Wikipedia talk:Short description/Archive 9

This is an archive of past discussions about Wikipedia:Short description. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 5

←

Archive 7

Archive 8

Archive 9

Archive 10

Archive 11

→

Archive 15

Hi folks, just a quick heads-up about a potentially confusing situation we're working on fixing: The Wikipedia apps (iOS and Android) have moved over to local descriptions, but editing can still take people to Wikidata. This will be fixed in the next app release. See phab:T257867 for more info. /Johan (WMF) (talk) 18:00, 8 December 2020 (UTC)

A short update: Because of technical reasons (the way authentication will work in the apps now, with OAuth instead of a cookie, which means more work for the developers to fix this since we're talking about two different wikis being involved), this will take slightly longer than anticipated, but should still be fixed soon. /Johan (WMF) (talk) 06:01, 20 January 2021 (UTC)

It is my understanding that Wikidata duplicates information populated from the lede, and the short description gadget also duplicates information imported from Wikidata to be used for the short description. We are even encouraged to use this duplicated information on this project page by being told it is public domain, so it is ok to copy it. It is also my understanding that this imported duplicated information is intended to be useful for the main purpose of "augmenting search results", as described in the nutshell. It is also my understanding that most users view on mobile, and we are instructed that our primary concern should be distinguishing articles with the same title.

I was questioning why we have a tiny bit of editorial content advice that seemingly contradicts the entire way Wikidata, and the short description gadget imports, and populates duplicated information. It can be found at: Wikipedia:Short_description#Writing_a_short_description, where it states, "avoid duplicating information that is already in the title". As, you can see this also contradicts the primary concern of distinguishing articles with the same title, as well as contradicting the stated main purpose of this project to augment searches, since most effective searches are based on strategic use of keywords, not how well they "mesh with the prose".

So, I did some digging to see how this bit got introduced, and found this diff where the editor implemented it without any discussion and no explanatory edit summary other than "clarify". It almost seems as though it got put in as a personal type of editorial preference as opposed to something consistent with the project goals. I propose that it should be removed [reworded]. Huggums537 (talk) 15:33, 16 January 2021 (UTC)

I don't see a conflict between these pieces of guidance. The short description is intended to "augment search results", not the searches themselves which AFAIK cannot preferentially search what is in the short description - the short description doesn't function as keywords indexed in a database, but just more words to search in a full-text search. And in the search results themselves, the title is already shown - if you want to distinguish between articles with similar titles, just repeating the title that is already displayed doesn't help with that. Nikkimaria (talk) 16:22, 16 January 2021 (UTC)

Nikkimaria, "the short description doesn't function as keywords indexed in a database, but just more words to search in a full-text search." Exactly. And, those "more words" being searched is usually information duplicated from the article itself. Since that is where the description comes from. So, your keywords effectively come from the article and go into the description. Also, I think repeating some of what is in the title does help distinguish articles/titles/subjects with similar titles, because a reader could think the link is related to another title with a similar name if we are using words not related to the title/article/subject of the article we are describing, but if we use words related to the article and title, they will likely know the link is also related to that article as well. Perhaps I should change my proposal from "remove" to "reword" to allow some information from the title to be used as opposed to duplicating the whole title? Huggums537 (talk) 22:20, 16 January 2021 (UTC)

Sorry, I'm not following. Why would repeating words from the title in a piece of text displayed immediately underneath the title be helpful? "John Smith: American baker" is distinguished from "John Smith: British boxer" without us having to say "John Smith: John Smith the American baker". Can you give a specific example in which repeating part of the title would help? Nikkimaria (talk) 22:27, 16 January 2021 (UTC)

Thank you for your question. Right, so using the examples you gave and my current proposal, I would use some of the information from the titles of these articles to craft my descriptions, but avoid duplicating the whole title. So, it would make use of the distinguishable information in the titles while omitting the non distinguishable information. It might look like this for the "John Smith: American baker" description: American baker (Born DOB), and similarly, the "John Smith: British boxer" description might be: British boxer (DOB-DOD). (I made both of these descriptions up solely from the titles you gave me since I couldn't find articles on them, which kinda proves my point that we sometimes need some information from the titles to create descriptions). The problem is that the current wording of the guidance discourages the usage of any information from the title, and I think this wording needs to be changed to allow for some information from the title to be used. Huggums537 (talk) 23:16, 16 January 2021 (UTC)

@Huggums537: Information may be duplicated in different places because they are accessed under different circumstances. Short descriptions are not the same thing as the short description helper and importing information from Wikidata, available under a CC-0 licence is not the same thing as writing a short description. Nevertheless, it is important to give guidance to editors who write short descriptions, and we have accurate guidance.

When editors on some platforms start typing a search term, they are presented with article titles and short descriptions together. See File:Mobile search for Worl.png for an example. Because they see both the article title and the short description, you should be able to work out that it is utterly pointless putting the same information into the short description as is already in the title. Of course, the editor sees a choice of article titles, which will be similar, so the extra information in the short description is what helps them decide which one is the article they want.

So yes, our primary concern should be distinguishing articles with the same title. That's overwhelmingly the main use of short descriptions on Wikipedia, and it would be stupid to suggest that we shouldn't treat it as our main concern when crafting short descriptions. --RexxS (talk) 18:08, 16 January 2021 (UTC)

RexxS, I understand that Wikidata, and short descriptions are in different places, writing descriptions are not the same thing as importing them, and copying public domain is not the same thing as writing them. However, none of these facts are relevant to any of my points since it doesn't matter if they are in two different places or not because the material comes from the same place - the article itself, no matter if you write it, or import it, the description is always going to have to be related to, and rely on information contained in the article itself, including the title. So, no I think we are getting inaccurate guidance on that bit. Also, I think it's not "utterly pointless" to include some information from the title as I reasoned in my reply to Nikkimaria above. I think I'll strike my original proposal, and switch from "remove" to "reword" to allow for the use of some information from the title, but forbid exact duplication of the whole title. That seems like a reasonable compromise to me. Huggums537 (talk) 22:20, 16 January 2021 (UTC)

P.S. I also noticed that even in the example provided here: File:Mobile search for Worl.png, both World war I and World war II could not avoid duplicating the title word, "war" in the descriptions. Huggums537 (talk) 00:33, 17 January 2021 (UTC)

@Huggums537: On the contrary, you clearly don't understand it at all. The point is about when the information is being used in different places, and is nothing to do with where the information is stored.

It is also nothing to do with a keyword search. Nothing. Zilch. There is no keyword search. Try to understand that.

"the material comes from the same place" No it doesn't. Originally, the short description was drawn solely from Wikidata. Then, until very recently, the short description was drawn from the local short description, and drawn from Wikidata if the local description wasn't set. The local short description may be supplied by use of the {{short description}} template or it may be generated by an infobox, for example. That's a long way from "coming from the same place"

"the description is always going to have to be related to, and rely on information contained in the article itself, including the title" That's nonsense. The first part is a truism, and the second part is patently false for the reasons I already gave. So, you're quite wrong. The current guidance, "avoid duplicating information that is already in the title", is exactly accurate. The article title shouldn't be part of the short description, and I defy you to find an useful example of where the title wording would be a useful part of its short description. How does it help anyone to see:

John Smith (baker)
American baker
John Smithe (banker)
English banker

when you could have

John Smith (baker)
Born 1960 in Chicago, US
John Smithe (banker)
Born 1970 in London, UK

How does repeating part of the title help to distinguish between the results displayed when you type John Sm?

P.S. How does the word war help to distinguish between World War I and World War II? Hint: it doesn't. You would convey exactly the same information by having "1939-1945 between Axis and Allies" and "1914-1918, starting in Europe", because we don't see the short description dissociated from the article title (which carries the rest of the information). --RexxS (talk) 01:00, 17 January 2021 (UTC)

RexxS, I see how you were able to reword the "John Smith" descriptions in a way that doesn't duplicate any of the title information, but many editors may not be talented as you are. Also, when you display it in a format where I can clearly see the examples, it does make me realize you have a fairly good point about it not really benefitting the reader as much by duplicating some of the information. However, that makes me wonder if the stated goal of the project should not be shifted towards making the descriptions be much less about "search results" and more about "readable annotations" like your example shows. The only problem is this would not help any of the people searching for "American bakers" or "English bankers" whatsoever. A Wikipedia search for both keywords yields this and this. So, how do these keywords have "nothing, zilch" to do with it? Also, I see nothing wrong with the war articles using the word "war" from the title. While the word alone, and of itself does not distinguish between the two articles, when it is the only word being used from the titles, and it's being used in conjunction with other elements of the description, it appears to be distinguishing just fine as it is, so I'm puzzled as to why you're insisting it doesn't. I might shorten them up a bit, like you did with "1939-1945 between Axis and Allies" that's not bad, but I sure wouldn't use "1914-1918, starting in Europe". and if we're that bent on avoiding words in the title, then lets ask all Wikipedians to have their thesaurus handy so we can change it to "1914-1918, conflict starting in Europe". Huggums537 (talk) 04:44, 17 January 2021 (UTC)

P.S. Another thought. Those saying, "how does it help to repeat information already in the title?" are essentially espousing the same sentiment of those who have opposed this project to begin with, and for very similar reasons. Much opposition to this project reasoned that short descriptions are useless because they really just repeat the same information already contained in the article, so they did not like the "redundancy" of the descriptions in the articles. Those not in favor of allowing some words from the title into the description also appear to have similar aversions to "redundancy" without recognizing the benefits. And, before anyone asks me to show them some benefits again, let me remind them that I already showed them two, and nobody wanted to listen. So, I'll not be repeating myself. Huggums537 (talk) 04:44, 17 January 2021 (UTC)

The guideline says "avoid duplicating information"; it doesn't say "duplicating information is prohibited". There may be circumstances (that I can't see explained above) where duplication would be desirable but the guideline provides correct guidance. Mentioning "war" twice is not something to worry about. Johnuniq (talk) 06:34, 17 January 2021 (UTC)

Thank you for bringing this up, because people seem to be reverting edits with the wrongful understanding that "duplicating information is prohibited" or just as an excuse to revert someone, and this is why I think the guidance needs to be reworded. Here's a couple examples: here the reversion seems to be only using the guidance as an excuse to revert because they reverted back to a revision that still contains the word "game" from the title, and here, the editors appear to have thought the guidance meant "prohibited" so, it forced them to use an alternatively less specific and/or accurate genre (action) to use in the description. Huggums537 (talk) 14:30, 17 January 2021 (UTC)

Both the "Adventure game" revert and the "First-person shooter" revert made the short descriptions better. I can't find the right guidance right now, but it can often help to think of the short description as completing a sentence of the form: "[article title] is/was a/an [short description]". – Jonesey95 (talk) 15:23, 17 January 2021 (UTC)

@Huggums537: People searching for "American bakers" or "English bankers" will use the category system, which is the job that was designed for: Category:American bakers, Category:English bankers – these are quickly accessible from a Google search outside of Wikipedia. The short descriptions were intended to help someone (particularly on a mobile platform) who was already browsing Wikipedia to use the internal search function to find a particular article, like Jphn Smith, using the fewest keystrokes by helping to distinguish the one they were looking for from other articles with similar titles.

We both know that duplicating a word like war from the title into the short description isn't going to cause problems. The difference between us, I suspect, is where we want the balance of the guidance to be. I want to avoid things like

John Smith (baker)
Person named John Smith who is an American baker
John Smithe (banker)
Person named John Smithe who is an English banker

- to take an extreme example, but I think you'll see what I mean.

From what I can gather, you don't want to see helpful descriptions being reverted simply because they repeat a word from the title in order to make the description less clumsy, and I sympathise with that. I'm willing to be persuaded that better guidance can be given, but removing it isn't an improvement IMHO. You'll find that if you try to make it too nuanced, it will become complex, and other editors will object to it because they prefer straightforward guidance. When John Lydgate coined the aphorism "You can't please all of the people all of the time", he could have been thinking about how to write guidance on short descriptions. --RexxS (talk) 15:56, 17 January 2021 (UTC)

Ok, well I suspect the differences between us are much less than you think, because I bet we can all agree that everyone (including myself) wants to avoid the extreme examples you provided. With that in mind, and seeing how you acknowledged that I want to avoid helpful edits being reverted just because they contain a word from the title to make them less clumsy, I would take your advice about offering "straightforward guidance" that is not too complex, and revise it to say something similar to the following:

• avoid making a copy of, or rewording the title. The use of a single word from the title is not considered "copying" or "rewo rding".Huggums537 (talk) 21:37, 17 January 2021 (UTC)

I'm sorry but I don't agree. The guidance is perfectly clear as it is, in my view, and really doesn't need a "deeming provision" as if it were a binding legal document. So far as I understand it, you came across a couple of reverts that you thought were based on a misreading of the guidance, whereas in fact they simply improved some existing SDs. MichaelMaggs (talk) 22:03, 17 January 2021 (UTC)

It's ok if you don't agree, and you're right that I thought a couple reverts were based on misread guidance, but I think that if those reverts truly were an improvement, then those editors should have no problem making a good argument for why they are good reverts in a proper discussion rather than simply relying on something as arbitrary as, "because the guidance says you can't duplicate that word". Also, I think of my proposal as a reinforcement of a practice that we already have in place as opposed to a "deeming provision". Why do I say this? Well, if we look at some level 1 and 2 vital articles we see at least one word from the title being used in every example: Human_history, Prehistory, Ancient_history, Post-classical_history (uses a word twice), Early_modern_period, Late_modern_period, History_of_art, and History_of_science. So, that's every single level 1&2 article in the history category with the exception of one single article: Civilization. If that's not enough to convince anyone , then I don't know what is. Huggums537 (talk) 03:19, 18 January 2021 (UTC)

I haven't seen a single example where the guidelines have resulted in a poor short description. That's what I'd be looking for, not just odd comments in edit summaries. MichaelMaggs (talk) 10:18, 18 January 2021 (UTC)

Ok, well if all the above isn't enough (including the bits about edit summaries), and all you are asking for me to do is search through 2,624,660 articles to find you some examples of poor descriptions, then that doesn't sound unreasonable at all. And, how could I possibly prove it was the guidelines that caused the poor descriptions even if I were to find them anyway? But, what I can show you are these 60,275 examples of very poor list descriptions (according to many experienced editors, see link in thread below) that all say "Wikipedia list article", and are honestly only there for purely technical reasons, so I would not say these poor descriptions are a result of the guidance, but there certainly is a 100% lack of guidance about it, which is just as bad. So, yeah the guidance isn't as solid as it appears. Huggums537 (talk) 14:46, 18 January 2021 (UTC)

Sarcasm isn't helpful. I wouldn't dream of telling you what do do. I've just tried to explain to the best of my ability why I don't agree with your proposal. MichaelMaggs (talk) 15:49, 18 January 2021 (UTC)

Ok, fair point about sarcasm, but you did say you would be looking for an example of where the guidelines have resulted in a poor short description: ("That's what I'd be looking for..."). I took this to mean that you might consider my proposal if I could show you any examples. If there was a misunderstanding, then I apologize for the sarcasm. Huggums537 (talk) 18:30, 18 January 2021 (UTC)

To take the first of your examples, List of United States Representatives from New York, what extra information would help the reader to determine whether or not this is the article they seek? If there is any then we should add it to the short description. If not then the description is almost immaterial and the current concise, accurate, consistent text seems as good as any. Perhaps we could clarify that "representative" here means a member of the United States House of Representatives, rather than say a diplomatic consul, or that they are elected by New York (state) rather than New York City. Certes (talk) 15:17, 18 January 2021 (UTC)

This is immaterial. If you followed my links, all these list descriptions will be removed at some point anyway. The point is that the guidance needs improvement, not the list descriptions. Huggums537 (talk) 15:39, 18 January 2021 (UTC)

Certes, I've been thinking about your question, and realized it's more material than I was first thinking. I say this because the examples I gave of bad descriptions is one of the main reasons I wanted to change the guidance, not just because I got reverted a couple times out of the many short descriptions I had been doing lately.

I think you have made some excellent suggestions for improving the descriptions, and even though it is a moot point for the list articles, they should be ideas worth remembering for future discussions on improving descriptions. Having this very discussion alone has caused me to have some good ideas on how to improve my own personal description writing style thanks to comments/examples given by RexxS and Jonesey95.

If you follow my links, (or if you've been a member of this project for a while) then you are aware that this project has received a lot of criticism over those list descriptions from many editors way more experienced than myself, and the response from the project has always been that these list descriptions are a "necessary evil" because 1} we have a security issue, and 2} we have a 2 million article quota to meet in order to solve that security issue. The concerned editors were then told not to concern themselves with the matter because all those terrible descriptions would be removed at some point.

However, what I have noticed as a newer outside observer coming in as a fresh member of the project, is that the same line of thinking that created the list article descriptions (just anything to meet that quota, template or bot as many as you can) has also created thousands of bad descriptions. When editors like MichaelMaggs tell me they haven't seen a single example of where the guidelines have caused bad descriptions, they would actually be right since it is this line of thinking, not the guidelines that causes it. This line of thinking is nowhere to be found in the guidelines, but it's this line of thinking that needs to be changed, therefore the guidelines needs to be changed by extension. Why does this line of thinking need to be changed you say? Well, we have met our quota, the wikidata has been turned off for mobile. It's time to stop the old line of thinking, and start moving toward better descriptions. I suspect that one main reason some may not be ready to accept my proposal, is because they are not yet ready to embrace and accept the new line of thinking as well. I think my proposal is a baby step toward that goal. I'm going to modify it, and put an RfC on the project page to see if others are interested in this conversation. Keeping in step with MichaelMaggs concern about "deeming provisions", I propose it be reworded to something like this:

* avoid simply making a copy of, or rewording the title. Sometimes using a single word from the title can't be avoided, or it's more practical, and is not considered "copying" or "rewording", but avoiding words in the title is usually preferred. Huggums537 (talk) 09:08, 26 January 2021 (UTC)

Huggums537 I was in two minds whether to reply to this. So far nobody has agreed with you about this alleged issue. It has not so far as anyone knows resulted in a single problematic short description - you've still not provided a single actual example of that. It's really not worth almost 4000 words of commentary, almost all of which is yours. Anyway, I've added "but don’t worry too much if you need to repeat a word or two for context" to the guidance. I'm hoping that will be uncontentious. MichaelMaggs (talk) 11:55, 26 January 2021 (UTC)

MichaelMaggs, your addition seems pretty reasonable to me, except it allows for more than I would have since some titles are only comprised of two words, but other than that, I would be happy to accept your version and be done with the discussion if there are no other objections. I think I'll remove the RfC template also since you seem to have resolved this... Huggums537 (talk) 12:10, 26 January 2021 (UTC)

@Huggums537 and MichaelMaggs: I think the addition is helpful and seems an acceptable compromise. I can see how strongly Huggums537 felt about the issue, and I believe it's worth spending 4000 words in debate if it helps a good-faith editor to gain a broader grasp of the issues. Thanks both, for your input. Cheers --RexxS (talk) 13:04, 26 January 2021 (UTC)

Please see: Template talk:Information page#"noreplace" is not working.  — SMcCandlish ☏ ¢ 😼  02:43, 13 February 2021 (UTC)

Should we be adding an explicit short description to articles which already have one (or more) via templates? For example, this good-faith edit adds a SD of "Year" to 1917, which already has SD "Calendar year" via {{Year nav}} and a less helpful SD of "1917" via {{Year article header}}. Certes (talk) 12:42, 17 February 2021 (UTC)

Certes, The purpose of embedded (template-created) SDs is to provide consistency and avoid the need to edit each page individually. I'd say it's OK to override them with a better, perhaps more specific, manual SD, but that shouldn't be done unless the manual version really is an improvement. Clearly it isn't in the situation you describe - although it appears to be. Looking at 1918, I see that the SD gadget displays only "1918" and not "Calendar year". The editor may have seen that and quite reasonably thought that "Year" would be an improvement. It seems that {{Year article header}} should be modified. MichaelMaggs (talk) 12:58, 17 February 2021 (UTC)

The article has two SDs. {{Year nav}} sets a SD of "Calendar year" and then {{Year article header}} sets a SD of "1918". The second one wins, which is wrong. {{Year article header}} should use noreplace but does not — GhostInTheMachine ^{talk to me} 14:01, 17 February 2021 (UTC)

I've amended {{Year article header}} to use noreplace and to use a SD of "Calendar year". --RexxS (talk) 15:34, 17 February 2021 (UTC)

Note that the cache is hanging on to the old "Local description", so the SD helper won't show "Calendar year" for a particular article until the article gets a null edit (or an edit) or the cache expires. I have a script that "touches" all of the articles in a category (i.e. makes a null edit), but I would prefer not to put the load on the server for just cosmetic edits. --RexxS (talk) 17:25, 17 February 2021 (UTC)

There were actually three SDs. The 2 above plus the local one. I have removed the local one so all is OK for this article at least — GhostInTheMachine ^{talk to me} 21:53, 17 February 2021 (UTC)

Quote:

Short descriptions should:

• be written in plain text, without HTML tags or Wiki markup
• use sentence case, starting with a capital letter, but avoiding a final full stop
• avoid initial articles (A, An, The)
• be brief: aim for no more than about 40 characters (but this can be exceeded when necessary).

This formatting policy is poorly written and self-contradictory. It needs to be discussed and reviewed. If the short description should be formatted as a sentence, why should we omit the full stop and the definite or indefinite articles? A sentence without a full stop is not a sentence. A sentence without a definite or indefinite article is not a proper sentence. On the other hand, if the short description should be formatted as a short explainatory note, why should we use sentence case? The short description in its current format looks super weird. I think we should vote on whether to adapt a sentence format or an explainatory note format, not something in between. 120.16.92.71 (talk) 11:56, 6 November 2020 (UTC)

Since I've started the discussion, I shall cast the first vote. I support the sentence format. 120.16.92.71 (talk) 14:19, 6 November 2020 (UTC)

The short description is not a sentence. The short description is intended to be the object of the implied sentence "[Article title is/was a/the] short description text". "Sentence case" means that the first letter of the short description should be a capital letter; that is all it means. We use a capital letter and avoid the full stop because the SD stands alone and as such should be formatted more like a section header. – Jonesey95 (talk) 16:52, 6 November 2020 (UTC)

Based on what you've said, the short description is merely a short explainatory note of the article title, then we should avoid using the sentence case too (e.g. Bolivia – inland country in South America). 120.16.92.71 (talk) 03:14, 7 November 2020 (UTC)

I agree - "sentence case" confused me for a while, as well. Since it means nothing other than "start with a capital letter", and the guidelines already state that, I've removed it. MichaelMaggs (talk) 12:33, 13 November 2020 (UTC)

The current article says:

Short descriptions should:

• use sentence case, starting with a capital letter, but avoiding a final full stop

But the Wikipedia Android app says, while editing description:

Start with a lowercase letter unless the first word is a proper noun.

Both guideles are inconsistent with each other. — Preceding unsigned comment added by PardhuMadipalli (talk • contribs) 09:02, 15 January 2021 (UTC)

@PardhuMadipalli: I assume you mean the app documented at mw:Wikimedia Apps/Team/Android/App hacking? You should probably ask there when they are going to update their help text. I believe the original intention was to update the Wikidata description, which does indeed start with a lower case letter unless it's a proper noun. The intention now is that the definitive short description is stored on English Wikipedia, so if the App changes the English Wikipedia short description, it should abide by the convention used here. --RexxS (talk) 21:29, 15 January 2021 (UTC)

RexxS, I have also noticed this. I think that the instructions need to be consistent between the two. Laterthanyouthink (talk) 13:05, 19 February 2021 (UTC)

For template designers, the Template:Has short description is available which returns the position of {{short description}} in the source wiki-text, or nothing if the template is not present (so it works directly with {{#if:{{has short description}}| ... | ... }} constructions. It finds {{Short as well as {{short and also {{ short, etc. with arbitrary amounts of whitespace after the opening curly brackets. That may be useful to compensate for editors' varying preferences for how they write templates. Is this worth documenting? --RexxS (talk) 22:16, 22 February 2021 (UTC)

What does it return for SDs that are embedded within an infobox or other template? MichaelMaggs (talk) 22:34, 22 February 2021 (UTC)

Nothing. The short description template isn't "embedded" until the wiki-text is expanded. If it's not in the source code, it won't show up. It exists so that an infobox or other template can determine whether a short description has been manually set for the article. For example the infobox call of the SD template might place the article in Category:Articles with long short description, but the actual SD in the wikitext might be quite short. It's not intended to be used instead of |noreplace as it uses much more resources because it has to load the whole page to search the text. --RexxS (talk) 02:40, 23 February 2021 (UTC)

Please see: Template talk:Disambiguation page short description#Short?. — Goszei (talk) 01:10, 2 April 2021 (UTC)

If short descriptions are supposed to be the top element of an article, why is AWB moving it below the hatnote, as in this edit? BilCat (talk) 05:00, 8 April 2021 (UTC)

The edit is probably from version 6.1.0.1 of AWB which dates back to September 2019. There was a change in November 2019 to ensure that the Short description is correctly moved to the top. There have been many other changes since then, but the "official" AWB version has not yet been updated. Editors should probably download the latest "snapshot" version or turn off the option to apply general fixes. — GhostInTheMachine ^{talk to me} 09:01, 8 April 2021 (UTC)

I'm the creator of the PearBOT 5 (returning to an encyclopedia near you soon) and the bot parameter which has only been used by said task. I've gotten a bit feedback about the bot parameter not being particularly useful both at Wikipedia talk:Shortdesc helper#Bad edit summaries and Wikipedia:Bots/Noticeboard#PearBOT 5 starting up again. I'm planning on removing the categorization for it from {{Short description}} if noone objects. The intention was for me to easily be able to find issues and make sure no reported issue occurred in multiple places, but this can be done in other ways. Instead some people have used it as the lowest priority backlog ever wasting editor time and spamming watchlists and cluttering category lists with a basically useless category. --Trialpears (talk) 21:52, 3 April 2021 (UTC)

Now removed. --Trialpears (talk) 21:59, 12 April 2021 (UTC)

This has come up in Phillip Adams (American football), a football player who earlier this week suddenly shot and killed a number of people and then killed himself. The initial idea is believed related to medical issues arising from his football career, but while this is not confirmed, there's no RS or the like that this was a planned attack or the like. For all purposes he can be called a murderer (mass murderer even) and appropriately cataloged that way, and the article doesn't hide this in the lede by explaining the situation around his death. But given that the current RSes and statements from authorities do not seem to paint this as a premediated event and something tied to a mental state, they are not treating him as a "murderer" and it would seem UNDUE and BLP to weigh that term equally with his football career.

We've had editors try to add "and murderer" both to the lede and (relevant here) to the shortdesc. While concise should avoid most problems, this is a case where there seems to be a need to also stress that these shortdesc should be neutral as well, and when dealing with BLP (or as in Adams, the recently deceased), that they must not violate BLP policy. It would seem this could be added as an additional bullet point under "Content" ("The description should be neutrally worded, and not violate core content policies like biographies of living persons.") --Masem (t) 14:09, 10 April 2021 (UTC)

Masem, I believe what you are looking for is already stated at Wikipedia:Short_description#Editing_procedures. MichaelMaggs (talk) 16:04, 10 April 2021 (UTC)

Okay I see that there, but I would have expected that as "content" issue to be under that section. Dunno if duplication there is necessary. --Masem (t) 16:43, 10 April 2021 (UTC)

Masem,  Done. Spent some time updating and restructuring the page. Should be more straightforward now, I hope. MichaelMaggs (talk) 15:56, 13 April 2021 (UTC)

Looks fine to me, thanks. --Masem (t) 16:14, 13 April 2021 (UTC)

I noticed a change recently when adding the Short description. In film articles, where I used to see "(Wikidata Import Edit and import)", I now see "American film (Override Export ?)". As I understand it the "American film" populated part is due to some automatically pulling info from the {{Infobox film}} template. This seems to be a fairly recent change, and I'm wondering if there was a discussion somewhere that explains this. — Ched (talk) 12:46, 17 May 2021 (UTC)

This is the edit that changed the behavior. – Jonesey95 (talk) 15:11, 17 May 2021 (UTC)

Thank you Jonesey95, I appreciate that very much. I'll take it up with him - just not sure where yet. — Ched (talk) 18:43, 17 May 2021 (UTC)

Just as a note, in order to avoid decentralized discussion (and have a productive one at that) I have replied at Template_talk:Infobox_film#Shortdesc. Primefac (talk) 10:16, 18 May 2021 (UTC)

 You are invited to join the discussion at Wikipedia talk:WikiProject Short descriptions § Talk page merge proposal. {{u|Sdkb}} ^talk 19:57, 23 May 2021 (UTC)

When editing on the mobile app the "Edit article description" button gives a screen with a character counter. The counter gives x/90 as the basic count. Going beyond 90 ("91/90") produces a message saying "Try to keep descriptions short so users can understand the article's subject at a glance". That is fine, but we do not have consistency. Is the "limit" 40 characters or 90? I recommend we change the guidance here to read "90". That will allow app-using editors to provide useful shorts. (E.g., shorts which appear in the app-articles as 1 or 2 lines, and short enough in the regular screen articles.) – S. Rich (talk) 15:14, 10 May 2021 (UTC)

There is no limit as such, but the short descriptions are intended to display under the article title in one line. On some mobile devices, the description seems to be truncated at or shortly after 40 characters. Hence the advise uses 40 as a suitable cut-off. A cut-off of 90 is much too long and needs to be altered. Is this Android or iOS? — GhostInTheMachine ^{talk to me} 16:27, 10 May 2021 (UTC)

i0S. And I agree 90 is too long. A tech person is needed to change the message on the app, but non-techs like me can change the advice. – S. Rich (talk) 16:46, 10 May 2021 (UTC)

90 is indeed far too long. Have a look at the screenshots on Wikipedia:Short description and you can see that mobile apps start to cut off the text at even lower than 40 in some cases. 40 is a reasonable compromise, and has a lot of discussion behind it. MichaelMaggs (talk) 14:40, 11 May 2021 (UTC)

The template:annotated link is extremely useful in See Also sections, where it expands a very short buzz-phrase (names of articles are typically just a very few words) into a link that is actually useful, because it appends the short description. For example, picking the first three that came into my head:

• Wikipedia – Free online crowdsourced encyclopedia
  • (38 but too short IMO, what happened to "that anyone may edit"?)
• White House – Official residence and workplace of the US president
  • (70 characters and rightly so: any less would be pointless)
• Big Ben – Clock tower in London, England
  • (59 characters and fails to explain that the Palace of Westminster is the seat of the Houses of Parliament)

The discussion in the preceding subsection suggests that some editors are more concerned about form rather than content. How does it matter if the short description runs to two or three lines on a competently written mobile OS? Are we aiming to convey information or are we trying to look pretty? Do we really want to limit the design because Safari is broken? How to iPhones cope with long article names? 40 is a laughable "compromise" - even Twitter 1.0 allowed 80. --John Maynard Friedman (talk) 22:58, 30 May 2021 (UTC)

The short description should be short enough to distinguish the topic from other near matches in a search list - not to fully describe it. There's no reason for example to include the "that anyone can edit" catchphrase for Wikipedia's short desc because the basic " Free, multilingual online encyclopedia" is more than sufficient (even in that "free, multilinguial" could be overkill). "White House" should be "Residence of the United States president" and nothing else, as to distinguish from other possible "white house"s. Etc. It it meant to be terse and concise, not to be verbose. --Masem (t) 23:03, 30 May 2021 (UTC)

No, all you have done is to restate the current policy. You haven't explained why it should be terse, why it should be 40 and not 90 or 120 or indeed as much as is needed to give readers a clear idea of what an article is going to be about and thus whether it is what they wanted to know. I can't help wondering if this is a relict from the early days of wikipedia when mobile data was limited and expensive. --John Maynard Friedman (talk) 23:11, 30 May 2021 (UTC)

People from marketing might like "that anyone may edit" but many of us prefer conciseness. It's not a beauty competition and the purpose of a short description is for disambiguation when searching on a mobile device. The purpose of {{annotated link}} is something else and it should not dictate what happens with short descriptions. Index of underwater diving is one of a few articles where that template causes significant delays and occasional breakage due to exceeding limits. Johnuniq (talk) 23:44, 30 May 2021 (UTC)

Let's not get distracted with the detail of my random example and the reference to 'beauty competitions" seems to be questioning my good faith. The point at issue is that "disambiguation when searching on mobile devices" is one of the uses and equally "should not dictate what happens with short descriptions" since they have other equally valid uses. On my Android phone, I have no issue whatever in seeing short descriptions longer than 40 characters. So are you really saying that the limit is not "for searching on a mobile device", but rather "for searching on a mobile device made by Apple, because Safari is broken"? --John Maynard Friedman (talk) 00:01, 31 May 2021 (UTC)

What has Safari got to do with it? Just look closely at the examples here and see where the text gets cut off. A bit less than 40 characters on the official mobile platform, a bit more on desktop. MichaelMaggs (talk) 08:21, 31 May 2021 (UTC)

Using new vector on Windows I get a cut off at 58 characters on my desktop and 45 on my laptop. The IOS Wikipedia app frontpage I get the lowest I've observed anywhere at 32 characters and for searches I get 42 characters. For the related articles on the bottom of the mobile site I see 41 characters on both Safari and Chrome for IOS. I won't get out my old android again today but when I did that a few months ago it showed very similar results.

That shows there are many places where truncation still is an issue, even though mobile search now can deal with multiline descriptions. I also find that more often than not a short description does a good job as a disambiguator with most subjects being manageable in the recommended 40 and almost all under 60. --Trialpears (talk) 08:52, 31 May 2021 (UTC)

I don't understand what is going on here, we seem to be seriously at cross purposes. On my Android phone (about 60mm wide using Chrome), I don't see 'any' short description under Abraham Lincoln, as shown on the project page. If I do a search, I see the article name and its short description, word wrapping happily into however many lines it needs. Output from {{annotated link}} does the same. So what and where is the problem? Surely this limitation is not being pressed because of a limitation in the Wikipedia app? That a tiny fraction of the readership uses? Please tell me I'm wrong! --John Maynard Friedman (talk) 11:01, 31 May 2021 (UTC)

I think it comes back to the question of what short descriptions are for. This page says they help users identify the desired article, and 40 characters are usually adequate for that. If I type "Tripoli" in the mobile search box I can pick Tripoli "Capital and chief port of Libya"; Tripoli, Lebanon "City in North Governorate; Tripoli, Greece "Place in Greece"; etc. That guides me to the correct article; the data I was looking for (population, foundation date, etc.) can come from there. If we want short descriptions to do something else then they may need to be much longer, but that should be a conscious choice with an agreed target function. Certes (talk) 11:09, 31 May 2021 (UTC)

Should I take it from your non-reply to my question, that my surmise (about the app being the reason) is correct, despite multiple places in the MOS saying that editors should not attempt to design pages to a specific form-factor? Talk about defending the indefensible!

Yes, I understand that one use of short descriptions is as you say, it assists readers who already know broadly what they want to identify the correct article. You and others continue to reiterate this point. I get it. But you are ignoring the other important use: to assist editors to identify articles that may be relevant to their enquiry but they don't already know about it. Take a look at Tragedy of the commons#See also: until changed to use {{annotated link}}, it was a list of unattractive buzz words. Now it is a directory of related concepts that attract readers to explore the topic in more depth. If these are all to be reduced to 40 characters, there is really very little added benefit. So please, think outside the narrow box and accept that the concept has gained wings. --John Maynard Friedman (talk) 15:49, 31 May 2021 (UTC)

It's good to see {{annotated link}} being used more widely. But your example is a list of subject definitions, not short descriptions – many of which you have written yourself in presumably deliberate conflict with the guidance at WP:SDCONTENT. You may disagree with the guidance, but unless and until you have consensus to get it changed, this is simply disruptive editing. You are creating unnecessary work for others who are now left to fix multiple descriptions that comply only with your personal preferences. MichaelMaggs (talk) 21:23, 31 May 2021 (UTC)

I agree with MichaelMaggs that the usage at the pages linked from Tragedy of the commons constitutes a misuse of short descriptions. I see that Jonesey95 changed one such SD to match better with current guidance here ([1]), which now produces

• Stone Soup – European folk story, the inverse of the tragedy

at the See also. This setup sufficiently serves the purpose of annotation in the See also, and the purpose of the SD at the article. — Goszei (talk) 22:43, 31 May 2021 (UTC)

First, when I wrote those informative short descriptions, I was not aware of the 40 character limit. Please WP:AGF. I am also in good faith trying to persuade fellow editors to see that there is a useful role for short descriptions if allowed to expand beyond its current near-trivial (in the mathematical sense) specification. No-one yet has addressed the problem of information discovery that I highlighted, you seem hypnotised by the original use as a disambiguation tool.

The change made by Jonesey95 to Stone Soup has resulted in a short description that is only meaningful when read in the context of the Tragedy of the Commons article. It is entirely meaningless in any other context. What happened to the folk tale? How does that make any sense?

MichaelMaggs has just changed the SD for the White House which (a) has removed its equal status as the office of the President and (b) has invalidated my example above, making it nonsensical.

I am not trying to force my will on anyone, that has never been my style and I'm not about to start. But I would hope that we can reach a reasonable conclusion through logical debate: re-iterating again and again that this is the policy is not an argument from logic. --John Maynard Friedman (talk) 23:04, 1 June 2021 (UTC)

Jonesey95 reverted the long sd for Stone Soup that you had added back to "European folk story". The words "the inverse of the tragedy" appear solely in the See also section of the Tragedy of the commons article. Adding additional text like that, outside the {{annotated link}} template, would be a good way to include any additional explanatory material that makes sense in the context of that specific section. MichaelMaggs (talk) 00:20, 2 June 2021 (UTC)

The purpose of the SD "European folk story" at Stone Soup is to differentiate it, in search results and in other brief listings, from the comic strip, the computer game, the magazine, and the children's book. – Jonesey95 (talk) 00:42, 2 June 2021 (UTC)

I can see that this discussion is going nowhere so I propose to bring it to a close with these remarks:

1. the present limit is not going to change without an RFC.
   1. it is clear from the foregoing discussion that such an RFC will not achieve consensus at this time, so I won't waste everyone's time by opening one.
2. the only reason to retain the present limit that anyone has produced is that the Wikipedia App cannot wrap the Short Description at line end
   1. the App is working as designed, this issue is a design error in the specification
   2. for some mobiles, the 40 character limit is already too long and even compliant SDs are being truncated
   3. the specification for the App is contrary to MOS strictures that editors should disregard form factor and not try to 'design' a page to a particular preferred layout.
   4. the App is used by a tiny minority of visitors and it seems extraordinary that this tail is wagging the dog
3. the present limit significantly reduces the value of SDs for any purpose other than its original idea as a disambiguator. The door has been slammed on any possibility of emergent novel uses
   1. The specific example of emergent use I raised is to improve the value of wp:see also sections using {{annotated link}} to act as a Pathfinder (library science) that will lead readers to discover related concepts. In principle the intent of this can be delivered by appending text to the ultra-terse SD but it creates needless repetitive work and reinventing wheels.
   2. Disambiguation is only useful when the reader is already aware of the topics being distinguished. It is not useful when no such prior knowledge exists.
4. most articles don't have short descriptions
   1. of those that do, the large majority are over 40 characters. It was this fact that led me unwittingly to create new SDs that were longer than the policy limit.
   2. most of the SDs that I added were taken from the offered 'copy' or 'copy and edit' from Wikidata. All were longer (sometimes a great deal longer) than 40 characters.

I have no expectation that this will change the minds of the defenders of the status quo but I hope at least it will crystalise the issues for a future debate because I have no doubt that this question will return in other contexts. --John Maynard Friedman (talk) 10:15, 2 June 2021 (UTC)

Coming late, I agree with this. Johnbod (talk) 19:19, 3 August 2021 (UTC)