Wikidata

Wikidata:Contact the development team/Archive/2016/03

< Wikidata:Contact the development team‎ | Archive‎ | 2016
This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

url datatype: protocol for phone number

Property:P1244 relies on tel: being available. Apparently, it's a simple configuration change to add this. Would you kindly do that?
--- Jura 03:38, 4 March 2016 (UTC)

I'm really not convinced that a URL type is the best data type for phone numbers. We don't have to use URL data types for everything that can be turned into a URL (e.g. we could store identifiers as URLs, but don't) and I also think "tel:" URLs are not very user friendly. I don't expect many people to know or expect that they have to prefix telephone numbers with "tel:" and I don't expect people to know the right format for "tel:" URLs either (20% of the uses for phone number (P1329) violate that property's existing format constraint and over 75% of all the uses contain spaces, which are explicitly not allowed in "tel:" URLs). People using the data are also very likely to want to display the number, not only create a link, and it's easier to turn the display form of a phone number into a URL than vice versa (enwikivoyage even has a module to do that already). - Nikki (talk) 18:07, 4 March 2016 (UTC)
I tend to agree with most of what you wrote, but came to a different conclusion:
If people can't get the format right, using a url format would allow them to get guidance on how to format it correctly. It also helps us make sure essential elements are not missing.
I'm not sure if phone numbers are entered by "users" in general. Some I fixed were entered by what seem to be professional promoters. I think we can expect them to follow quite quickly.
The data can still be displayed as a number, but obviously, all the malformed ones would be eliminated.
Given that we announced that in advance, I think we should make good on our announcement.
--- Jura 05:18, 5 March 2016 (UTC)
If you can come to some kind of agreement on this we can add the protocol. Maybe a post on project chat? --Lydia Pintscher (WMDE) (talk) 10:59, 6 March 2016 (UTC)

SPARQL graphs

To check if it's ready for other sites, would it be possible to activate wikidatasparql: at Wikidata?

This way we can make sure that mw:Extension:Graph/Demo/Sparql works while the other issues mentioned in phab:T126741 are being sorted out.
--- Jura 10:41, 4 March 2016 (UTC)

I think the performance and stability issue are so severe that we shouldn't even do this on Wikidata (or test.wikidata for that matter). This wiki is still a production site that users expect to work reliably and fast. Cheers, Hoo man (talk) 11:11, 4 March 2016 (UTC)
Obviously we wouldn't want it then. Is it an issue just for the page or the site in general? I think we might do a few hundreds of charts, but nothing near the potential use at Wikipedia.
--- Jura 11:32, 4 March 2016 (UTC)
If it just runs a SPARQL query once and then stores the picture, I don't see how this could be an issue, see https://phabricator.wikimedia.org/T126741#2051994
--- Jura 12:15, 4 March 2016 (UTC)
At this point the risk is too high I fear. I am trying to figure out the way forward though for stored queries that can be used on Wikipedia in general given all the pieces we have now. --Lydia Pintscher (WMDE) (talk) 11:01, 6 March 2016 (UTC)

Date formatting

Tracked in Phabricator
Task T127820

Hello. We are experimenting with automatic retrieval of dates of birth and death over at Norwegian Nynorsk Wikipedia (nn). This works fine except for one thing: When getting the values from Wikidata, the dates are not formatted properly for our language. They should have the format "j. F Y", but end up as "j F Y", without the period. I've found that imported dates on Norwegian Bokmål Wikipedia (no:, although I think their language code is nb) includes the period (they also use the abbreviated month name). So it seems the date format is customizable somewhere. The question is, where? If anyone could point me in the right direction, or simply fix it for us, that would be great! --Ranveig (talk) 10:02, 25 February 2016 (UTC)

@Ranveig: try using nn:Modul:Wikidata, where you can specify format at the "datetime" part. But maybe there is another solution. --Edgars2007 (talk) 10:36, 25 February 2016 (UTC)
I was about to say I'd tried that and it didn't work, but it seems there was one instance that I missed. Thanks for the extra push! Would still love the basic problem to be sorted out, though. --Ranveig (talk) 11:31, 25 February 2016 (UTC)
Might I also add that the reason I'd like to use the value from Wikidata directly is that that gives the actual date as it appears on Wikidata. In cases where only the year is given, the module gives either January 1, as in w:nn:Israel Ruong, or November 30 of the previous year, as for w:nn:Ingrid Geithus (this was how the problem was discovered). So far I haven't been able to find a solution that gives us correct dates in all instances. --Ranveig (talk) 11:41, 25 February 2016 (UTC)
Tracked in Phabricator
Task T103378
Oh, right. There should be a note, that people shouldn't use Lua for dates, as there we're some problems with them. Will try to find the phab ticket about it. --Edgars2007 (talk) 12:18, 25 February 2016 (UTC)
@Ranveig, Edgars2007: You should in LUA be able to adapt how the module displays the date depending on the precision. "If precision == date -> format = 'j. F Y', elseif precision == month -> format = 'F Y' etc etc". -- Innocent bystander (talk) 08:07, 29 February 2016 (UTC)
Thank you all for your assistance. In the end I had to get my friendly neighbourhood hacker to, erm, hack the module so that it stopped displaying made-up dates and the occasional wrong year. It seems that when only a year is given, sometimes the day and month are stored as 00-00 (which is then converted to November 30 and the previous year) and sometimes as 01-01. So in our module, the wrong years are identified and a 1 added. When/if this is fixed, we'll presumably have to unhack it.
To go back to my original question - I'd still be very grateful if someone could figure out where in Mediawiki it says that Wikidate dates in Nynorsk should be expressed without the period when they have graciously given to Bokmål. It's clearly possible to display it correctly. I just have no idea where this could be fixed. --Ranveig (talk) 13:06, 7 March 2016 (UTC)

Visual Editor feedback

I was trying to edit Wikidata:Property_proposal/Generic with the VE but it gives some errors. I used the feedback tool integrated in VE to report the errors, and it notified me that it had posted my message on VisualEditor/Feedback, however that page doesn't exist. Where did my message go? --Micru (talk) 18:34, 6 March 2016 (UTC)

Here mw:VisualEditor/Feedback. Matěj Suchánek (talk) 18:36, 6 March 2016 (UTC)
Yeah, here. Sjoerd de Bruin (talk) 18:38, 6 March 2016 (UTC)
Also relevant: phab:T125400. Sjoerd de Bruin (talk) 18:41, 6 March 2016 (UTC)
Thanks for the pointers! It was really confusing that it just disappeared...--Micru (talk) 22:13, 6 March 2016 (UTC)

Bug T115792

Is there any progress at Phabricator:T115792? Jeblad (talk) 08:02, 7 March 2016 (UTC)

A possible workaround could be to offer a link that explicitly sets "no": https://www.wikidata.org/wiki/Special:NewProperty?lang=no&uselang=nb
--- Jura 08:49, 7 March 2016 (UTC)
@Jeblad: try https://test.wikidata.org/wiki/MediaWiki:Wikibase-newitem-summary/nb at MediaWiki:Wikibase-newitem-summary/nb.
--- Jura 17:17, 7 March 2016 (UTC)
Adrian said he'll look into it. --Lydia Pintscher (WMDE) (talk) 09:25, 9 March 2016 (UTC)

Automatically add the label service in query service

Tracked in Phabricator
Task T129100

Something easy and convenient for the query service could be to add an option to automatically insert the code for the labelling translation service, with a button for example. author  TomT0m / talk page 11:41, 7 March 2016 (UTC)

Created phab:T129100 for this. Mbch331 (talk) 17:37, 7 March 2016 (UTC)
You could use the auto complete feature. Just type 'ser' and press ctrl + space. --Jonas Kress (WMDE) (talk) 19:17, 9 March 2016 (UTC)

SPARQL request returns item deleted last December

Couldn't figure how this request manages to return deleted item Q21814419. -- LaddΩ chat ;) 02:49, 9 March 2016 (UTC)

@Smalyshev (WMF): Something for you I believe. --Lydia Pintscher (WMDE) (talk) 09:26, 9 March 2016 (UTC)
  Resolved --Smalyshev (WMF) (talk) 22:50, 9 March 2016 (UTC)

Unable to delete alias

On Ertuğrul (Q243245), there is a Turkish alias "Ertuğrul Gazi" which I believe to be erroneous and caused by a charset encoding issue. Unfortunately I cannot delete it. When I try to delete it, it disappears, but my change is not recorded in the page history, and the entry reappears when I refresh the page. Bovlb (talk) 22:21, 8 March 2016 (UTC)

I solved it by using the API to force the aliasses for tr (wbsetaliases). (Removing the alias with the API wasn't possible either.) Mbch331 (talk) 22:33, 8 March 2016 (UTC)
Thanks. That solves my immediate problem, but there seems to be a UI/API bug here. Alternatively, if this entry was invalid UTF-8 introduced in an earlier codebase, maybe we should sweep for other examples. Cheers, Bovlb (talk) 16:09, 9 March 2016 (UTC)
Maybe Wikidata:Bistro#Help_.21 (reported by Ludo29) and Wikidata:Project_chat#Cannot_remove_an_entry (reports by Vachovec1,Hunnjazal,Milicevic01) are about the same problem. Is this for AKlapper to sort?
--- Jura 11:45, 10 March 2016 (UTC)
Nikki added it to phab.
--- Jura 11:58, 10 March 2016 (UTC)
That phabricator ticket is about something else. Mbch331 (talk) 12:23, 10 March 2016 (UTC)

Wikidata Query Service map display

Thanks a lot for this new feature. It's really great! The only issue I have is the size and shape of the marker. When you have many results it makes the map illegible (see the map of ski resorts for example). Could you change it to a small red dot please? Or add an option to chose another marker? Thanks. — Ayack (talk) 17:21, 9 March 2016 (UTC)

This is done just wait for the next deployment --Jonas Kress (WMDE) (talk) 19:16, 9 March 2016 (UTC)

Great! Thank you. — Ayack (talk) 19:48, 9 March 2016 (UTC)
  Done --Smalyshev (WMF) (talk) 22:45, 9 March 2016 (UTC)

Issues with adding or removing qualifiers

New issues that started today:

  • Button "remove" below a qualifier (while editing a statement) suppresses the entire statement, see this test on sandbox;
  • After creating a new statement, a new button "+ add qualifier" appears when hovering the statement; however, when using that button to add a qualifier, the edit box to input the value is on top of the "save" button and... there is no way to save that qualifier value;
  • After creating a new statement and purging the cache of the page, hovering the statement no longer display the new button "+ add qualifier". Statement needs to be edited before "add qualifier" gets displayed.

LaddΩ chat ;) 13:16, 11 March 2016 (UTC)

Thanks for the report. I can't reproduce the first issue. Can anyone else? The second and the third issue are related and the tickdet for it is at phabricator:T128317. The add qualifier button shouldn't be there. --Lydia Pintscher (WMDE) (talk) 09:43, 16 March 2016 (UTC)
@Lydia Pintscher (WMDE): I just figured what I do wrong: the "remove" link under the qualifier is for removing the entire statement... The "remove (qualifier)" link is at the extreme right side of the screen, on the same line as the qualifier. Poor UI design, IMHO, but it's not a bug, sorry about this. -- LaddΩ chat ;) 13:34, 16 March 2016 (UTC)
No problem. UI issues like that are also important for me to know. --Lydia Pintscher (WMDE) (talk) 20:43, 16 March 2016 (UTC)

saving information about qualifiers

is only possible by clicking edit for a second time because the save button appears only then and is not visible to save immediately. This is a new bug for some days now. Perhaps you can fix it, because it is one unnecessary click more, which was not before--Oursana (talk) 03:50, 12 March 2016 (UTC)

I am sorry. I've been trying to understand what the issue is and reproduce it. I can't find the issue you are having. Can you explain it more please? --Lydia Pintscher (WMDE) (talk) 09:46, 16 March 2016 (UTC)
Ahhhh now I understand. It is the same problem as in the topic right above. Jonas is looking at it. --Lydia Pintscher (WMDE) (talk) 11:13, 16 March 2016 (UTC)

Special:Search/Results doesn't find Results (P2501)

output
Search results

To search for Wikidata items by their title on a given site, use Special:ItemByTitle. To search by label in a given language, use Special:ItemDisambiguation.
Results 1 - 20 of 8,290

    Results (Q18703883)
    8 KB (14 words) - 18:49, 11 March 2016

    Results (Q1718079)
    8 KB (30 words) - 03:10, 4 September 2015

    Results of the Archbold Expeditions. No. 9. A new race of Hyosciurus (Q19609458): scientific article
    5 KB (64 words) - 10:02, 22 March 2015

    Results of the Archbold Expeditions. No. 1. A new genus and species of squirrel from Celebes (Q19606906): scientific article
    5 KB (84 words) - 15:15, 31 March 2015

    Results May Vary (Q915327): album
    15 KB (122 words) - 18:17, 5 February 2016

    oxidation-reduction process (Q14349764): A metabolic process that results in the removal or addition of one or more electrons to or from a substance, with or without the concomitant removal or addition of a proton or protons.
    5 KB (42 words) - 20:02, 27 January 2016

    extracellular matrix disassembly (Q14877674): A process that results in the breakdown of the extracellular matrix.
    7 KB (16 words) - 21:35, 26 January 2016

    Journal of Negative Results in BioMedicine (Q6295634): journal
    7 KB (59 words) - 23:17, 26 February 2016

    extracellular matrix organization (Q14819468): A process that is carried out at the cellular level which results in the assembly, arrangement of constituent parts, or disassembly of an extracellular matrix.
    6 KB (36 words) - 21:34, 26 January 2016

    actin cytoskeleton organization (Q14873990): A process that is carried out at the cellular level which results in the assembly, arrangement of constituent parts, or disassembly of cytoskeletal structures comprising actin filaments and their associated proteins.
    9 KB (39 words) - 22:06, 26 January 2016

    chromatin organization (Q15322272): Any process that results in the specification, formation or maintenance of the physical structure of eukaryotic chromatin.
    7 KB (29 words) - 21:41, 26 January 2016

    Glorious Results of a Misspent Youth (Q3109150)
    7 KB (90 words) - 08:37, 6 September 2015

    cytoskeleton organization (Q14864151): A process that is carried out at the cellular level which results in the assembly, arrangement of constituent parts, or disassembly of cytoskeletal structures.
    5 KB (30 words) - 22:22, 26 January 2016

    response to toxic substance (Q14330803): Any process that results in a change in state or activity of a cell or an organism (in terms of movement, secretion, enzyme production, gene expression, etc.) as a result of a toxic stimulus.
    5 KB (40 words) - 11:04, 29 January 2016

    response to virus (Q14864931): Any process that results in a change in state or activity of a cell or an organism (in terms of movement, secretion, enzyme production, gene expression, etc.) as a result of a stimulus from a virus.
    5 KB (44 words) - 01:09, 30 January 2016

    protein glycosylation (Q14839643): A protein modification process that results in the addition of a carbohydrate or carbohydrate derivative unit to a protein amino acid, e.g. the addition of glycan chains to proteins.
    11 KB (38 words) - 13:57, 26 January 2016

    response to cytokine (Q14860535): Any process that results in a change in state or activity of a cell or an organism (in terms of movement, secretion, enzyme production, gene expression, etc.) as a result of a cytokine stimulus.
    5 KB (43 words) - 02:34, 26 January 2016

    microtubule cytoskeleton organization (Q14330650): A process that is carried out at the cellular level which results in the assembly, arrangement of constituent parts, or disassembly of cytoskeletal structures comprising microtubules and their associated proteins.
    9 KB (40 words) - 22:10, 26 January 2016

    cellular response to mechanical stimulus (Q14859645): Any process that results in a change in state or activity of a cell (in terms of movement, secretion, enzyme production, gene expression, etc.) as a result of a mechanical stimulus.
    9 KB (42 words) - 01:15, 30 January 2016

    response to organic substance (Q14861556): Any process that results in a change in state or activity of a cell or an organism (in terms of movement, secretion, enzyme production, gene expression, etc.) as a result of an organic substance stimulus.
    5 KB (41 words) - 00:47, 30 January 2016

View (previous 20 | next 20) (20 | 50 | 100 | 250 | 500)


That used to work better, even when search couldn't find items.
--- Jura 19:55, 15 March 2016 (UTC)

I think this is phab:T110648. Sjoerd de Bruin (talk) 07:36, 16 March 2016 (UTC)
I recall when that was opened, but back then one could search for properties, but one just wouldn't find relevant items. Now properties seem to be gone. It doesn't seem to mention recent activity either.
--- Jura 08:17, 16 March 2016 (UTC)
@aude: Could you have a look? Searching for IMDB still has the identifier among the top results. --Lydia Pintscher (WMDE) (talk) 09:54, 16 March 2016 (UTC)

API change rank and order

I've added a new code value settlements. I would like to change rank of old value on "deprecated" or, may be, to change the order by moving the new value in the first place. Advise how to act in this case, and examples in the API request for 1. change of rank and 2. change of the order, please. Игорь Темиров (talk) 18:48, 15 March 2016 (UTC)

There is currently no specific API to change the rank of a statement. What you can do is to use wbgetclaims to query an existing statement, change the rank and pass the otherwise unchanged statement to wbsetclaim. Similar thing with the order. wbsetclaim does have an index option. Note that this is an option of the API, not an attribute of the statement. Query the statement with wbgetclaims and pass the unchanged statement to wbsetclaim with a specific index set. Does that help? --Thiemo Mättig (WMDE) 10:22, 16 March 2016 (UTC)

superscript/italic in labels, and unicode

Hi devs, I have a part technical part social question related to labels and formated strings (or not).

A few people started using non common unicode character for a little bit of typography in labels. As labels are used in Wikipedia articles, it can be important that they respect typography rules for some people. For example after a chat I labelled Napoléon Iᵉʳ Napoleon (Q517)      in french. The problem with that it's that it's a trick as these character are not supposed to be used like that as the superscript letters are dispached in several distinct unicode blocks with distinct intended usage. As a consequence there is not a complete alphabet in Unicode with superscript letter, and the rendering can be inconsistent. I also just used some italic unicode letters using this page ( http://qaz.wtf/u/convert.cgi?text=Capillariidae ) for the title of that article : Proposal of a new systematic arrangement of nematodes of the family Capillariidae (Q23199049)      because scientific name are supposed to be in italic. This works fine in that case except the letters unicode names are labelled MATHEMATICAL and that the name is a biological one.

The point beeing : unicode seems pretty beneficial for Wikidata, but this is not perfect. Can this team and/or wikimedia get in touch with Unicode to make a request to make possible to be able to perfectly handle those usecases ? I guess Wikimedia could have some importance for them. author  TomT0m / talk page 15:39, 18 March 2016 (UTC)

Only the scientific names up to the rank of genus are usually given in italics. The family name in your example wasn't printed in italic (Q344098). --Succu (talk) 20:10, 18 March 2016 (UTC)
See http://www.unicode.org/faq/ligature_digraph.html (specifically, the last three questions under "Presentation Forms" and the second and fourth questions under "Plain Text"). Unicode's goal is not to make text formatting systems obsolete, it simply provides the characters to which text formatting can be applied. I'm pretty sure that any request for things like superscripts and italics - even from Wikimedia - will be met with a reply along the lines of "That's outside the scope of Unicode. If you need to apply text formatting, you should use something designed for formatting text.". - Nikki (talk) 16:38, 18 March 2016 (UTC)
The case of taxon names is pretty similar to the maths case and not only stylistic: there is semantic encoded in the fact that they are in italics. The case for superscript is that it's not to put arbitrary text in superscript, it's to respect standard typography for abbreviations - and as well it encodes semantics. author  TomT0m / talk page 17:11, 18 March 2016 (UTC)
Anyway, if it's not possible, would that be possible to add a rich text label feature in Wikidata ? That would make the labels way more complicated however ... it does not seem to be in RDF spirit, for example, to encode rich text in a label. author  TomT0m / talk page 17:11, 18 March 2016 (UTC)
I wouldn't say it's that similar - Unicode encodes characters because the character itself is semantically different (because Unicode is only concerned with identifying distinct characters). Formatting applied to strings is beyond what they're trying to do even when it has a semantic meaning, e.g. superscript in maths also has a semantic meaning, but that has not led them to encode more superscript characters. :)
There was actually a request just a couple of weeks ago asking for things like italics and superscripts, see Wikidata:Project_chat/Archive/2016/03#Wiki_mark-up_in_string-type_properties. There is some support for it and Lydia seems open to the idea of potentially adding a new data type if there is a clear enough use case for it. I'm not sure if there's any chance of making it possible for labels, they're already really restricted compared to statements (no qualifiers, no references... :().
- Nikki (talk) 19:50, 18 March 2016 (UTC)

WDQS Query results and IE11

Hi! For some days the table view displays no results using IE11/Win7 (Works fine with Firefox 45). --Succu (talk) 13:22, 16 March 2016 (UTC)

@Jonas Kress (WMDE): Can you have a look please? --Lydia Pintscher (WMDE) (talk) 20:44, 16 March 2016 (UTC)
Should be fixed when https://gerrit.wikimedia.org/r/#/c/277993/ is deployed, sorry for your inconvenience --Jonas Kress (WMDE) (talk) 13:48, 17 March 2016 (UTC)
Danke. --Succu (talk) 15:57, 17 March 2016 (UTC)
Now I get an empty view, Jonas Kress (WMDE). Only the table header, the navigation and search controls are shown, but no results. :( --Succu (talk) 12:41, 21 March 2016 (UTC)
Please be patient the patch you need still has to be deployed. --Jonas Kress (WMDE) (talk) 13:58, 21 March 2016 (UTC)

  Done

SPARQL query shows deleted item

@Smalyshev (WMF):This query returns a deleted item: 

PREFIX ps: <http://www.wikidata.org/prop/statement/>
PREFIX p: <http://www.wikidata.org/prop/>
PREFIX wdt: <http://www.wikidata.org/prop/direct/>

SELECT ?item ?YouTube ?YouTubeEntity WHERE {
	?item p:P2397 ?YouTubeEntity .
          ?YouTubeEntity ps:P2397 ?YouTube .
     filter(SUBSTR(str(?YouTube),1,2)!="UC")
           }
Try it!

I know there's a ticket about this, but that's a restricted ticket, which I can't access, so I can't log it in the ticket. The deleted item is Q23038801. Mbch331 (talk) 21:36, 19 March 2016 (UTC)

  Updated Smalyshev (WMF) (talk) 22:10, 19 March 2016 (UTC)
It might help to do create a short public viewable ticket with a summary stating that occasionally there is an issue with deleted items appearing and these can be deleted on demand.
The actual resolution depending on a blocking/not viewable ticket.
--- Jura 11:23, 20 March 2016 (UTC)

formatterURL for Wikidata ID

P2617 (P2617) is now available. Please see discussion at Wikidata:Property proposal/Authority control#archINFORM.net availability for an example of how it should be used. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:46, 17 March 2016 (UTC)

Hey Andy. This seems like a very bad modeling. The identifier on that other page should just be added like any other identifier. It just happens to be the same as the item ID (in most cases - I don't know how it handles merges for example). Doing this with true/false also lets us run into all kinds of export and UI issues. --Lydia Pintscher (WMDE) (talk) 17:02, 17 March 2016 (UTC)
@Lydia Pintscher (WMDE): I'm not clear what you mean by "The identifier on that other page". Please explain. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:56, 17 March 2016 (UTC)
With "The identifier on that other page" Lydia means the (primary) identifier for the given concept on the other page (this is usually modeled with an external identifier property these days).
I also agree that using an external identifier property is the best way to model these things, as external identifiers have the proper semantics and tooling support for this use case. Even if the external identifier value is the id of the item, it's still an external identifier and should be modeled as such. Cheers, Hoo man (talk) 20:41, 17 March 2016 (UTC)
Hi, I'm the main developer of "that other page" (archINFORM). The intent of the proposal has a more general background: Wikidata becomes the definitive provider of authority controls. Most content providers will include earlier or later Wikidata-ID's into their database entries (and could set up easily resolver links like the proposed one from archINFORM). In the past a similar way already established for the German GND-Identifier, called BEACON ( https://de.wikipedia.org/wiki/Wikipedia:BEACON )
Isn't it better to transfer the job of maintaining the relationships to the specific content providers, than doing that in Wikidata?
Maybe there are some conflicts in existing UI and export modules, but this shouldn't be the reason to decline this idea. --Arch2all (talk) 10:21, 18 March 2016 (UTC)
@Arch2all: I think it's great that Wikidata IDs are being used to model and identify data within other projects. Obviously there are several ways to model this on Wikidata, as we have more knowledge about these ids already in place than what we have with identifiers in general. It might be tempting to "re use" the information that is already in Wikidata (the item id) by inventing a new way to model these things, but I don't think we should do that, mainly because I'm concerned about consistency and tooling support.
I fail to see that the advantages of having a new and fundamentally different way to model/ handle a subset of identifiers outweigh the disadvantages mostly in regards to consistency. Also this new form of modeling would only be correct if the external data set is built around Wikidata in a way that ties the (actual current) id of the Wikidata item to the concept within the external data set, or in other words: If I merge an item and the id changes, that would need to be reflected in your data set immediately, otherwise such a modeling would be problematic in these cases. Cheers, Hoo man (talk) 17:21, 18 March 2016 (UTC)
Imagine the external data provider changes his website/database structure. All related wikidata entries will be broken! If the provider offers a Wikidata resolver like the arcHINFORM proposal suggests, this won't be a problem. --Arch2all (talk) 17:51, 18 March 2016 (UTC)
@Arch2all: Sure, but how would the proposed modeling help in that case? For the case you describe it makes no difference whether we have an external identifier that contains the item id or something that magically maps to a URL using the item id. I still don't see what the problem with having external identifiers whose value is identical to the item's id is. Cheers, Hoo man (talk) 18:09, 18 March 2016 (UTC)
I agree with Hoo here, I don't see the benefits to creating a new way to model these. It creates more complexity for us and for people using the data when the only difference is that the ID is most likely the same as the current Wikidata one - only "most likely" because we can't guarantee that any external website is actually in sync with Wikidata (even Wikidata's own query service which updates almost instantly manages to get out of sync sometimes). - Nikki (talk) 23:07, 22 March 2016 (UTC)
If that is indeed what Lydia meant, it is a pity that no-one has yet made that point in the property proposal I linked to above, which has been open - on a well-visited page - since 28 January. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:55, 18 March 2016 (UTC)
You should have taken this to the project chat, as what you are suggesting here is way more than just a new property: You were suggesting a whole new way to model something that is fundamental to Wikidata. I guess that many also didn't understand the purpose of your request, thus didn't comment. Cheers, Hoo man (talk) 17:21, 18 March 2016 (UTC)
The property proposal was not mine. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:25, 18 March 2016 (UTC)

Updates for Special:ConstraintReport

It seems that the weekly update of constraints hasn't run for months, e.g. since August, looking at Special:ConstraintReport/Q2345 listing an error for "Value type [...] class: award; relation: subclass" on Q2345#P1411 that was changed in August [1].
--- Jura 11:23, 20 March 2016 (UTC)

Would run the update?
--- Jura 11:23, 20 March 2016 (UTC)

@Hoo man: would you? --Lydia Pintscher (WMDE) (talk) 08:53, 21 March 2016 (UTC)
@Jura1:   Done, I've re-generated the data late yesterday and just applied it. - Hoo man (talk) 10:39, 22 March 2016 (UTC)

Adding qualifiers after you have created or edited a claim

If you have created or edited a claim, you can after saving see a link to add a qualifier to that claim. If you click this link then you can add a qualifier as normal, however you cannot save this addition (there is no save button) until you click to "edit" the statement, which fortunately does not remove what you have just done. However it is not I suspect what is intended. I have no preference whether the fix should be to disable adding qualifiers without first clicking to to edit the statement, or to allow qualifiers to be saved without having click this - however one or the other should be done. This applies in Firefox 45.0 on Xubuntu Linux, I haven't tested on other systems/browsers. Thryduulf (talk: local | en.wp | en.wikt) 20:15, 21 March 2016 (UTC)

According to phab:T128317 this has been resolved and the fix will be deployed soon. Sjoerd de Bruin (talk) 20:18, 21 March 2016 (UTC)
@Thryduulf, Sjoerddebruin: I've just applied the fix to Wikidata, it should be fine again. Cheers, Hoo man (talk) 21:38, 22 March 2016 (UTC)

Translation extension: $wgTranslateBlacklist

If we remove values from the above configuration ("en"), can do the translations discussed in Wikidata:Translators'_noticeboard#Source_language?
--- Jura 06:19, 23 March 2016 (UTC)

As far as I know, this is not possible or has at least not been properly tested by the language engineering team. Feel free to propose such a change, but it's quite likely that additional work or at least testing will be required before that change can be implemented. Cheers, Hoo man (talk) 13:16, 23 March 2016 (UTC)
I brought it up on the translators' board and apparently it had already been asked for other wikis. Can we attempt to test it, e.g. at https://test.wikidata.org ? Either it's a simple change or too complex to do in the short term.
--- Jura 16:09, 23 March 2016 (UTC)
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Contact_the_development_team/Archive/2016/03&oldid=319287858"