in charge from the get go. Off2riorob (talk) 10:53, 16 August 2010 (UTC)

Rob gave me the okay to refactor his comments into the discussion section and place a rephrasing in the summary. It'd be nice to keep that section free of signatures and strongly phrased opinions. Thanks! Ocaasi (talk) 14:20, 16 August 2010 (UTC)

This all makes sense to me. That's good work all round. The hat note at the top of the summary section is generally good. If there is a bit I am unsure about, it is where it says "Not representative of consensus." Obviously I don't expect everyone to agree with every point, but I would hope that there would be consensus that this is an neutral summary of the issues people had with pending changes. Yaris678 (talk) 00:57, 17 August 2010 (UTC)

"Not representative of consensus" is supposed to relax the criteria for inclusion without compromising on impartial phrasing. In other words, not everyone has to agree that item x or y was a flaw, or that item p or q shouldn't be a feature request. This shouldn't make it a free-for-all, though. I think the no-consensus part is to emphasize that this summary isn't binding. It won't be what the community officially uses. It's just a pretty good list. Do you have a different way to say it? Ocaasi (talk) 02:10, 17 August 2010 (UTC)

I have made a change to the hat note. Feel free to revert and discuss if you don't like the new words. Yaris678 (talk) 08:01, 17 August 2010 (UTC)

Looks fine to me. Below is a list of issue that people have variously mentioned that either need rephrasing or screening for duplicates/relevance. Anyone care to look through it? Ocaasi (talk) 08:20, 17 August 2010 (UTC)

I've merged the points below into the summary as best I can. I wasn't sure what "reviewing the reviewer" meant though. Does it refer to the current inability to see what reviews an individual reviewer has made? Yaris678 (talk) 15:03, 19 August 2010 (UTC)

Thanks for doing that. I crossed off the list, and then consolidated (or at least moved) them by sub-topic. "Reviewing the reviewer" had to do with the ability to oversee someone's reviewing decisions. Off2riorob had claimed that wasn't possible. Gerard pointed out that there is a specialpages log which records every reviewing action. The in-between is that it's 'less convenient to monitor edits', which I think I incorporated. Ocaasi 19:06, 19 August 2010 (UTC)

(indent)

Cool. I think the spliting things up made sense, but I didn't think the sub headings captured entirely what the sections were so I expanded a couple.

In terms of general changes...

1. Do you think it would be a good idea to have the work summary somewhere else. e.g. Wikipedia:Pending changes/Work summary. It difficult to track changes to it because there are so many other edits on the page, with people adding comments. The work summary can still be transcluded to the Closure page though.
2. Do we need a "way forward" section, where people suggest what we do now? Beyond what has already been suggested in the features section. This section would naturally contain a lot of mutually exclusive ideas, but it would be good to list those ideas. For example, someone might add "The Pending Changes feature should be restricted to a small number of low-traffic pages until we have decided how to deal with the above issues".

Yaris678 (talk) 07:13, 20 August 2010 (UTC)

• Good changes on section headings.
• I personally don't mind having the edits mixed together. Though transclusion would work, we might be overstepping our custodial duties by moving the page to where it would be inconvenient for editors to get to. I think we're close to finished in terms of adding new ideas to the list anyway; I can't conceive of too many more being added.
• I do think it's almost time to try and shift the discussion towards some more concrete proposals. With only 1 month to figure this out, 1 week almost passed, and almost 2 weeks needed for a fair poll, I think we have about 7-10 days to try and reach consensus about next options, or to try and create a few 'packages' which people might vote on. I think we could consider drafting different options ranging from:

1. close
2. close until fixed
3. continue without expansion
4. continue without expansion until fixed
5. continue with expansion

• I think these will be more useful for voting/discussion than a simple close/continue, which is so general that it can't really yield a meaningful response.
• Doing this would require:

1. figuring out a list of 'necessary' fixes, which most people would agree on before continuing/expanding
2. fixing as many of them as possible now
3. talking to the developers to see how long others will take
4. figuring out what the scope of an expansion would be (i.e. how many articles:5k, 10k; which kinds: low traffic, blp, etc.)
5. any necessary summary statistics to back up the views
6. a timeframe for the continued trial (2 months, 6 months, 1 year)
7. new metrics to evaluate a successful continued trial

• We also need to bring some other editors into this, particularly whoever set up this page and then left an opinionated horde to hash it out themselves (not that it's not going well).
• Not sure if that's what you had in mind, but let me know. Maybe we should open up a new section on talk for this. Ocaasi (talk) 11:36, 20 August 2010 (UTC)

(indent)

I wouldn't be so sure that the list is nearly done. I notice that User:Cenarium has tagged a couple of points as needing a citation or clarification.

I agree that the way forward bit should be discussed in another section.

Yaris678 (talk) 17:49, 20 August 2010 (UTC)

Potential issues for summary

• Watchlist issues:

Wikipedia:Pending_changes/Feedback#Thoughtless_implementation

Wikipedia:Pending_changes/Feedback#Poor_implementation.2C_chaotic.2C_extra_work_for_little_or_no_benefit

Wikipedia:Pending_changes/Feedback#Succinct_yet_descriptive_list_of_issues_and_suggestions #6

Ocaasi (talk) 17:29, 20 August 2010 (UTC)

What school year?

There are two comments about the school year and this is also mentioned in 'Open questions'. This is rather confusing since I can say the school year wasn't ending during the trial in New Zealand or Malaysia. And I'm pretty sure Australia, South Africa, Singapore... According to Academic term the school year in India begins in June, so this was more at the beginning rather then the end of the school year there. I presume it's referring to the US, perhaps also some other northern hemisphere temperate countries but as it stands, it's an overly broad brush claim that clearly isn't true in a number of instances. Nil Einne (talk) 19:15, 17 August 2010 (UTC)

June to August sees a very noticeable drop in vandalism, every year, and then from early September onwards it's noticeably all downhill. I believe this is mainly due to the holiday seasons in the USA, Canada, and most of Europe, which is where most school vandalism comes from (most vandalism is from schools). Australia, New Zealand and South Africa also have significant holiday during this time. -- zzuuzz ^(talk) 19:29, 17 August 2010 (UTC)

Zzuuzz hit it on the head. It's pretty much common knowledge that when the (US/Canada) school year ends, the rate of vandalism drops dramatically. Since this state only lasts for three months, however, it's not indicative of everyday vandalism (which would normally be higher), which I feel invalidates the trial as biased. —Jeremy ^{(v^_^v Carl Johnson)} 19:43, 17 August 2010 (UTC)

I was semi aware there was generally a low in vandalism, but I think people are perhaps missing my point. IMHO it would be far better to talk about what actual matters (drop in vandalism levels) during this time then make a clearly untrue broadbrush claim about the end of school years. I've modified the claim here [2] accordingly. If someone want's to add back mention of school year, please clarify what school year you're referring to, e.g. US/Canada, temperate northern hemisphere countries. Don't get me wrong, I'm not saying the reason isn't because of the end of the school years in said countries, but the reasons don't really matter in this discussion. And if you want to discuss reasons, you have to make sure you are clear about what you're talking about, and not make misleading claims like imply it's the end of the school year everywhere during this time, which it clearly isn't. (In case you haven't realised I strongly dislike clear untruths or overly broad-brushed claims of this sort.)

Also, while I don't know much about the Australian and South African school years, I would argue it's misleading to claim there is a significant holiday for New Zealand schools during this time. While I admit I'm not that familiar with the normal practice of schools here, AFAIK, supported by our article, the only holiday for schools in NZ at the time is the normal inter-term break. But the length is only 2 weeks which is the same for all inter-term breaks. So realisticly, you're going to catch one of these breaks for any 3 month period. As it's the mid school year, it may be that schools have more significant exams before the break, as well as more assignments etc during this break then for the other inter-term breaks which could have an effect but in terms of the break itself, I don't know if you can say it's any more significant then the rest. (This compares to Malaysia where are are two semesters and the inter semester break is generally longer then the intra-semester break.) New Zealand universities which usually operate on a two semester basis do have their mid-semester break during the June-July period (exams before), but this isn't what people usually mean when they refer to schools in New Zealand.

Nil Einne (talk) 20:01, 17 August 2010 (UTC)

Two weeks out of school is significant enough to be noticed :) We get the same effect around Easter time. I think the causal link between school holidays and the drop in vandalism is beyond question. It can be safely blamed on North America and Europe, which is where most school edits come from - usually during biology and history lessons. I'm not sure it invalidates the trial, or even had much effect, though it certainly skews any interpretations about the effect of PC in reducing vandalism or affecting the number of edits. -- zzuuzz ^(talk) 20:15, 17 August 2010 (UTC)

But realisticly any trial 3 month trial is going to catch one of those two weeks. And in fact a 6 month trial is going to catch 2 of those 2 weeks. A 2 month trial has a good chance of catching such a 2 week period although obviously isn't guaranteed. In other words, I'm not really sure if there's any relevance of the school holidays in NZ other then the fact we have them and they're likely to be present during most trials.

Note as I said above, I'm not denying or questioning the link. I'm questioning making broad brushed claims that aren't true (the school year wasn't ending in all countries during the trial).

To put it a different way, there is a world outside the US and the temperate northern hemisphere and it's IMHO unwise and even offensive to act like they don't exist by making claims that only apply to one area without making it clear you're only referring to one area. Particularly in some sort of general summary which should be neutrally worded and unambigious. See [3] which is the 'Open questions' section that primarily started this discussion and had the specific but clearly overly broadbrushed claim "Did the timing of the trial--as the school year ended--skew results?".

As I've said, the best thing here is either don't talk about the school year at all (IMHO the better solution) or if you do, specify who's school year you're referring to.

Nil Einne (talk) 20:24, 17 August 2010 (UTC)

Indeed. No complaints about some clarification of this phenomenon. -- zzuuzz ^(talk) 20:31, 17 August 2010 (UTC)

Clarification is indeed fine, and sought after. Don't think this is some blanket violation of WP:WORLDVIEW. Well, maybe it is, but that's just because the fine youngsters of the northern-western hemisphere can be real twits when left unattended in the library. Take it as a compliment that your brethren aren't the source for such brilliancy's such as [Wikipedia is a biiiiiig penis because John and Tony are gay gay gay]. So don't feel too left out. Ocaasi (talk) 21:37, 17 August 2010 (UTC)

Also, per two of your specific criticisms: No one suggested it was the developer's secret intention to bias the trial by starting it in the summer. Correlation & causality--stuff can happen with anyone planning it to happen, and especially without anyone planning it to happen deceptively. Further, it's possible the developers wanted the study in the summer, because they knew that the effects of vandalism would not overburden the system in its first trial. Two, the majority of the English-speaking world is out of school between July and August. A global worldview is important for articles that have global relevance. This study is primarily about the performance of Pending Changes on English Wikipedia, which though not owned or centered in any location, is inherently biased by where more people speak English. It's no offense, just a numerical fact that the majority of students who speak English are out of school during these months, and the places where people are in school during these months are disproportionately less likely to use English wikipedia. Ocaasi (talk) 00:03, 18 August 2010 (UTC)

Poll and advertising

What do you think of starting the poll on 31 August 2010 (00:00 UTC) and ending it on 14 September 2010 (00:00 UTC) ? Also, how do we measure consensus ? As for advertising, I think we should use a sitenotice already in the discussion phase, then for the duration of the poll. Cenarium (talk) 14:50, 18 August 2010 (UTC)

• We could ask the Crats to close it through a chat using similar consensus as per closing of an AFD. Or is it to be more of a headcount? We could also go on a RFA close again by the crats through a crat chat with the successful percentage at the minimum end for RFA success? Off2riorob (talk) 14:58, 18 August 2010 (UTC)
• I agree with the idea of approaching the bureaucrats to close the !vote, as they are expected to be intimately familiar with the concept of consensus (regularly closing RFA's). As for the advertising, the currently developing sitenotice should suffice. Tyrol5 [Talk] 19:49, 18 August 2010 (UTC)
• I think the vote should give three options: End trial. Continue without expansion. Continue with expansion. In any scenario, possible changes/improvements would still be under consideration. I think it's important to distinguish between a 'longer' trial and a 'broader' trial. Broader would mean > 2000 pages, up to 5k or 10k, although possibly limiting high traffic pages to a few hundred, or focusing the expansion on BLP pages. I guess I think we need to clarify as much as possible first, then provide a few options or packages. I don't think a simple stop or continue will suffice.

Modified Proposal: Vote to close. Or vote to continue with conditional expansion. Conditional expansion allows P.C. to continue 'as is' with a cap of 2000 articles for a period of 6 months. Expansion beyond 2000 articles to a max of 10,000 would be conducted only after the following issues were remedied: 1. accept/unaccept scheme overhauled 2. good-edit/vandalism confusion settled 3. guidelines for which types of pages should use PC clearly established in the documentation 4. Useful statistics are identified which correlate with 'successful' results in a variety of specific circumstances 5. etc.,

Speaking of which: Will some of the identified issues be fixed in the next month? If not, then the vote to close is almost pointless, because it won't be addressing the system as it will be implemented. In other words, in addition to "forming consensus", should we be working on addressing all of the 'cons' and 'feature request' in advance of the poll?

Ocaasi10:45, 19 August 2010 (UTC)

I don't understand some of your points...1 to 5..? It needs to be kept imo as simple as possible, the understanding is there that we will tweak and improve some of the issues raised as the tool goes forward if there is community support for the tool to continue. Off2riorob (talk) 20:21, 19 August 2010 (UTC)

The points are just improvements that would have to happen before the continued trial could expand to any more articles. The list is just a guess at what consensus for the most needed features/fixes would be. Ocaasi (talk) 03:07, 20 August 2010 (UTC)

Ah, ok, we do need imo to find out from someone if the tool is continued in use what is the scope of use for the tool going forward, we should ask on of the main involvees, I forget who they are? Presently I am left asking, support, what is it that I am actually going to be supporting? Off2riorob (talk) 03:13, 20 August 2010 (UTC)

Yes, that is very much the issue. I think at the very least, it would just be a vote to support continuing what has been happening the last two months, no more, no less. I think that's missing an opportunity to get a better idea of what changes should happen and what people want to do with pc. In the first section on this page, Yaris and I have been brainstorming about what might happen next. I don't know who to contact about this among the main involvees. Maybe RobLa Ocaasi (talk) 11:42, 20 August 2010 (UTC)

I have put a couple of feelers out to get a bit more feedback, I like your simple point. A vote comment to continue what has been happening with no major expansion, not with a limit as such but no large organized topic focused expansion. A comment I did get was that presently the capacity is not possible for a community roll out as it would be unworkable anyways. Off2riorob (talk) 12:01, 20 August 2010 (UTC)

Maybe a useful approach would be to give the continue option a conditional possibility for small expansion. Something like: "Vote to close, or to continue as is with the possibility of expanding to up to 5k articles conditional upon the following X issues being fixed." Or something like that. Is that more confusing? Ocaasi (talk) 12:37, 20 August 2010 (UTC)

If you mean 50–250k articles instead of 5k, sure. There are quite a number of unwatched BLPs this was meant to take care of, 5k would barely put a dent in our lack of protection of them. NW (Talk) 12:53, 20 August 2010 (UTC)

Few things: I don't think there's community support to expand PC as of yet, at least not until its interface and documentation kinks are worked out. I'm not sure that it's even technically possible to have it running on hundreds of thousands of articles. I have no personal issue with it being expanded to BLPs. If there was a minimal expansion in the next few months, I would assume BLPs would be a target. This is all still just a test of the interface itself and whether or not the community likes it. That said, BLPs seem like a natural use for the system. Are you suggesting that this trial hasn't served it's purported function, that it was originally designed to fix the BLP issue and not to compare as an alternative to semi-protection? Ocaasi (talk) 13:03, 20 August 2010 (UTC)

Yes, that is the case. It appears you are a new user who might not be familiar with the history behind Flagged Revisions, so here's a quick summary: BLPs had been a major issue for the project since the Seigenthaler incident. Flagged Revisions had been promised for ages. Back around January 2009, we finally got around to having a massive poll on the matter. The option that received the highest plurality of support was "use flagged revisions on all BLPs", docking about 60%. The "GodKing" declared that to be sufficient consensus and asked for the Foundation to enable flagged revisions. After a number of subsequent polls, which established a new method of flagging called "flagged protection", that could be used as an alternative to semi-protection but was intended to be used far more widely. (see Wikipedia:Targeted flagging) However, the developers were very slow on the matter and it took about a full year for us to finally get a working version of flagged revisions, retitled pending changes. However, during 2009, a number of news media stories were run about how "Wikipedia intends to institute a reviewing system for many articles, ditching its instant-open-editing model." Jimbo later began giving statements along the lines of "No, that's not what we are doing at all; we are actually opening up articles that had been locked for a long time." That's spin, nothing more. I don't think that the community's intent for flagged revisions ever changed. NW (Talk) 13:19, 20 August 2010 (UTC)

I was familiar with the various incidents but not their interconnection. Thanks for the summary. So a few questions then: Was the Jan '09 flagged revisions 'consensus' for pending changes or merely patrolled revisions? Even though it took a while, isn't it about the same now that the system is being tested e.g. the problems still needed to be worked. Also, how targeted is "BLP"? Some have commented that far more articles contain information relating to living individuals than those with a name in the title.

Let me just double check the details. Flagged protection=Pending changes; Patrolled Revisions = Pending changes - the pending part; Targeted Flagging=Pending Changes but only on low traffic/watchlist BLPs? Is there any technological difference between TF and PC, or is it just a matter of scope? Ocaasi (talk) 13:49, 20 August 2010 (UTC)

Flagged revisions was a broader form of Flagged protection=Pending changes, which would have automatically applied to articles in Category:Living people. Targeted flagging was a proposed method of using Pending changes. Patrolled revisions was related to all of these, but a separate feature that still is not available on enwiki.

The January 2009 discussion was consensus (as declared by Jimbo) for flagged revisions (later renamed to pending changes) and patrolled revisions. BLPs in context of this discussion was understood to be limited only to Category:Living people, but it could have been expanded to other articles outside of that on a case-by-case basis. NW (Talk) 14:06, 20 August 2010 (UTC)

Well maybe there'd be support to continual the trial and expand to 10k low-watched BLPs. Is that even worth discussing? Also, what's your preference, targeted flagging or targeted patrolling? Ocaasi (talk) 14:34, 20 August 2010 (UTC)

The latter isn't even an option right now; patrolled revisions was never enabled, and wouldn't really help matters by itself even if it were. Patrolled revisions was always meant as a supplement to Flagged revisions/Pending changes. Anyway, if you move your 10k figure about an order of magnitude to bring it more in line with Scott MacDonald's proposal, then that would be my preference. NW (Talk) 14:44, 20 August 2010 (UTC)

The idea with patrolled revisions was principally to better monitor underwatched BLPs (I say underwatched because unwatched means essentially nothing, plenty of accounts are inactive and people don't watch all day anyway). As it's totally passive it would be easily approved by the community, but an active preventive implementation on tens of thousands of pages would be hard to pass, and as Scott points out on the main page largely incompatible with this implementation. Cenarium (talk) 17:42, 20 August 2010 (UTC)

• - As regards closure methods - I posted a note at the crats noticeboard about closure and got no interest at all. I suppose it is not actually their specified job. Perhaps we can just ask an experienced administrator. I don't think it is a big issue. Off2riorob (talk) 23:24, 20 August 2010 (UTC)