On 26.04.2015 22:28, Gerard Meijssen wrote:
It is a matter of perspective. From my perspective a value exists or
not. Depending on that I may want to process. When you state novalue
there is a value of novalue and that is not the same as there not being
a value in the first place.

Ah, I see. I think any query interface should allow you to find both: things with a novalue-claim and things with no claim at all. You can then pick your perspective on these two things as you like.

However, it would be an error to treat "novalue" as a kind of "some value", and it would be an even bigger error to treat "novalue" as a specific value (that can be equal to other such values). For example, a WDQ tree query should never go through "novalue" (and not even through a "somevalue" a.k.a. "unknownvalue"), as I am sure you will agree.



Wikidata-l mailing list

Reply via email to