Last modified: 2014-11-15 21:49:00 UTC
Some Wikidatans have come up with a system of machine readable constraint templates that are added on the talk page. There's a list of these at https://www.wikidata.org/wiki/Category:Properties_with_constraints, and a summary at https://www.wikidata.org/wiki/Wikidata:Database_reports/Constraint_violations/Summary which shows different violations. Right now bot authors have to write code to manually parse these templates, it would be great if we could have these constraints stored in the data model itself. The software wouldn't necessarily need to enforce them, just store them.
A RFC would be good with a little bit of detail on how to do this. My first idea would be to include statements for properties, but I am unsure if this would be sufficient?
https://www.wikidata.org/wiki/Wikidata:Project_chat#Property_Metadata as a start
*** Bug 45676 has been marked as a duplicate of this bug. ***
(In reply to comment #2) > https://www.wikidata.org/wiki/Wikidata:Project_chat#Property_Metadata as a > start https://www.wikidata.org/wiki/Wikidata:Project_chat/Archive/2013/10#Property_Metadata -- Archived link.
I think it could be good if 'Domain' and 'Range' statements link to query pages (when we have these) rather than Qitems.
We're getting close to finishing this feature. One open important question is: Do we need qualifiers and sources on all this or are simple property/value pairs enough? Usecases please :)
Just normal statements - I don't see a need for us to qualify simple Wikidata-internal information.
.. The French have a system of nationality whereby you can be "French" because of the law without being considered to be French. So the Nationality would be French and the qualifier would be a pointer to this law. NB this is just one use case.. There are many more.
(In reply to Gerard Meijssen from comment #8) > .. The French have a system of nationality whereby you can be "French" > because of the law without being considered to be French. So the Nationality > would be French and the qualifier would be a pointer to this law. > > NB this is just one use case.. There are many more. That is for item pagess. This bug is about statements on property pages :)
There are more countries like this.. When you flag it as being applicable for a country, you can query it.
(In reply to Gerard Meijssen from comment #10) > There are more countries like this.. When you flag it as being applicable > for a country, you can query it. Gerard, I think you are getting it wrong... this is to allow entering statements on property pages like this one: https://www.wikidata.org/wiki/Property:P27 If you try to add statements to that property page, you will see that it is not possible at the moment.
These statements effectively become a machine readable definition of a property: * Domain * Range * Subproperty of * Symmetric property * Inverse property of The only case where I can see a need for additional info is to provide a link to the discussion where that statement (definition) was agreed. i.e. a link to a wikidata discussion. If that is a problem then maybe we could ask editors to enter a summary when they change a property or a statement about a property. Could we start to do this for changes to the property name as well?
I agree sourcing such properties is not relevant. I believe that qualifiers might be a good thing, for example for a "start date" when Domain or Allowed values need to be different starting from a certain date (e.g. after geopolitical changes). I hope that the community will agree to add some more text properties, like: purpose, usage... lack of proper property documentation proved pretty painful in some cases.
done now but still needs reviews before deployment
Is there a patch or series of patches in Gerrit for this that I can monitor to see how review is progressing?
(In reply to Zell Faze from comment #15) > Is there a patch or series of patches in Gerrit for this that I can monitor > to see how review is progressing? What's blocking it at the moment is the new serialization format. There current state of that is that there are a number of small things needing fixing to make sure we don't break anything.
