Wikipedia talk:What Wikipedia is not/Archive 54

Idea for a compromise

The most contentious point of divergence here is whether these events should have their own articles, not whether they should be covered at all, so I propose we collate them as sections in a list article of some sort. The sections that end up too long after the buzz has died down can be spun out in their own articles. This way we will have less articles prematurely created and less incentive for users to pad the articles with condolence quotes and such to make them seem more notable. For example, see 2017 Atlantic hurricane season. Hurricane Katia was discussed in news even here, 9000 km away, and it doesn't have its own article. And it doesn't need one. Daß Wölf 00:13, 29 October 2017 (UTC)

• It is a reasonable and fair compromise that I would support Daß Wölf but unfortunately it will never pass; there simply is no compromise that can be reached with the other side of the discussion. They need the trivial biographical information on the alleged terrorist, the "background" info that simply lists other notable (but unrelated) incidents or states a country is on "edge", and an impact section that is comprised of sources that mention the incident maybe in a sentence or as a part of a "trend". When you analyze these articles, 80% of it is window-dressing and the rest is actually about the incident. It all could be easily and orderly summarized, as you recommended, in a list. Many editors have proposed similar solutions during AFD and merge discussions but it is opposed by a sizable enough group who claim such a move would result in the loss of "valuable information". Hence why we have had too many "no consensus" outcomes on this topic.TheGracefulSlick (talk) 06:02, 29 October 2017 (UTC)
• The length of material in some of these individual events is not appropriate for a list article. Collapsing an article that is a few pages in length to a 2 liners in a list is really equivalent to a deletion. These events generally meet WP:GNG and WP:EVENTCRIT - with copious coverage. What isn't clear on some of them (ignoring bad faith nominations of not recent events) is whether they will have a WP:LASTING effect - this isn't clear since they are recent and evaluating LASTING on a 1 month old event is difficult and requires CRYTALBALLING -- hence WP:RAPID is relevant.Icewhiz (talk) 06:59, 29 October 2017 (UTC)

What about Murder of Reuven Shmerling? Prima facie, it looks like a standard short article, but when you remove the repetitions and a few minor details, you're left with this:

I've removed the wife's name, the name of the news agency that reported the arrests, and the list of people who expressed condolences. Even this is already 30% shorter than the Hurricane Katia entry. How is this material not best suited for a list article? Daß Wölf 17:50, 29 October 2017 (UTC)

• Agree with Icewhiz, and I would also observe that this isn't a compromise. Coretheapple (talk) 13:05, 29 October 2017 (UTC)
• Agree with Icewhiz and Coretheapple.E.M.Gregory (talk) 13:22, 29 October 2017 (UTC)
• This is why it sucks we are not using Wikinews. Detailed information written Wikistyle could go there, the event listed in these tables with the intra-wiki link to the Wikinews article and other sources. Unfortunately, we simply cannot get editors to use what WMF already provided. --MASEM (t) 13:37, 29 October 2017 (UTC)

Wikinews is infested by policies. For example, Wikinews articles are supposed to be "finished" at a certain point and locked against further editing, which is obviously contrary to the encyclopedia article model that assumes, most likely correctly, that kibbles and bits about the JFK assassination will be coming out millennia from now. Also, Wikinews demands much too much from the first editor to get an article over a threshold. The more policies a project has, the more likely your work is going to be thrown away; in other words, the more policies a project has, the more it sucks. Since policies are the one thing on Wikipedia that deletionists never line up to take out, no matter how badly written and counterproductive they are (WP:NOT being a case in point) the only way to delay the day when both Wikipedia and Wikinews suck too bad to use at all is to try to delay the lobbying against news and half a dozen other categories of coverage that goes on nonstop here. Wnt (talk) 15:10, 29 October 2017 (UTC)

• One recent AfD nomination, is at least two-thirds speculation, mostly by journalists, so I dispute Icewhiz's point as to the volume of sound factual info - in many instances - it is little more than a para. It is ironical to me that 'film articles' will often list as few as 4 or 5 reviews, whilst this 'terrorism' incident is deemed to need about a dozen journalists' speculations. In this instance, most of the speculations are valueless, since relevant expertise is not demonstrated and all of the speculations tend to be in a particular direction (is it ISIS? We don't know but it could be). We don't write-up comments about Mozart Sonatas written by minor journalists who can't read music but, in this topic area, the equivalent is precisely what we are often doing. I always had thought that 'LASTING' impact was meant to be demonstrated before keeping an article. Time and time again, it is being argued in this topic area that the ABSENCE of lasting impact has to be demonstrated before deletion. The compromise suggestion made above is what SHOULD happen, it won't get support though - these minor events are already anyhow being added to umpteen lists in addition to the main article. More lists sometimes than there are editors watching this topic area. Pincrete (talk) 14:34, 29 October 2017 (UTC)

  The 2014 Dijon incident is discussed in journal articles as late as 2017 as well as being given as an example of a vehicle ramming attack in subsequent attacks in 2017 - so there is a lasting impact. Whether it was or was not terrorism is immaterial to notability - coverage is.Icewhiz (talk) 14:42, 29 October 2017 (UTC)

  But importantly, there's a different between just be name-dropped, and actual continued coverage. That's what WP:N/WP:NEVENT requires. --MASEM (t) 14:52, 29 October 2017 (UTC)

• Yes, there is a difference. 2014 Dijon attack has analysis by academics including David C. Rapoport, and in the policy journal Combating Terrorism Center#CTC Sentinel. And a number of reported articles that analyze this attack and its impact in the context of similar attacks, such articles have run in such major, reliable media as the Globe and Mail: "History of lone-wolf vehicle attacks suggests risk of emulation is very real" and Financial Times: France emerges as main focus for terrorists, New York Times, In the Age of ISIS, Who’s a Terrorist, and Who’s Simply Deranged?, National Public Radio: Vehicle Attacks Like Berlin's Are Nothing New, And Are Likely To Continue, and were published years after the attack happened. Deletion, and deletion by reduction to a brief mention in a list, would be a major loss of useful information.E.M.Gregory (talk) 15:10, 29 October 2017 (UTC)

• Like I said above, these editors would never let such a fair and reasonable compromise pass. It benefits the encyclopedia but that is not their number one concern. It is the loss of "useful information" (dang I was one word off): the trivia, the speculation taken from one POV, and a "trend" or "context" pieced together within multiple name-drops. I am almost fearful to nominate articles for deletion or propose a merge; by the time we reach a "no consensus", the article is left in a worse condition by editors who hastily throw in news sources to claim there is further coverage. So sadly, the only way we can achieve a positive change is by more editors like Masem participating at AFD to help end the deadlock we experience or admins must exert more confidence in ignoring moot arguments.TheGracefulSlick (talk) 16:29, 29 October 2017 (UTC)
• Can't check the FT, but the Globe & Mail, NYtimes and NPR sources are namedrops, and provide no significant coverage of the event beyond noting it happened (eg if we were evaluating notability of a subject, they wouldn't qualify as they are not secondary to the event). --MASEM (t) 17:01, 29 October 2017 (UTC)

• No, they are not mere "name drops," they are analyses of the Dijon attacks, variously as one of a group of attacks where the mental status of the attacker is in quesiton, one of a a gorups of car ramming attacks, one of a group of apparent copycat attacks. Several more analyses of this attack along similar lines are in the Dijon article.E.M.Gregory (talk) 18:19, 29 October 2017 (UTC)

• The mention of the Dijon attack among other attacks as a type of pattern or specific type of attack does not make the Dijon attack notable, but lends to the pattern/type of attack being notable. (It's not transforming any elements of the Dijon attack into a novel view of that attack, only merely noting it falls within the pattern). Its lends more appropriate to include it in a list or a full article on these types of attacks rather than an article by itself. --MASEM (t) 18:32, 29 October 2017 (UTC)

• Besides being more than just namedrops, continual references to this incident are exactly what generates a reader's interest in looking the referenced incident on Wikipedia. The BBC just mentioneed Dijon in 2014... Wwhat was that?Icewhiz (talk) 18:28, 29 October 2017 (UTC)

• There is zero mandate that we be there to provide information on a news topic just because a reader is searching for it. We are not a newspaper, and readers coming here for news content are not coming here for the right purpose. But we can still use things like redirects to help provide search hits to the right list article or the like where the event may be listed as part of a larger notable topic. --MASEM (t) 18:32, 29 October 2017 (UTC)

• Encyclopedias exist to look up topics. If an event is continually mentioned, it is an event likely to be looked up. This is evident in the page views of the Dijon 2014 incident that peak with subsequent car attacks.Icewhiz (talk) 18:37, 29 October 2017 (UTC)

• Being useful to users is what we are here for. Of course hits validate the usefulness of a page. Speaking as a long-time user who started editing after years of relying on WP to know stuff I was curious about, I'd be ticked if I wanted to look up a TERRORIST ATTACK and all Wikipedia had was a lousy mention on a list.E.M.Gregory (talk) 20:39, 29 October 2017 (UTC)

• To understand why this is a dysfunctional proposal, consider that the indictments have just been issued in teh Murder of Reuven Shmerling. According to the indictment, the two murderers plotted and carried out the murder to take "revenge" for the death of a specific Palestinian terrorist who died in a failed stabbing attack in the summer of 2015. It wold be useful to be able to link the Shmerling murder article to that attack because I imagine that I am not the only person who would like to know something about the terrorist these two terrorists wanted to avenge. Terrorist attacks often come back into the news cycle, for all sorts of reasons. And out users expect us to have articles on them (evdenced, as Icewhiz points out above, the the page hist the 2014 Dijon attack gets every time a terrorist rams a car into a cluster of pedestrians.E.M.Gregory (talk) 21:18, 29 October 2017 (UTC)
• Support, since it's what I already suggested. I do not agree with Icewhiz that this isn't viable because a whole class of editors will never compromise. I think that verges on WP:ASPERSIONS. In point of fact, we have thousands and thousands of list articles, including of events, specifically because editors compromise and realize, collectively, that not everything worthy to be included in some form in the encyclopedia needs its own article. There is no reason this subject in particular will be magically different and somehow immune from the same reasoning.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  04:01, 30 October 2017 (UTC)

• SMcCandlish I already expressed my support but I must ask, how would this be enforced? Do we open talk page discussions at the articles and decide if it is appropriate for a list? I have tried that on a few occasions and I am accused of "stealth deletion" tactics; it ends in "no consensus" like the AFDs. I guess my main concern is: how would we prevent this proposal from being more of the same? The same editors are still going to oppose any move whatsoever to convert incidents into lists, regardless of what is decided here.TheGracefulSlick (talk) 05:45, 30 October 2017 (UTC)

  WP:PM and, failing that, WP:AFD, probably. I think that's how we've dealt with all "forced" mergers, where the sensible community consensus is for listification, but a camp of topically-"invested" editors is resistant. Example: merging away the formerly large mass of independent articles on fictional characters not independently notable; and numerous non-notable, local, K–12 school articles into articles on school districts.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  07:41, 30 October 2017 (UTC)

• Daß proposes summarizing and including certain terror attacks on lists. The problem is that unfortunately this can't be done. Take, for instance, the 2017 Buckingham Palace attack article. An attempt was made to include a very brief summary of it on a list, the List of terrorist incidents in London, but two editors persistently deleted ^{[1][2][3][4][5]} the Buckingham Palace attack entry because «only suspected to be terrorism related, needs a court verdict to confirm it was». Observe the 3RR rule was not even followed. So, whenever attempts are made to summarize, other editors are on standby to simply memory-hole the whole entry. Unfortunately, experience shows that the proposal would result in the immediate and wholesale destruction of a very large quantity of valuable encyclopaedic content that has, in every case in the last few months, survived, through independent review, a persistent AfD campaign. XavierItzm (talk) 10:59, 30 October 2017 (UTC)

• • What User:XavierItzm writes is very true. The most glaring example is Palestinian stone-throwing, where a rule exists that incidents where a stone thrown at a car killed the occupants or maimed them for life cannot be added to the page unless there is an article about them. The rule effectively forces the creation of such articles.E.M.Gregory (talk) 11:24, 30 October 2017 (UTC)

    Simple and obvious solutions: 1) change the list's inclusion criteria, by RfC if necessary, so it's not over-inclusive or over-exclusive. 2) Deal with revert-warriors at noticeboards.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  12:32, 30 October 2017 (UTC)

    Yes, but in some cases the revert warriors are correct (not always) in rejecting a notable incident in a list. For instance if the list is labelled terror, and there is doubt whether it was terror, plain murder, insanity, or an accident - multiple lists might reject such an item - even though the item itself might, due to actual coverage (possibly, depending on time elapses since the event, even LASTING), meet GNG and NEVENT. Keeping a notable event, meeting policy guidelines, standalone allows the coverage on wiki to grow without constant reversion and in the case of non-alert contra-reversion warriors loss of information.Icewhiz (talk) 12:41, 30 October 2017 (UTC)

So we have to have a separate article (inevitably conjecturing that the incident was the worst terrorist event since the year dot), because it actually doesn't meet objective criteria for inclusion in a list of terrorist events? Yes that makes enormous sense - if one ignores every RS, BLP, NPOV policy that WP has. We are here to record transient speculations are we? Pincrete (talk) 12:52, 30 October 2017 (UTC)

Hypothetically, if an airplane flies into a building and no one takes immediate responsibility, the event would be notable but not necessarily terrorist in nature (it might be speculated to be such, and speculation might evolve as it is investigated). And lest you say notable per WP:AIRCRASH the same would be true if a truck ran over a large bunch of people and we don't know yet why it happened. The criteria for notability of events is independent of whether they are or are not terrorist. Inclusion in a particular list - is usually dependent per a closed set of criteria - which in the case of a non-closed case could evolve and leading the event to bounce between list to list (e.g. we still do not know why Malaysia Airlines Flight 370 happened - we might know. Or we might think we know. etc.).Icewhiz (talk) 13:02, 30 October 2017 (UTC)

The notability for events is about enduring coverage, which cannot happen in the first 24-48 hrs of an event occurring. Granted, there are a few types of events that we have articles on immediately because we know that due to the nature of the event, there will be analysis of it : earthquakes, hurricanes, commercial aircraft disasters (regardless if terrorism was involved), etc. Most other events don't have this assured post-event coverage, including those that may be confirmed as terrorism, and in such case, per NEVENT, we should wait to learn what the event's context and larger importance is before inclusion on WP. But editors instead tend to rush in, against NOT#NEWS and other policies, to start articles immediately as something bad happens (arguing international coverage == notable, which is nowhere in our guidelines), and this ends us with dozens of useless reaction statements that pollute such articles. If editors could just wait 2-3 days before writing on events where enduring coverage is not clear, that would be a much better situation, as then it would be easier to slot those that aren't notable into such tables and lists. --MASEM (t) 13:29, 30 October 2017 (UTC)

When should Grenfell Tower fire have been created? In the news item on the main page. 2017 Finsbury Park attack? An event not notable for its body count (and to begin with, the fatality was possibly a heart attack victim (for a couple of weeks)) - but notable for its target (Muslims) - and yes, again on the main page. Many events are difficult to slot in 2-3 days. Some are even hard to slot at a month. There is little harm if a non-notable incident (due to lack of LASTING) is "live" on-wiki for a few months - nominating these close to the event is really a waste of everyones' time (and leads to crystallballing and assertions from all sides). Things would go much smoother at AfD if borderline events (not clear PROMO, SPAM, or really ROUTINE stuff) got nominated a year+ after the event.Icewhiz (talk) 13:45, 30 October 2017 (UTC)

The harm is that you get new editors, not aware of the intricacies of policy, creating new articles on any old event - I've seen articles about simple traffic accidents created as new items because of this. We also shouldn't be driven by trying to feature content on the main page through ITN through this; ITN is not a news ticker, but editors seem intent in pushing it in that direction. I've generally found that articles on events, massively edited in those first few days and written in a manner that doesn't fit an encyclopedia, rarely get the love and care needed in the long term to bring it to a quality article (keeping DEADLINE in mind).

In 'my' perfect world the solution is that news articles would live at Wikinews first on their creation, and when there was enough assurance they were notable, the appropriate content could be brought over into enwiki. That's not clearly happening nor expect us to every have consensus to do that, but we need better checks and balances on en.wiki's end so that news articles are reviewed within a fair and reasonable time after an event, recognizing that unlike most other topics which aren't necessarily fixed in time, we generally can judge the enduring coverage of an event easily by Google searches or the like. Give articles on news events 2-4 weeks from the event to show if the topic's enduring, and if its not, figure where to slot it in larger tables or list, or otherwise delete. --MASEM (t) 13:56, 30 October 2017 (UTC)

Wikinews is dead. It also isn't the best place for evolving coverage of an event (because news items are supposed to be closed). There are also advantages to updating events in real-time (something I've done, but on really clearly notable events). How about this one - a conditional future AfD - tag the event, and have the AfD itself run 9 months after the event (not article creating) start date (if placer doesn't remove the tag). For any event where RAPID is going to be called, this would typically resolve RAPID issues.Icewhiz (talk) 14:22, 30 October 2017 (UTC)

9 months is far too long. Most of the problematic articles in events can be determined within a few weeks (in the sense one can do a google news search, not necessarily by quality of the article). This "gate" idea should review what sources are in such offsite searches, not what is presently within the article, so we're not judging quality but potential based on if the topic is truly enduring. I do agree there are some events that easily are notable within a few hours of them happening and can bypass this.

However, this still isn't sufficient if we have editors arguing in the case of the examples give (the Dijon attack, etc. ) that the type of coverage and namedropping some years later equates to "notable". The same arguments would happen at those "gates" as they do not with AFD. I'm all for allowing new event articles to have at least some chance, but we have to recognize that as an encyclopedia and not a newspaper , we only want to keep event articles that have long-term influence, not just because they happened. As I mentioned before, we have to imagine that if we were starting these event articles 40 years later, all other sourcing being the same, would they really be notable? Most of the time, that answer is really no. That's the line between us being an encyclopedia and us just happening to cover widely-reported-on events. --MASEM (t) 14:35, 30 October 2017 (UTC)

Dijon 2014 is a clear case of disagreement between editors which would not be resolved by a gate. However many AfDs are run against newish events in which case this would be pertinent. An AfD on a 2-3 day old (or even 2 weeks) event which isn't clearly non-notable is usually pretty pointless and devolves into a philosophical crystal-balling discussion on why it will or will not be notable and what degree of continuing coverage will or will not emerge.Icewhiz (talk) 14:45, 30 October 2017 (UTC)

I agree its a point of contention and is a key example of the issue with how WP:N and WP:NEVENTS should be evaluated. The event is long and truly over, so we have 20/20 hindsight to evaluate the "enduring coverage" of it, but that leads to the argument of what enduring coverage actually is (see arguments above). I think, judging by all commenters so far, that if we didn't have the sources that came a few years later, the event likely would be fairly seen as non-notable per NEVENT. The argument now remains if those additional sources really are true enduring coverage or not. Until that type of question is resolved, there's little else to argue for this "gate" check.

What much of this comes down to is that editors that seem very interested in news events want special allowances that other topic areas do not have. Given that we are not a newspaper (and several recent polls show there's no consensus to change from how it is currently worded), we want news event articles to be approached as an encyclopedic manner, which demands more a long-term view written with good hindsight. Most news articles are being written based on crystal-balling that there will be enduring coverage, which we don't allow for any other topic. I understand the need to write these articles early in the event, so I can appreciate the desire to have this crystal balling at the start, but that needs to be checked in a reasonably short period after the event has tailed off, if we are going to be allowing these to be created as such. --MASEM (t) 15:53, 30 October 2017 (UTC)

Very interested in terrorist attacks perpetrated by brown people, to be precise. They don't care about regular stabbings or shootings. It is a very specific agenda here. It is not a specific dispute regarding NOTNEWS. AusLondonder (talk) 13:53, 31 October 2017 (UTC)

A different idea for a compromise: Provisional keep

A different idea for a compromise when an article on a crime, possible terrorist attack or other new event is started very RAPIDly, is the model I just found at 2014 Endeavour Hills stabbings. The article was started on 23 September 2014, the day of this stabbing attack on an Australian police station. Wikipedia:Articles for deletion/2014 Endeavour Hills stabbings was started on 26 September, it was brief, collegial, discussion and amicably closed after 7 days. What interests me in particular is the comment by User:Lankiveil :"Provisional Keep, I concur with the nominator that this will likely be forgotten in about three months once the news cycle has moved on, we'll probably be in a better position to gauge if this was just a flash in the pan WP:NOTNEWS case then." The response by article creator '''tAD''': "I have no problem with that. If nothing comes from the inevitable research by authorities into Haider, this will be little more than a man killed by authorities after attacking them" The AfD was closed as keep.
My suggestion is that we encourage the use of Provisional keep as an iVote in breaking news situations. In something like the Wikipedia:Articles for deletion/2017 Clovis library shooting, I can see taht I would have been totally comfortable writing something like, Provisional keep: Revisit in six months when we can assess WP:INDEPTH and WP:PERSISTENCE. E.M.Gregory (talk) 19:14, 30 October 2017 (UTC)

• One thing I would like to see is avoiding using AFD on breaking news items until after a fixed period had passed (but allowing for CSD for bogus or other content-violations like BLP to happen as necessary), but with the understanding that all articles should be reviewed by a human editor, and if deemed non-notable, then send to AFD for a larger review/judgement call which may include merging into existing lists. Certain news items where the topic is clearly notable before the end of that period can be kicked out from the queue. In other words, we recognize editors are going to write news articles as they happen (its nearly impossible to stop this) but have a way to check there work after enough time, and only trigger the AFD after that time there are still questions. I would also agree that if the item is very much slow moving that takes months to develop, then allowing a "Provisional keep" to put that article back in this queue would be reasonable. The only thing is that I think the time period should be shorter, on the order of 2-3 months. --MASEM (t) 19:29, 30 October 2017 (UTC)
• Or editors could wait that same amount of time instead of churning out articles when they have no idea if the subject will have long-term significance. We could also userfy/draftify the articles until they have/if they have demonstrated long-term significance.TheGracefulSlick (talk) 19:53, 30 October 2017 (UTC)
• I like the idea, but in your example it was forgotten about. Perhaps we need to create a category for such AfDs? If its a provisional keep in September 2017 its tagged with "Category:AfD Provisional keeps until February 2018" or something of the sort. While I understand Masem's point on the shorter period of time, it gets a little hairy with gag orders. I also agree with TheGracefulSlick's point about creating a draft for these, however you run into the issue of a new user creating the article, then what? A lot of these solution on stopping creating them are contingent upon them not being created anyway, which seems to be our biggest issue. - Galatz^Talk 00:39, 31 October 2017 (UTC)

"A lot of these solution(s) on stopping creating them are contingent upon them not being created anyway, which seems to be our biggest issue" - second that! This proposal seems to turn policy on its head, we provisionally keep something until it demonstrably lacks long-term, in depth coverage or significance? Pincrete (talk) 08:00, 31 October 2017 (UTC)

I'm not sure this is actually a change in policy (how could you apply NOTNEWS for content that you don't know whether it will have long term impact?), and it definitely is a good idea. I see it as a direct application of WP:IMPERFECT. The idea that we *should not* have an article until we are sure of its enduring notability, seems backwards with the way we have always worked; and it is not in the spirit of the WP:EDITING policy. Our modus operandi has always been "write it dirty, and fix it fast".

The desire to wait till the dust is settled so we can make a proper scholastic work is tempting, but it dilapidates our largest strength: focusing the outbursts of attention from many editors, and direct them into shaping a first rough draft, to be polished later and made compliant with guidelines. This is how our first million articles were written, and it made us lots of good.

Conversely if we freeze article development at that point, and wait till the heat of the moment passes before we're in a position to start building a perfect article, only the overburdened regulars and the trolls with a stake in the topic will be left to work on the article for the long term. And we will have lost the readers that could have benefited of a curated selection of the best available sources at all times during the development of the event.

Yes, this way of working implies lots of clean-up after the fact, but that's why the deletionist mindset was an asset to the project; on the other hand, preventing content from ever being included has not worked as well, except for the very obvious vandalism. The best content in controversial topics appears when some editors try to push a point of view in, and the others push to take it out until only the very well supported facts have been distilled. Vetoes to content of a specific kind do not work as well as this adversarial back-and-forth. A systematic reminder to review articles that were created in this dirty way would be our best option.

But hey, if you want to work for a more reliable encyclopedia, with articles carefully compilled from the beginning by experts from well-researched and stable sources, I hear Citizendium is still up and running! Meanwhile here we've learned that we work best when we throw shit at the wall and see what sticks, doing the clean-up afterwards instead of . Our core policies of WP:NPOV and WP:CONSENSUS were designed to produce the best results under those circumstances. Diego (talk) 11:14, 31 October 2017 (UTC)

The GNG and the SNGs are geared towards demonstration of enduring notability, either by showing it is there (GNG), or the likelihood that already exists or will come to be by some criteria of merit that is enduring. News topics are the black sheep compared to this, and one of the reasons we have to recognize how a newspaper differs from an encyclopedia - the former is worried about the here and now, the latter about the long-term.

That said, the IMPERFECT idea still is necessary to consider, and since we can't stop creation of articles without becoming a closed wiki (for all purposes), we should be looking to find a process that "blesses" a article on a current event as having enough legs after a reasonable fair period to keep the article so it can be developed more, or otherwise slot those away into lists or deletion if it turns out to be a non-story. As long as we have WP:NOT#NEWS (and very recent polls and RFC show there's no consensus to change/improve/remove it) we need a method of transforming content written as a newspaper into content written for an encyclopedia. --MASEM (t) 13:34, 31 October 2017 (UTC)

That attitude is the problem. "Blesses". Goodness. Wikipedia editors do not need the blessing of other Wikipedia editors to improve Wikipedia. If it demonstratedly meets WP:42 it doesn't need prior approval, doesn't need premission, doesn't need review. Processes already exist to deal with inappropriate articles, and community consensus is fine here. These proposals all suffer from the same flaw, which is preventing editors from doing work based on one's own criteria of what work is valid or is not valid. Instead, let people create stuff. If there's a problem with an article, we can have discussions (AFD, for example). But we need to stop trying to interfere with editors trying to be helpful. Stop the bad-faith stuff ahead of time, but the rest of this is not useful. --Jayron32 14:04, 31 October 2017 (UTC)

Maybe this is right, though I still think that some months out from a event and one can show that it only had a burst of coverage for a day or so and then never was talked about again (a quick check from google news), that we shouldn't be keeping the article. That still leaves the question of why there's conflict at the AFD for the topics in the OP, and this might come down more to establishing the difference between the burst of news and reactions that happen from an event (which are not contributors towards notability), and the enduring significant coverage from secondary sources that we'd expect after an event to demonstrate notability. --MASEM (t) 15:18, 31 October 2017 (UTC)

I think User:Jayron pegs it. WP:42. 'nuff said.E.M.Gregory (talk) 15:36, 1 November 2017 (UTC)

Suggestion: inclusion criteria for developing stories

As a follow-up to the above discussion, let me float this wording:

• “Suspected terrorist attacks are not presumed notable for stand-alone articles unless at least one victim dies. The events that do not meet this threshold may be included on appropriate lists.”

This provision would apply to coverage of current event, where the WP:LASTING impact is unknown. The restriction, if implemented, would not apply to historic events. Likewise, it would not apply to events where terrorism-related charges have already been brought. Even though this sounds crass, it may give us some sort of a “hard bottom” to base the inclusion criteria on for developing stories. Is this something that's worth discussing? K.e.coffman (talk) 02:44, 4 November 2017 (UTC)

• No. Just no. Quit trying to create policy that is aimed at preventing good-faith, referenced contributions. WP:42 is sufficient for stand alone articles, WP:V is sufficient for additions to existing articles. If Wikipedia is dying as people say it is, the reason that it would be dying would be behavior like this that attempts to stop people from good-faith non-promotional contributions. --Jayron32 03:13, 4 November 2017 (UTC)

• Hear, hear.E.M.Gregory (talk) 10:27, 5 November 2017 (UTC)

• Jayron32 you literally have had nothing productive to say during this whole discussion. Do you even understand why we need to refine policy? Too many articles have been created with WP:SYNTH and POV-pushing just so they appear notable to the untrained eye. Either say anything of worth to this discussion or let others do it for you.TheGracefulSlick (talk) 04:57, 4 November 2017 (UTC)

• We DON'T need to refine policy here in the way you are proposing. We just need to trust existing processes to deal with this. Just because someone doesn't worshipfully agree with your premise doesn't mean the should be dismissed as having nothing worthwile to say. Your concerns are valid. Your solution is not the only possible solution, and I am arguing that this solution is harmful in ways you so not appear to be considering. "POV pushers" are less of a threat to Wikipedia's health than are entrenched editors who are willing to alienate potential new contributors in a vain attempt to "prevent" such POV-pushers from contributing. Perhaps the Parable of the Tares has some wisdom to add to this discussion "So the servants of the owner came and said to him, ‘Sir, did you not sow good seed in your field? How then does it have tares?’ He said to them, ‘An enemy has done this.’ The servants said to him, ‘Do you want us then to go and gather them up?’ But he said, ‘No, lest while you gather up the tares you also uproot the wheat with them." To wit: When you try to weed out your "POV pushers" by this method, you do more harm by also tearing out the good-faith new contributors. The objection I have is well thought out, and consistent. Your proposal is not productive because its effect will be to drive away good-faith users while at the same time will have little effect on stopping POV-pushers, as you call them. --Jayron32 05:09, 4 November 2017 (UTC)

• No. Just no. Quit trying to create policy that is aimed at preventing good-faith, referenced contributions. WP:42 is sufficient for stand alone articles, WP:V is sufficient for additions to existing articles. If Wikipedia is dying as people say it is, the reason that it would be dying would be behavior like this that attempts to stop people from good-faith non-promotional contributions. Exactly as User:Jayron32. writes.E.M.Gregory (talk) 10:27, 5 November 2017 (UTC)

• Are you aware this isn't my proposal Jayron32? I have supported the idea for lists above but that is it. I just asked that you offer anything productive to this discussion instead of dismissing the several proposals presented here from both sides of the issue. Obviously, there is a problem and current policies are not giving us a clear solution. Clarifying policy on a topic that is a magnet for POV-pushing, synth, and BLP violations is a good reason to have a discussion. No one is driving good-faith editors away by offering solutions; unfortunately, a few of them -- the ones the clean up the mess -- have left this topic altogether because of all the disruptive editing that has gone unaddressed.TheGracefulSlick (talk) 20:26, 4 November 2017 (UTC)
• User:Jayron32, above you characterised some of us as 'elitist' and capitalised that our 'elitism' HAD TO STOP. You also characterised all new article writing as being done by 'sincere' people who are "trying to be helpful", to contrast presumably with all us nasty, insincere, elitists who are actually doing the donkey work clearing up the PoV pushing, BLP violations, crap sourcing, coatracking and synth. You also claim that some of us are saying "if it appears in the news, Wikipedia must never talk about it", something which I have never heard any editor say or even imply. What is being asked for is some clarification - in a world of multiple online news sources - of what 'depth' of coverage actually means, when innumerable sources cover the same story for a day or two, adding their own 'spin' and 'speculation', without any evidence of actual traditional journalistic 'research'. This to me is simply 'acreage' of low value coverage, not depth or breadth (a variety of viewpoints). Simply saying "if it's verifiable it's OK" is the response of an ostrich IMO. It is a simple mathematical certainty that every (relatively trivial), sometimes tentatively, terror related article, needs a lot of effort to 'clean up'. The article will almost certainly not be updated, because only the 'hot-news' is ever widely reported, so we are left with a stub of yesterday's headlines and a good dose of instant comment and speculation from every online news source anywhere in the globe. SHOULD existing procedures be able to deal with this? Yes, obviously. DO they? No IMO, and the more NEWS stories we do, the worse the problem will get. You don't have an answer, I don't have an easy answer, no one has an easy answer, but the people who have recently expressed concern (about our coverage of terrorism and other 'hot' news), and tried to find an answer include Evergreen Fir and Doug Weller. If you don't think there IS a better answer, you could at least try being a tad less offensive to we poor "donkeys being led by lions", or should that be the other way around? Pincrete (talk) 22:43, 4 November 2017 (UTC)

Comment Some modification of this would be required, such as the addition of the word "normally" as a minimum. A 'hard and fast rule' would not be workable since there are very notable 'foiled' attempts or non-fatal incidents. I also wonder if this might be counter-productive since this could be interpreted as "any attack with one death can have an article". The "where terrorism-related charges have already been brought" is also problematic, since in the UK certainly, it is not uncommon for someone initially to be charged with the most serious offence which could conceivably be brought, and then have the charges reduced, I would not be surprised if this happens with 2017 Finsbury Park attack, or 2017 Buckingham Palace incident, one to attempted murder, one to "threatening behaviour", as in both cases terrorist intent may be very hard to prove. The underlying problem IMO is that we seem to be no closer to resolving how/why/in what way WP is NOTNEWS and how that should be applied in this topic area. Pincrete (talk) 19:41, 4 November 2017 (UTC)

• Oppose. Victim count has little to do with notability. Most 1 kill events will not be notable in present day Iraq. They might be notable in more peaceful locales. Notability does not hinge on body count, but rather on in depth and continuing coverage that is driven by other factors (trends, identity of victims, or such quirks such as using an Uber driver failing to navigate to Windsor castle, driving to Buckingham palace, and attempting to draw a 4 foot sword in 2017).Icewhiz (talk) 19:53, 4 November 2017 (UTC) Corrected per comment below. Yes, a sword. That's the reason this whole thing was covered internationally - the "terror" is incidental to the bizarre particulars (but the bizzare can be notable, e.g. Phineas Gage)..Icewhiz (talk) 22:14, 4 November 2017 (UTC)

He didn't 'use' a sword, he 'reached for' a sword (and was sprayed immediately if I remember properly), if we are going to base 'notability' on 'quirks', that's even sillier than 'body count'. Has any source ever written more than half a sentence about the sword? Otherwise I agree with you. Pincrete (talk) 22:02, 4 November 2017 (UTC)

Were getting a bit off track but Icewhiz are you really comparing an incompetent sword-reaching (not wielding) uber driver to a landmark medical case still studied today? Seriously? Sensationalism has bitten you good.TheGracefulSlick (talk) 01:06, 5 November 2017 (UTC)

Surely Gage is much more notable, yet the coverage of for Gage is mainly due to non-medical reasons (just as in sword equipped Uber driver - the grounds for coverage are not terror). I'll one-up a Jack Churchill at you - he wouldn't have passed notability without coverage from the whole "bagpipe, longbow, and broadsword" angle. Both are more notable than the Uber sword-drawer at present, of course.Icewhiz (talk) 06:21, 5 November 2017 (UTC)

• Examples of attacks that are patently notable despite the fact that no one died (except sometimes the perp,) include: the 2016 Ohio State University attack, 2007 London car bombs, 2006 UNC SUV attack, 2015 University of California, Merced stabbing attack, 2007 Glasgow Airport attack, 2017 Orly Airport attack, and many more, including our many Category:Failed terrorist attempts, many of which - like the poorly named Wikipedia:Articles for deletion/June 2017 Brussels attack just brought to AfD by Slick - are low casualty only by chance or due to incompetent bomb-building by perp. This discussion is mere WP:FORUMSHOP by User:TheGracefulSlick who, failing to get as many attack article deleted as she wishes to, wants to change the rules to favor her POV.E.M.Gregory (talk) 10:47, 5 November 2017 (UTC) strike my erroneous assumption.E.M.Gregory (talk) 15:26, 5 November 2017 (UTC)

I think all contributors are agreed that 'body count' is probably not a viable guideline. However since TGS did not open the thread and has not opened any section of it, it's difficult to see how she can be forum-shopping. Also, while you happily throw around accusations of PoV to her and me, you rarely identify what that supposed PoV is. I disagree strongly with you about whether innumerable 'stubs' is the most coherent and effective way of informing the reader about terrorism in Europe (I rarely venture outside European subjects), whether this provides context. Apart from that, I don't even know what our PoV supposedly is. Pincrete (talk) 13:01, 5 November 2017 (UTC)

• E.M.Gregory do you want to retract your false accusations? I did not open this thread nor did I even make a proposal. This just looks like an attempt to throw more mud at me. And I'll bite, what PoV am I supposedly pushing? For whatever reason, it certainly isn't to further this conversation or better the encyclopedia.TheGracefulSlick (talk) 15:17, 5 November 2017 (UTC)

• Oppose With all due respect to K.e. Coffman, whose view I respect greatly in many areas, I think that would not be helpful for the reasons stated by Jayron. Coretheapple (talk) 13:44, 5 November 2017 (UTC)
• Oppose Something like the 2010 Times Square car bombing attempt didnt even occur yet is definitely notable. Number of deaths really shouldn't matter. - Galatz^Talk 14:08, 5 November 2017 (UTC)
• Oppose. Has nothing to do with body count, only to do with in-depth coverage in independent, reliable, secondary sources.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  11:05, 8 November 2017 (UTC)

NOT#MENU

• I would support including "Not menu" into the page; it's related to "Not advert" / "Not indisriminate" but it would not hurt to call it out. Sample: :*"Each of these breweries has a menu that features a selection of American-style cuisine including pizza and burgers, steaks, salads and desserts. Each region (and sometimes, each store) has its own selection of beers and beer logos. It is rare to find two beers of the same name or taste in two different stores!" Etc. .

We might as well add this to the guideline. K.e.coffman (talk) 01:40, 7 November 2017 (UTC)

• Oppose – Key points about the foods and fare that companies and food manufacturers create and purvey should not be wiped entirely from articles. Conversely, I also understand that articles do not need to go into great intricate detail covering every foodstuff a company/organization is involved in. These types of matters in articles should be judged on a case-by-case basis, rather than by the ambiguous ten-character phrase, wp:notmenu. North America¹⁰⁰⁰ 10:57, 7 November 2017 (UTC)
• I also disagree with the inclusion. Its too tough to say because certain things are going to be different. Like McDonald's kosher vs non-kosher locations in Israel could be considered under a NOTMENU. Referring to differences relating to beef in India would also need to be removed. Half the content on Big Mac would have to be removed. I think this is something that needs to stay with editors judgment. - Galatz^Talk 17:43, 7 November 2017 (UTC)

• Support. We are not a restaurant menu, nor a listing for various products. I don't think this redirect was highly needed, but it doesn't hurt. It would help if we could see examples of problematic content that were removed as 'menu-like' though. --_{Piotr Konieczny aka Prokonsul Piotrus| reply here} 09:51, 8 November 2017 (UTC)
  • If you think we need to see such examples to evaluate the intent and impact of such a rule, why do you support it in absence of that evidence?  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  11:03, 8 November 2017 (UTC)
• Oppose as redundant; see my comments above this unnecessary !voting section for details.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  11:03, 8 November 2017 (UTC)
• Comment We fall somewhere in between. If we are talking a notable restaurant, we're likely going to mention the general types of foods, and if they have a signature dish they are notable for, that will be included, but we obviously don't need their menu. If we are talking a food manufacturer, a list of their major products but not every variety or packing type. There's a level of resolution here that we should be careful with. --MASEM (t) 14:27, 8 November 2017 (UTC)
• Support as this goes hand in hand with WP: Indiscriminate and WP:Trivia, two things we established early on. SwisterTwister talk 23:20, 10 November 2017 (UTC)

Discussion of "not a laboratory" proposal

This is a tricky thing. I think we need to make a distinction between research that observes what happens in WP vs. research that involves surveying users (not messing with content) vs research that intervenes in WP and sees what happens. The first shouldn't bother anybody (and is done all the time); the second can get into spammy territory and abuse of Talk pages but doesn't harm content, but the third is problematic on a couple of levels: (a) editors can become research subjects without their informed consent and b) content might be generated that has nothing to do with the serving the mission of WP. Some examples of each bucket include:

• • 1st bucket (observations)
    • so much of this is done and even gets published, and is tracked at the WMF Research blog

• • 2nd bucket (surveys)
    • May 2006 Wikipedia:Administrators'_noticeboard/Archive55#Advertising
    • August 2009 Wikipedia:Administrators'_noticeboard/Archive200#Using_Sitenotice_for_Survey
    • March 2011: Wikipedia:Administrators'_noticeboard/Archive222#Researchers_requesting_administrators.E2.80.99_advices_to_launch_a_study
    • July 2014: Wikipedia:Administrators'_noticeboard/Archive263#Allowing_email_survey

• • 3rd bucket (interventions)
    • May 2008: Wikipedia:Administrators'_noticeboard/Archive144#Proposed_research_test_for_better_editing
    • May 2012 Wikipedia:Administrators'_noticeboard/Archive235#Articles_created_for_the_good_of_education (about T. Mills Kelly, the teacher who had his class write fake articles)
    • Sept 2012, an experiment in de-P about giving "edelweiss" awards to random (?) new editors to encourage participation... discussion there archived here and (here is the accompanying paper, and an article in Die Zeit)
    • August 2016 here - the thing about someone creating a bot to write articles and uploading a bunch of them to see what would happen..

There was another recent "intervention"-y kind in the last year or so but I cannot find it... -- Jytdog (talk) 16:39, 5 October 2017 (UTC)

• @Jytdog: This, and by extrapolation this, were you thinking of? Round about April just gone. — fortunavelut luna 23:32, 5 October 2017 (UTC)

Yes! The "anthropology" work the class was doing was discusssed here. Thank you! Jytdog (talk) 23:49, 5 October 2017 (UTC)

Some points of wording I would see to include:

• We do not want researchers to add content to mainspace (and any other space?) that would not normally be added by any other well-intentioned editor that conforms to WP's content policies. Eg, adding purposeful hoaxes to see how far those fly is completely unwelcomed. But adding well-sourced content on a notable topic that just doesn't happen to have an article is less of a problem. (Hence why we tolerate college class "improve a WP article" assignments.)
• We do not want researches to try to sway editors. You can ask editors to participate in a survey or interview them, but attempting to alter behavior ala the "edelweiss" award listed above, that's a problem.
• COI has to be emphasized. It is not a problem that funded research that involves editing WP is bad, it is only when the the targets of that editing have clear ties to the agency providing the funding.
• Any such research approaches should be announced and approved to an appropriate venue. If for some reason there is a need to keep the research "secret" ( ala a blind test) from normal editors, we should request them announce and get confirmation via the ORTS, and tag affected articles with that ticket (which they don't need to spell out anything, but a OTRS member would be able to confirm the "research" was accepted via ORTS.) --MASEM (t) 20:46, 6 October 2017 (UTC)
  • And i would add - "editors participating in a research project are obligated to follow the same content and behavioral policies and guidelines as all editors; this includes engaging with the editing community." or the like. The Carolineneil account clearly was under the misapprehension that they could just do as they pleased under the false notion that "this is an experiment using WP so get out of my way". Jytdog (talk) 21:03, 6 October 2017 (UTC)
    • Yes, that was a point I think I was thinking about but forgot to include. If there are multiple editor accounts involved in a valid research project, there needs to be at least one person that is going to take responsibility for them all, if those other accounts do not appear to otherwise take such actions. Obviously this is getting past what would be at NOT, but it should be a framework to be followed. --MASEM (t) 21:25, 6 October 2017 (UTC)
      • Seems like we're getting awfully close to calling for a WP:IRB. GMG^talk 21:34, 6 October 2017 (UTC)
        • Not with so much formality but in that vein. I think we want to 1) have awareness of any research that requires the research team to edit WP in some way to conduct the research, even if the awareness needs to be kept to a limited number of people and 2) have enough of an idea of what the edits will do on WP to make sure it doesn't mess up en.wiki's goals. We won't have a way to force academics to follow this, but we will be a lot more lenient and understanding of mistakes if these are done upfront. --MASEM (t) 22:14, 6 October 2017 (UTC)

drafting

While many observational studies are conducted on Wikipedia, Wikipedia is not a laboratory. Breaching experiments done in order to make a point are not acceptable, nor is adding content in order to study its effect on editors or readers. The editing community actively maintains content and anything that is added to WP will end up absorbing volunteer effort; volunteer time is the lifeblood of Wikipedia and abusing this resource for purposes outside of the mission of Wikipedia is not acceptable. Experiments involving the editing community require informed consent of the editing community in order to be ethical. People who edit Wikipedia in order to conduct experiments are not exempt from Wikipedia's content and behavioral policies and guidelines, and need to follow community norms for interacting with other editors to discuss behavior or content. If experimenting editors are paid for their work in Wikipedia they must disclose that per the WP:PAID policy, and if they are not paid they still should disclose any outside interest that is driving their activity in Wikipedia, per the WP:COI guideline. See also Wikipedia:Research and Wikipedia:Research recruitment as well as m:Research:Index

thoughts? Jytdog (talk) 06:11, 8 October 2017 (UTC)

I'm concerned that the anger that's understandably been generated by one recent poorly-conducted experiment is driving unwarranted negativity about experiments in general. Perhaps I can counter with a couple of more positive examples:

• In this experiment (see Signpost report), a random selection of articles on tourist destinations was improved, and it was found that tourism to those destinations increased. Some articles were improved, no articles were made worse, no bias was introduced, and we learned something about Wikipedia's impact. Seems like win-win all round, yet this would be prohibited under "nor is adding content in order to study its effect on editors or readers."
• Suppose I want to know which of two welcome templates is superior. I decide to alternate between the two templates as I welcome my next hundred new users, then go back and look at whether one is better at promoting continued engagement. Seems harmless and useful, yet based on the proposed wording, it would be unethical for me to do this unless I first obtained informed consent. Seems like an unnecessary hurdle.

I appreciate that we might clarify that researchers are expected to adhere to the usual P&G, including WP:POINT, WP:PAID and WP:COI. But I think "Wikipedia is not a laboratory" needlessly discourages the best tool we have for learning about ourselves and the project. Adrian J. Hunter^{(talk•contribs)} 08:22, 8 October 2017 (UTC)

I certainly would not agree with trying to ban "adding content in order to study its effect on editors or readers", so long as this is done in accordance with the 5 pillars etc. The Finnish paper on Spanish tourism is a good example. But why were all contributions to the Dutch WP reverted? Poor Dutch perhaps? Johnbod (talk) 11:54, 8 October 2017 (UTC)

The Dutch translator translated non-free text from highly promotional tourist brochures. See here. Adrian J. Hunter^{(talk•contribs)} 14:23, 8 October 2017 (UTC)

So content was added to four WPs using tourism brochures as sources .... and the authors claim to show that this kind of promotional editing makes people buy more. Everybody in the paid editing editing ecosystems must be delighted and now has a "scientific" paper they can wave at clients. Nice job screwing the WMF movement by "proving" that promotional editing can "stick" and can lead to higher sales. ARGH. Jytdog (talk) 16:30, 8 October 2017 (UTC)

Negative, that problem was limited to the Dutch Wikipedia edits, which were all reverted. Content in the other three Wikipedias was translated from the English and Spanish Wikipedias. Adrian J. Hunter^{(talk•contribs)} 20:59, 8 October 2017 (UTC)

As I understand it they used the same content and same sources everywhere. So you will need to prove your "negative". Jytdog (talk) 21:03, 8 October 2017 (UTC)

Copied from the signpost comments:

And on the German Wikipedia (edits included , , , ), there was some milder pushback regarding e.g. lack of references, travel guide language, and unencyclopedic illustrations, to which one of the researchers responded, stating later that there had been an attempt to address the issues. See de:Benutzer_Diskussion:Oltau/Archiv/2014#Spanischkurs and de:Benutzer_Diskussion:Mefk81#Spanischkurs. Regards, Tbayer (WMF) (talk) 22:06, 6 August 2017 (UTC) Jytdog (talk) 21:06, 8 October 2017 (UTC)

Translating one of those diffs: The central coast is characterized by its long and wide beaches of golden sand and the famous and spectacular dunes of Maspalomas, as well as salty water ponds and a palm tree oasis. Jytdog (talk) 21:11, 8 October 2017 (UTC)

I'm just going by the paper, which says, "Added text was translated mostly from the corresponding Spanish or English language Wikipedia pages." It also says that 93% of the added text was still there at the beginning of the year following treatment (I presume that excludes the Dutch additions), so it can't have been too bad. You're right that there were a few issues on the German Wikipedia, but the experiment expanded 120 articles. I know of no non-automated process that could generate that much expansion without a few issues. You'd get vastly more problems if 120 articles had been expanded organically. Adrian J. Hunter^{(talk•contribs)} 03:13, 9 October 2017 (UTC)

Adrian, come on. If you are going to take the word of inexperienced editors, that their work complies with WP policies and guidelines, I have a bridge to sell you. (and i think you already know that many accomplished academics arrive at WP and demonstrate quickly that they have no idea how this place actually works). And expansion for expansion's sake (much less with bad content) is nobody's ideal of a good thing. You appear to be sticking to your guns even after you have been given direct evidence that the content they added was really bad and this is just weird.... Please don't keep advocating for this badly done research project that has harmed what we do on at least two levels. Maybe this comes from an instinct to protect other academics... I don't know what is going on with you but what you are doing and writing here is weird. Jytdog (talk) 16:51, 9 October 2017 (UTC)

You quoted that there was "some milder pushback" on the German Wikipedia and provided four problematic diffs. This is not compelling evidence that the bulk of the expansion to 120 articles across the French, German and Italian Wikipedias was, considered in aggregate, "really bad". On the contrary, the reported persistence of 93% of the content is evidence that it was judged acceptable by other Wikipedians. Have you found other evidence I'm unaware of? Adrian J. Hunter^{(talk•contribs)} 05:35, 10 October 2017 (UTC)

You have made it clear that you are committed to doing no work to examine what happened but rather to taking these authors at their word, and that you will continue "defending" on that sole basis. That is not a discussion that is worth my time. Jytdog (talk) 10:59, 10 October 2017 (UTC)

The "milder pushback" comment came from the Signpost discussion, not the published paper. In any case, that this particular study didn't follow our guidelines doesn't invalidate the results (that expanded Wikipedia articles are correlated to increased tourism, so the articles have influence in our readers), and it doesn't prevent that other similar studies could follow our content guidelines more closely and therefore improve the encyclopedia while performing their experiment. Diego (talk) 11:56, 10 October 2017 (UTC)

Moreover the conversations show that they got most of the content from the Spanish Wikipedia itself (so many of the errors were made because such content was live, and had not been removed from that project), and that they engaged in constructive discussion, trying to learn and amend their failures. So I fail to see how this is any more disruptive than a pack of wilful newcomers being given a task to expand a topic - which is something that the WMF encourages. Diego (talk) 12:12, 10 October 2017 (UTC)

The study was flawed in several ways. Most importantly it was correlational only and correlation is not causation. The study was also not blinded so there is a huge risk of confirmation bias in the experimental design. "Proving" that promotional editing pays, is so damaging to us, and so useful to the paid editing ecosystem that sells the very idea that it is essential to use Wikipedia for promotion - that if you don't have a shiny Wikipedia page you are losing business. So, so damaging. That you fail to see the damage, is something I can only try to correct.

And again these researchers did not edit WP with the aim of improving Wikipedia. They came here to "test" their hypothesis - they needed to get content to "stick" in order to do their shitty experiment. Same as the chemistry editor. They did not give a flying fuck about Wikipedia or its mission. They used Wikipedia. Abused it, actually. Jytdog (talk) 12:33, 10 October 2017 (UTC)

This was a translation activity between the different language projects of content we already had included, it's nowhere similar to promotional paid editing.

They came here to "test" their hypothesis - they needed to get content to "stick" in order to do their shitty experiment. Same as the chemistry editor. I've asked before, if this has the potential to improve the project, how is it a bad thing, and why should we care about the editor's motivations to improve the project? I fail to see how this is damage, because to fail to advance a coherent argument and merely present us with some indignant doom-predicting warnings.

They did not give a flying fuck about Wikipedia or its mission. They used Wikipedia. And you haven't answered my main question. How was what these people do any different in this respect to a bunch of students editing Wikipedia to complete their class assignments, which is considered good for the project? Diego (talk) 12:52, 10 October 2017 (UTC)

User:Diego Moya - "How was ... this ... different from a bunch of students editing Wikipedia to complete their class assignments, which is considered good for the project?" Students editing Wikipedia to complete their class assignments is not always good for the project. Often such assignments are very problematical. Does that answer your question? Class assignments are only good under precise terms, which, unfortunately, in my experience, are often not met, so that class assignments should not be considered good. Is there another question? Robert McClenon (talk) 00:34, 19 October 2017 (UTC)

And yet, the community hasn't banned student class assignments but merely offer guidance on doing them well, and a mayor player -the WMF- actively encourages them, so we consider that the benefits (expanding the encyclopedia and training new editors) may outweight the problems. So there's really no substantial difference that merits treating the two phenomena differently. Yes, this answers the question. Diego (talk) 09:15, 19 October 2017 (UTC)

First, the WMF is assuming, in my opinion incorrectly, that the class assignments may result in improvement of the encyclopedia. In my opinion, they often do not. (The ones that I am aware of did not.) Second, the experiments in question are not intended to improve the encyclopedia or its use. Robert McClenon (talk) 00:48, 20 October 2017 (UTC)

Students in the education program also come to Wikipedia with a COI. They often (not always) behave exactly like paid editors do and like these people did. Students have written to me: "Stop changing my edits! My teacher needs to see this so I can get graded!" You can change "teacher" to "client" and "graded" to "paid" and everything is the same. It is about serving an external interest for the editor. Same thing with these "researchers" - "I need to get these edits to stick so the experiment will "work"."

COI is what it is. We cannot eliminate it but it needs to be managed.

(and btw the education program is not widely loved. it is more or less tolerated. kind of like paid editing is, but with less revulsion with regard to motivation but perhaps more discomfort with regard to the burden it puts on the volunteer community to clean up after them and to deal with them.) Jytdog (talk) 13:14, 10 October 2017 (UTC)

Ok, we then agree that this needs to be managed. This leaves us a lot of leeway in how to manage it, though, besides the total ban proposed by this thread. In fact the draft proposed below by Robert McClenon and GreenMeansGo seems much more nuanced and practical, allowing experimentation in a controlled way. Diego (talk) 14:49, 10 October 2017 (UTC)

Both the original version and the version that is said to propose a total ban were actually nuanced, but there was a lot of reacting to an unseen draft and then to a first draft. The current version is better, but no one was ever proposing a total ban on experiments. What was being proposed was a total ban on undisclosed experiments, and that is still being proposed. At least, it is my intent that undisclosed experiments be banned. Robert McClenon (talk) 23:02, 17 October 2017 (UTC)

• Some drafting is going on at User:Robert McClenon/NOTLAB. Jytdog (talk) 16:37, 8 October 2017 (UTC)
• Maaaybee bringing this down to a slightly more cordial level, anyway you look at it, if it's research involving human subjects (editors), then it needs to be held to a higher standard of oversight, as it is in every corner of academia. Maybe their review boards are giving it the go, because they don't understand the impact on human subjects, and all they see is an an online encyclopedia, but they're wrong. At the very least, the community should be given the opportunity to review such research, and determine that it's harmless, rather than being a free public lab rat, and trying to clean up any mess after the fact. I'm not saying all such research is harmful, or that any particular design is; all I'm saying is that it should be the community, not self-interested researchers, with anybody's guess what level of understanding of the project, who get to make that decision. GMG^talk 12:50, 10 October 2017 (UTC)

• Yes, that seems the way to follow; informed consent is a golden rule for all ethical experimentation. The best proposal so far is that by user Johnbod to create the figure of a community liaison for experiments, that researchers should contact with to ensure they know the right procedures. Diego (talk) 12:58, 10 October 2017 (UTC)

I've gone back and forth about what the appropriate forum might be, and I can't come to any strong position other than it depends. In some cases, a WikiProject discussion might be perfectly appropriate for designs with very limited scope. For example, I don't see why a user survey among WikiProject Medicine participants, regarding interpretation of MEDRS, should need to go farther than WikiProject Medicine for approval. However, a very broad design may need to go to Village Pump or AN for approval, and if someone brought a topic up in a forum that was too narrow for the scope of the design (like a WikiProject), it would be up to participants there to escalate the proposal to a higher level of oversight, by, for example, moving the ongoing discussion to a broader forum, or posting a notification in a broader forum regarding the ongoing discussion.

But I think the important thing is that there is a clear expectation set out that prior approval of some kind is expected for any design that a reasonable editor could construe as invasive or disruptive. GMG^talk 13:10, 10 October 2017 (UTC)

• I think things are winding down to a proposal over at User:Robert McClenon/NOTLAB. Anyone is free to weigh in on wording or scope prior to a formal proposal. Comments welcome on the associated talk page. GMG^talk 21:53, 16 October 2017 (UTC)