Ah, Simon, I missed your email before writing mine as it was in a different
thread in my client. Sorry.

On Wed, Mar 2, 2016 at 9:47 AM Simon Razniewski <[email protected]>
wrote:

> Hi Tom,
>
> Thanks for your message, you are touching upon a very crucial point, the
> justification of completeness. For several properties, completeness is
> clearly impossible ("SignificantEvent", "Occupation", "AwardsReceived" -
> what is significant, what counts as an occupation, and didn't everyone once
> win something like the 2nd place in the school chess tournament?).
>
> Still, for some properties, we believe stating completeness to be
> justified, or at least no more unjustified than entering data. Are Malia
> and Sasha all children of Obama? To the best of common knowledge yes, and
> if we are to doubt this, we might similarly doubt whether Malia is a child
> of Obama at all, so stating completeness looks to me no less justified than
> stating that Malia is a child of Obama.
>
> Regarding the other aspect you mention, the time variance of data, the
> best we can think of is to qualify completeness statements with an "as-of"
> timestamp.
>
> "Malia and Sasha are all children of Obama, as of 2nd of March, 2016".
>
> Besides that there may be a portion of well-known history that is not
> going to change any more, such as "founders of Apple", "Kings of the
> Italian Monarchy", "Participant nations in the 2012 Olympics" and similar,
> where time does not play a role.
>
> Deciding and agreeing for which properties completeness assertions make
> sense and for which not would probably require a significant effort by the
> domain experts (i.e. you, the Wikidata authors), the above examples are
> only guessed.
>
> Do you know potentially of any guidelines/agreements in Wikidata to decide
> which properties ("child") are more well defined, and which ones
> ("significantEvent") are not?
>
> Thanks and best wishes,
> Simon
> _______________________________________________
> Wikidata mailing list
> [email protected]
> https://lists.wikimedia.org/mailman/listinfo/wikidata
>
_______________________________________________
Wikidata mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikidata

Reply via email to