Wikidata:Contact the development team/Archive/2015/07
|
This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion. |
Numbers with units
It's been several months since I've heard anything on this - has there been any progress? --Rschen7754 15:33, 21 June 2015 (UTC)
- User interface work is still needed. We had no UI developer for a while. We just got a new one so this will pick up speed again now. --Lydia Pintscher (WMDE) (talk) 13:52, 22 June 2015 (UTC)
- @Lydia Pintscher (WMDE): Do you want to include the problem of the uncertainty at the same time ? Snipre (talk) 18:51, 22 June 2015 (UTC)
- I wanted to but I've realized that this will just delay it even further. So we'll tackle it but probably not in the first version. --Lydia Pintscher (WMDE) (talk) 15:11, 23 June 2015 (UTC)
- @Lydia Pintscher (WMDE): Ok thanks. But it can be good before launching the first version of the Numbers with units datatype to see if it will be a good solution to already include in the code fields to store the uncertainty. Snipre (talk) 07:37, 27 June 2015 (UTC)
- Yeah You're right. Definitely need to keep that in mind. --Lydia Pintscher (WMDE) (talk) 19:59, 4 July 2015 (UTC)
- @Lydia Pintscher (WMDE): Ok thanks. But it can be good before launching the first version of the Numbers with units datatype to see if it will be a good solution to already include in the code fields to store the uncertainty. Snipre (talk) 07:37, 27 June 2015 (UTC)
- I wanted to but I've realized that this will just delay it even further. So we'll tackle it but probably not in the first version. --Lydia Pintscher (WMDE) (talk) 15:11, 23 June 2015 (UTC)
- @Lydia Pintscher (WMDE): Do you want to include the problem of the uncertainty at the same time ? Snipre (talk) 18:51, 22 June 2015 (UTC)
Please can we have automatic labels and descriptions based on transcriptions and statements
There are a lot of languages which do not have descriptions yet. I did a search for "Peri" and it threw up a load of items with no descriptions in English, many with no labels in English. As I went through these adding labels and descriptions I noticed some patterns:
Lots of places and people in Italy, Indonesia, Estonia, Romania and Turkey named Peri. (plus a lot of wikinews articles in Serbian about Katy Perry or Kejti Peri as they say in Serbian). In pretty much every case the best label in English is the label in the language where the original wikipedia article is (most of these are in one language only). There is a concept "Peri" but it is already linked to "Fairy" and isn't a problem. I didn't find one item (so far - I'm only half way through and then there all the "La Peri" items) where you couldn't just reuse the label in the original language. While some could be translated even those cases tended to be proper names(Peri Photography Centre in Turku, Finland) so they could just as well stay untranslated.
- Please modify the search function to show the label in another language if there is no label in English.
- Please modify the item interface to show the label in another language (or labels in other languages) as a proposed label with a simple one click to accept workflow.
For labels I have not found a single case where the label could not be easily rendered from a few basic statements. Instance of and Country; Occupation and Country of Citizenship etc. By adding these basic statements I could be adding labels in a hundred languages instead of just in one.
- Please replace the descriptions with descriptions which are automatically generated from statements
- Where statements are missing change the wikipedia mobile interface to invite users to add basic statements.
This will needs some thought as to the workflow -
- Can we have a way for wikidatans to develop and edit the flow diagram to guide wikipedia mobile users and respond to their answers?
- Filceolaire (talk) 23:07, 22 June 2015 (UTC)
- We maybe also have to agree about how we in the best way choose good labels and descriptions to Wikisource-items. Items like Fredman's Epistles (Q20203267) and Fredman's Epistles (Q5383621) tend to have very simlair labels and descriptions. Different descriptions makes its easier to search, but it does not make it easy to manually read a statement. Imagine in has edition or translation (P747) in Bible (Q1845) 400 claims, having the same label: "The Bible". -- Innocent bystander (talk) 07:04, 23 June 2015 (UTC)
- The search function already uses your fallback languages (babel) in the search function now. Sjoerd de Bruin (talk) 07:23, 23 June 2015 (UTC)
- Does babel:de-0/fr-0/ru-0/ja-0/zh-0 add more fallback-languages? -- Innocent bystander (talk) 08:21, 23 June 2015 (UTC)
- Depends on what you see here, try it. Sjoerd de Bruin (talk) 10:54, 23 June 2015 (UTC)
- Ok, it worked (at least in that page). In real life, I have seen content in fi and fit, even if that wasn't on that page. - Innocent bystander (talk) 11:10, 23 June 2015 (UTC)
- My language is English then pt, fr, de, it, es, nl as babel languages but I don't see labels from these in search. I would love to see fallback labels in these languages but I would also like to see fallback labels in any other language if these were missing. Filceolaire (talk) 17:48, 23 June 2015 (UTC)
- As Sjoerddebruin said language fallbacks are already happening in many many places. You just can't see them because English is the last in all fallback chains. I am trying to find ways to improve this but it is not easy. We can't do fallbacks that are specific to a user for example as that kills caching. Doing specialized fallbacks for certain topics is also something I don't want. --Lydia Pintscher (WMDE) (talk) 13:36, 1 July 2015 (UTC)
- Thanks Lydia. As I understand it you don't want the fallback to be customised based on the users details or the items details as that will slow stuff down. A possibility worth looking at might be that if there is only text in one language then you show that text together with a romanisation as the fallback for everyone. Cases where there is text in only one language make up a lot of the cases where there is no English text. Filceolaire (talk) 18:46, 1 July 2015 (UTC)
- That would be an option and we have in fact been thinking about this. The reason we decided against it back then was that it'd be very confusing why a label suddenly vanishes when someone adds a second label to an item that previously showed a fallback. What do others think? --Lydia Pintscher (WMDE) (talk) 20:01, 4 July 2015 (UTC)
- Thanks Lydia. As I understand it you don't want the fallback to be customised based on the users details or the items details as that will slow stuff down. A possibility worth looking at might be that if there is only text in one language then you show that text together with a romanisation as the fallback for everyone. Cases where there is text in only one language make up a lot of the cases where there is no English text. Filceolaire (talk) 18:46, 1 July 2015 (UTC)
- Ok, it worked (at least in that page). In real life, I have seen content in fi and fit, even if that wasn't on that page. - Innocent bystander (talk) 11:10, 23 June 2015 (UTC)
- Depends on what you see here, try it. Sjoerd de Bruin (talk) 10:54, 23 June 2015 (UTC)
- Does babel:de-0/fr-0/ru-0/ja-0/zh-0 add more fallback-languages? -- Innocent bystander (talk) 08:21, 23 June 2015 (UTC)
- The search function already uses your fallback languages (babel) in the search function now. Sjoerd de Bruin (talk) 07:23, 23 June 2015 (UTC)
- We maybe also have to agree about how we in the best way choose good labels and descriptions to Wikisource-items. Items like Fredman's Epistles (Q20203267) and Fredman's Epistles (Q5383621) tend to have very simlair labels and descriptions. Different descriptions makes its easier to search, but it does not make it easy to manually read a statement. Imagine in has edition or translation (P747) in Bible (Q1845) 400 claims, having the same label: "The Bible". -- Innocent bystander (talk) 07:04, 23 June 2015 (UTC)
- multiple items with similar labels and statements in one language are something that autolabels can detect and address. Editions of the bible can be distinguished by their date of publication for instance. All those Wikinews articles in Serbian about Katy Perry (there are about a dozen of them) could be distinguished by their publication date too if that statement was added and the label got auto updated when the statement was added. Q20203267 could get the label "Fredmans epistlar " from the Swedish label with the description <edition or translation of (P629):Fredman's Epistles (Q5383621), language of work or name (P407):Swedish (Q9027)> from the statements. My guess is that auto descriptions are easy for 95% of all items and for 99% of the items actually accessed. As the algorithm for labels is improved these percentages will I hope get even better. Filceolaire (talk) 17:48, 23 June 2015 (UTC)
- I am not sure the "publication date" is good to add in all Wikisource-items, since the pages on Wikisource sometimes is regarded as a work separate from the one published on paper. Should a new date then be added every time somebody edit the page, or should we choose the first edit of the page? And often is there no corresponding WP-items for the Wikisource-items, and the WS-items then carry more information than it should. :) -- Innocent bystander (talk) 18:35, 23 June 2015 (UTC)
- Note that I didn't suggest adding a publication date to the wikisource translation example above - I suggested the description read <translation of:original name><language:Language of the translation>. That should work 99% of the time I think. Filceolaire (talk) 02:08, 27 June 2015 (UTC)
- I am not sure the "publication date" is good to add in all Wikisource-items, since the pages on Wikisource sometimes is regarded as a work separate from the one published on paper. Should a new date then be added every time somebody edit the page, or should we choose the first edit of the page? And often is there no corresponding WP-items for the Wikisource-items, and the WS-items then carry more information than it should. :) -- Innocent bystander (talk) 18:35, 23 June 2015 (UTC)
- multiple items with similar labels and statements in one language are something that autolabels can detect and address. Editions of the bible can be distinguished by their date of publication for instance. All those Wikinews articles in Serbian about Katy Perry (there are about a dozen of them) could be distinguished by their publication date too if that statement was added and the label got auto updated when the statement was added. Q20203267 could get the label "Fredmans epistlar " from the Swedish label with the description <edition or translation of (P629):Fredman's Epistles (Q5383621), language of work or name (P407):Swedish (Q9027)> from the statements. My guess is that auto descriptions are easy for 95% of all items and for 99% of the items actually accessed. As the algorithm for labels is improved these percentages will I hope get even better. Filceolaire (talk) 17:48, 23 June 2015 (UTC)
Replacing statements
I have today replaced some statements, that were wrong. (They were user-mistakes imported from Wikipedia, why I did not marked the old statements as deprecated.) I had some severe troubles with this in IE11/W7. It demanded several reloads of the page to be able to edit the page. -- Innocent bystander (talk) 10:40, 1 July 2015 (UTC)
- What sort of trouble were you having? - Nikki (talk) 12:36, 1 July 2015 (UTC)
- The browser stalled and the statements did not change as intended. -- Innocent bystander (talk) 13:49, 1 July 2015 (UTC)
- There were some issues we fixed very recently where removing the last reference of a statement could cause this. The fix will be here in a few days. I hope that fixes your issues. If you still see it next week please let me know. --Lydia Pintscher (WMDE) (talk) 13:54, 1 July 2015 (UTC)
- The browser stalled and the statements did not change as intended. -- Innocent bystander (talk) 13:49, 1 July 2015 (UTC)
Concept URI
There is a link to "Concept URI" to the left, next to every item. Is there anybody who can help me get a link to a xml-version of that page instead? -- Innocent bystander (talk) 13:58, 4 July 2015 (UTC)
- For the supported formats (JSON and various RDF formats), you can just add the file extension to the URL it links to (e.g. ".json" to get JSON), but XML is no longer supported there (see phab:T93353 and phab:T99560) so you can't get XML from that any more. To get XML you would probably need to use the MediaWiki API instead. - Nikki (talk) 15:01, 4 July 2015 (UTC)
- Yes, can anybody help me get such an API-link somewhere on the page? You may edit my own .js or .css-pages if you like! -- Innocent bystander (talk) 16:06, 4 July 2015 (UTC)
- We're trying to find a way to improve the whole thing and make it usable. Comments/ideas/patches welcome at phabricator:T102155. --Lydia Pintscher (WMDE) (talk) 20:08, 4 July 2015 (UTC)
- Yes, can anybody help me get such an API-link somewhere on the page? You may edit my own .js or .css-pages if you like! -- Innocent bystander (talk) 16:06, 4 July 2015 (UTC)
Century dating—localization
Please see WD:Project chat#Century dating—localization.
In short: in some cases, certain dating (for date of birth, date of death, what have you) is only available at the century level (e.g., "12th century"). At the moment, when a Wikipedia uses {{#property: ...}} to pull information like that, the result comes out as "12. century", which is a bit of a strange hybrid of German and English forms. Other wikis might like this to come out in other forms:
- lawiki: "saeculum XII" (would be fine, though ablative "saeculo XII" would be even better)
- frwiki: "XIIe siècle"
Is this possible, either natively or at translatewiki? After all, if the date of involved is, say, 15 January 1208, it correctly comes out on lawiki as 15 Ianuarii 1208.
Please note: at least at lawiki, Lua has not been implemented, and it's not clear whether it can be any time soon. (Small wiki, limited numbers of people able to help on it, etc.) We'd far prefer an approach native to the interface, if possible. Thanks. StevenJ81 (talk) 18:27, 26 June 2015 (UTC)
- Can you please clarify something for me? When you look at the date here on Wikidata it is correct for you? When you display it on a Wikipedia it is shown in which language? Or is it just the formatting that is wrong? --Lydia Pintscher (WMDE) (talk) 13:20, 1 July 2015 (UTC)
- Formatting is odd, sample at Property_talk:P569#Date_output. --- Jura 13:29, 1 July 2015 (UTC)
- There was some other oddity with the input form discussed in the misnamed Project chat thread:
- an input of "15. century" results in 1500-01-01 +/- 100 yr (Special:Diff/145618879. Q2161498 seems better. --- Jura 13:36, 1 July 2015 (UTC)
- Great. Thank you. I opened phabricator:T104447. --Lydia Pintscher (WMDE) (talk) 13:51, 1 July 2015 (UTC)
- That ticket is only about the display of the date on Wikipedia (etc) when using data from Wikidata and not about the display on Wikidata itself, right? From what I can tell, Wikipedia, Wikisource, etc, are displaying the date in exactly the same way as Wikidata itself, which in all the cases I've tried is wrong. I added a date to Wikidata Sandbox (Q4115189) and tried changing the interface language on Wikidata and previewing {{#property:P571|from=Q4115189}} on several Wikisources. The result:
- Wikidata in English: "15. century" - wrong, should be "15th century" (English doesn't use a dot for ordinal numbers)
- English Wikisource in English: "15. century", as above
- Wikidata in German: "15 century" - wrong, should be "15. Jahrhundert" (German does use a dot for ordinal numbers and the text isn't translated)
- German Wikisource: "15 century", as above
- Wikidata in Latin: "15. century" - wrong, should be "saeculum XV" or at least "saeculum 15" (Latin doesn't use a dot for ordinal numbers, the text isn't translated and the number is in the wrong position)
- (or "saeculo XV" if the ablative case is a possibility, but at a lower level of importance) StevenJ81 (talk) 20:41, 1 July 2015 (UTC)
- Latin Wikisource: "15. century", as above
- Wikidata in Japanese: "15. century" - wrong, should be "15世紀" (Japanese doesn't use a dot for ordinal numbers, the text isn't translated and there shouldn't be a space between the number and word)
- Japanese Wikisource: "15. century", as above
- - Nikki (talk) 14:34, 1 July 2015 (UTC)
- That ticket is only about the display of the date on Wikipedia (etc) when using data from Wikidata and not about the display on Wikidata itself, right? From what I can tell, Wikipedia, Wikisource, etc, are displaying the date in exactly the same way as Wikidata itself, which in all the cases I've tried is wrong. I added a date to Wikidata Sandbox (Q4115189) and tried changing the interface language on Wikidata and previewing {{#property:P571|from=Q4115189}} on several Wikisources. The result:
- Great. Thank you. I opened phabricator:T104447. --Lydia Pintscher (WMDE) (talk) 13:51, 1 July 2015 (UTC)
- An option to generate an ISO looking output would be nice too. --- Jura 15:04, 1 July 2015 (UTC)
- Just added to hold the auto archive for a time. StevenJ81 (talk) 15:46, 9 July 2015 (UTC)
- An option to generate an ISO looking output would be nice too. --- Jura 15:04, 1 July 2015 (UTC)
collection of usecases
I´ve been adding dates to lots of items simply by c&p from various sources in different languages. My defalult language is set to "Deutsch". The following table gives an overview, of some values I´ve entered and the way wikidata reacts. V= accepted; X = rejected. For sure some values that are absolutely according to the rules in German are not accepted, while others that don´t in fact are accepted. So you might have a look and see if this is helpful
| Datumseingabe | |
|---|---|
| 1. jan. 2010 | v |
| 1. jan 2012 | v |
| 1 januar 2010 | x |
| 1 january 2010 | v |
| 1. Jänner 2010 | x |
| 1 feb 2010 | v |
| 1 februar 2010 | x |
| 1. febr. 2010 | x |
| 1. Febr. 2010 | x |
| 1. Feb. 2010 | v |
| 1. feb. 2010 | x |
| 1 febbraio 2010 | x |
| 1 february2010 | v |
| 10 märz 1964 | x |
| 24 september 1937 | v |
| 25 dezember 1910 | x |
| 25 dicembre 1910 | x |
| 25 Dez 1910 | x |
| 25. Dez 1910 | x |
| 25. Dez. 1910 | v |
| 20. century | v |
| 20 century | x |
| 20. Jahrhundert | x |
- Thank you! That helps. --Lydia Pintscher (WMDE) (talk) 20:02, 4 July 2015 (UTC)
Item number in list of aliases
I noticed that items display own number in list of aliases when added as property value. For example, for Illarion Golenishchev-Kutuzov (Q4141432). --EugeneZelenko (talk) 22:46, 5 July 2015 (UTC)
- I think that is phab:T104273. - Nikki (talk) 23:39, 5 July 2015 (UTC)
Parse property talk (test)
To run https://test.wikidata.org/wiki/Special:ConstraintReport would you parse the property talk pages? --- Jura 15:58, 6 July 2015 (UTC)
- Lol .. nevermind. --- Jura 17:13, 6 July 2015 (UTC)
Actually, could you do it here? We changed a few constraints that generated errors on a large number of items. --- Jura 06:43, 8 July 2015 (UTC)
- According to the documentation, the property talk pages get parsed once a week. --Pasleim (talk) 16:02, 9 July 2015 (UTC)
- Isn't it a script that could also be triggered manually? --- Jura 16:04, 9 July 2015 (UTC)
- Yes. I'll poke Marius to see if he can run it. Might take a while as people are traveling for Wikimania and things. --Lydia Pintscher (WMDE) (talk) 22:52, 9 July 2015 (UTC)
Done I've updated the table with data acquired around 00:00UTC. Cheers, Hoo man (talk) 00:22, 10 July 2015 (UTC)
- Thanks. Looks much better now that all the population numbers don't appear in red. --- Jura 06:02, 10 July 2015 (UTC)
- Yes. I'll poke Marius to see if he can run it. Might take a while as people are traveling for Wikimania and things. --Lydia Pintscher (WMDE) (talk) 22:52, 9 July 2015 (UTC)
- Isn't it a script that could also be triggered manually? --- Jura 16:04, 9 July 2015 (UTC)
Strange behaviour on watchlist
I noticed a strange behaviour on my watchlist (noticed it before, but didn't copy the lines). Jura1 made 2 edits to Tom Holland, with a difference of 0 bytes according to my wachtlist when collapsed:
18:08 Tom Holland (Q2023710) (2 bewerkingen | geschiedenis) . . (0) . . [Jura1 (2×)]
When I uncollapse them, I see 2 edits which count up for 825 added bytes:
18:08 (huidig | vorige) . . (+405) . . Jura1 (Overleg | bijdragen | blokkeren) (Claim aangemaakt: vader (P22): Dominic Holland (Q5290554)) (Label: Widar [1.3]) 18:08 (huidig | vorige) . . (+425) . . Jura1 (Overleg | bijdragen | blokkeren) (Referentie toegevoegd aan claim: vader (P22): Dominic Holland (Q5290554)) [terugdraaien] (Label: Widar [1.3])
That's not all that's strange. I then click the text "2 bewerkingen", which normally would result in a diff of the two edits. It only shows a diff for the first edit, which also shows that nothing has been changed (how can the change than be +405 bytes?) and there is 1 more edit. This is the diff I get: https://www.wikidata.org/w/index.php?title=Q2023710&curid=1951681&diff=227946182&oldid=227946182 When I check the history it shows 2 edits by Jura1. What's going on here? Mbch331 (talk) 16:52, 9 July 2015 (UTC)
- Strange indeed. Thanks for reporting. Will look into it but it'll take a while because of traveling. --Lydia Pintscher (WMDE) (talk) 22:54, 9 July 2015 (UTC)
- Noticed it again today on my watchlist. A bot did two edits on Ramses Shaffy (Q1343144). The text "2 bewerkingen" which should lead to a diff of both edits, only lead to the first edit and the sum is 0, while manually adding the bytes altered gives a totally different sum. Mbch331 (talk) 06:13, 16 July 2015 (UTC)
- I think this happens when an already existing claim is added through the API in stead of the UI. Nothing gets added, because the claim already exists (which shows as an empty diff for that edit), but because of API use the watchlist still thinks a claim was added. And then the watchlist doesn't understand it anymore and shows the strange behaviour. Mbch331 (talk) 06:23, 16 July 2015 (UTC)
- Noticed it again today on my watchlist. A bot did two edits on Ramses Shaffy (Q1343144). The text "2 bewerkingen" which should lead to a diff of both edits, only lead to the first edit and the sum is 0, while manually adding the bytes altered gives a totally different sum. Mbch331 (talk) 06:13, 16 July 2015 (UTC)
reference URL (P854) is useful to add references ..
Is there a way to have it suggested as a qualifier? Usually only "stated in" and "retrieved" get suggested. --- Jura 06:02, 10 July 2015 (UTC)
- That looks like phab:T77972. - Nikki (talk) 08:14, 10 July 2015 (UTC)
- It does. Apparently it's a low priority .. --- Jura 08:38, 10 July 2015 (UTC)
- Ah, nice to see a task of me here. Would love to have a default set of properties for references. Sjoerd de Bruin (talk) 12:35, 16 July 2015 (UTC)
- @Sjoerddebruin: The default sets exists: see help:sources. Snipre (talk) 20:53, 19 July 2015 (UTC)
- Ah, nice to see a task of me here. Would love to have a default set of properties for references. Sjoerd de Bruin (talk) 12:35, 16 July 2015 (UTC)
- It does. Apparently it's a low priority .. --- Jura 08:38, 10 July 2015 (UTC)
Translation extension
I´ve tried to save a translation, but i´ve got errord likr "Das Speichern der Übersetzung ist fehlgeschlagen: [c4d3208a] Database query error". May be someone can explain what to do.--Giftzwerg 88 (talk) 09:38, 20 July 2015 (UTC)
- Shouldn't you rather contact the developers of the extension in Phabricator or their IRC if it exists? What page did the error occur on? Matěj Suchánek (talk) 13:45, 20 July 2015 (UTC)
property aliases
When Lydia proposed that aliases for properties should be unique I objected that having non-unique aliases is useful because it is a way to make sure the property suggester suggests all similar properties so users can pick the exact one they need. Thinking about it there is an alternative that would work better.
Make the property suggester show any 'sub-properties' of a property, as well as the property whose alias has just been typed in.
Sub properties are identified by the "subproperty of (P1647)" property on the property.
Doing this would work better than my hack with aliases because it would work in all languages. Filceolaire (talk) 13:34, 21 July 2015 (UTC)
- The entity suggester already suggests "father" (P22) if you type "child". It doesn't matter that "child of" is unique. --- Jura 13:41, 21 July 2015 (UTC)
New edit links on Special:WhatLinksHere/Q5
Somehow these new links on the above page don't work .. and I suppose wont ever work: https://www.wikidata.org/w/index.php?title=Q23&action=edit
Is there a way to not show them? --- Jura 11:15, 23 July 2015 (UTC)
- I have opened an issue on phabricator. ·addshore· talk to me! 11:48, 23 July 2015 (UTC)
- A patch is already up and being merged :) ·addshore· talk to me! 12:36, 23 July 2015 (UTC)
- Amazing .. lightning speed. Thanks! --- Jura 12:57, 23 July 2015 (UTC)
- A patch is already up and being merged :) ·addshore· talk to me! 12:36, 23 July 2015 (UTC)
Ymer
Not sure this belongs here, but can't find a better place for it... Item Q10723584 links (correctly) to pages on the da, sv, and en WPs. Curiously, the da article displays (red) interwikilinks to the nl, fr, and de WPs, although there are as far as I can see no interwikilinks in that article nor in the wikidata item. Any idea what is going on here? Thanks and apologies if this is not the right place for this. --Randykitty (talk) 10:56, 24 July 2015 (UTC)
- Are you using the "Indholdsoversættelse" beta feature? Sjoerd de Bruin (talk) 10:59, 24 July 2015 (UTC)
- Yes, and that is indeed the problem. I switched it off and the incorrect redlinks are gone. (BTW, the fr link was to "français", lower case). Weird. --Randykitty (talk) 11:35, 24 July 2015 (UTC)
- Looks like lower case: français is standard i French. It Is English Which Uses Upper Case Letters For Languages: French
- -- Innocent bystander (talk) 12:32, 24 July 2015 (UTC)
- It does, but other interwikilinks (on the en and da WPs) display "Français". So it looked to me like an interwikilink that had been added by hand, but I couldn't find it in the source code of the article. --Randykitty (talk) 12:56, 24 July 2015 (UTC)
- Yes, and that is indeed the problem. I switched it off and the incorrect redlinks are gone. (BTW, the fr link was to "français", lower case). Weird. --Randykitty (talk) 11:35, 24 July 2015 (UTC)
"Select an article to preview" buttons have disappeared
The buttons to select a language article to preview have disappeared from my UI. I have the 'move' tool enabled and the buttons for that are still visible - usually these are next to the select buttons. Joe Filceolaire (talk) 09:55, 25 July 2015 (UTC)
- For me the have disappeared as well. Usually these kind of things return automatically. Mbch331 (talk) 10:04, 25 July 2015 (UTC)
- The URL of the image was not working anymore, I hope I replaced it with the good image. Sjoerd de Bruin (talk) 10:16, 25 July 2015 (UTC)
- fixed. Joe Filceolaire (talk) 19:28, 25 July 2015 (UTC)
- The URL of the image was not working anymore, I hope I replaced it with the good image. Sjoerd de Bruin (talk) 10:16, 25 July 2015 (UTC)
The problem described in phabricator:T59529 is back. --Succu (talk) 09:00, 30 July 2015 (UTC)
- Thanks. I've reopened the ticket. In such cases you can also reopen the ticket yourself if you want :) --Lydia Pintscher (WMDE) (talk) 18:58, 30 July 2015 (UTC)
- This should be fixed now. Aude (talk) 17:17, 31 July 2015 (UTC)
Constraint violation that isn't a violation
I checked the constraints for I See Fire (Q15272117) (See Special:ConstraintReport/Q15272117) and got 1 violation. nominated for (P1411)=Satellite Award for Best Original Song (Q919999) isn't valid, because Q919999 isn't a award (Q618779). However Q919999 is a Satellite Awards (Q40024) which is a subclass of Q618779. So in my point of view, that would still be a valid claim and the report is wrong. Mbch331 (talk) 14:36, 30 July 2015 (UTC)
- Satellite Award for Best Original Song (Q919999) is a instance of subclass of award (Q618779), however, the current constraint on Property talk:P1411 requires that Satellite Award for Best Original Song (Q919999) is a subclass of (subclass of) award (Q618779).--Pasleim (talk) 17:48, 30 July 2015 (UTC)
- My bad. Didn't read the constraints well enough. Mbch331 (talk) 18:44, 30 July 2015 (UTC)
- Changed Satellite Award for Best Original Song (Q919999) to subclass and problem is solved. Mbch331 (talk) 18:47, 30 July 2015 (UTC)