Indeed. Thanks for the example. I'll probably incorporate it in my
talk at WikidataCon.
As far as I know there is no general method for nudging towards
consensus for cases like these. The onus appears to me to be on whoever
is entering the information to look for similar situations and model
them all the same. (In this case it appears that a recent change to the
Nobel Peace Prize was made to remove it being a subclass of Nobel Prize,
actually reducing commonality.)
But what can be done in the future? One way to go is to ask that
editors be more careful when editing items that might belong to a group,
and try to model them the same as other members of the group. Another
way to go is to ask that editors be more careful when editing items that
have parts/instances/subclasses and check that all the other items are
modeled the same way.
I prefer something similar to the second way, where editors of classes
and properties (or just about anything that is going to be the common
target of a property, but instance and subclass and subproperty seem to
me to be the most important such properties) are asked to be careful to
specify the relationship between the class or property and the other
items that target it. So whoever does major editing on Nobel Prize
should add a comment on the relationship between the various Nobel
Prizes and Nobel Prize. (Having such information is quite common for
concepts in Cyc.)
Actually Nobel Prize isn't the greatest example for my preference
because there doesn't seem to be any Wikidata items for the even the
famous Nobel Prizes. Suppose there was a Wikidata item for Einstein's
Nobel Prize in Physics. Then its relationship to Nobel Prize would
provide guidance for the relationship between the Nobel Prize in Physics
and Nobel Prizes itself.
I find modeling deficiencies like this in lots of places in Wikidata.
That's not a severe problem if you have the resources of Google to throw
at curating Wikidata information. But if you don't have this level of
resources available for curating Wikidata information then these sorts
of infelicities are a significant barrier to using Wikidata.
Peter F. Patel-Schneider
On 9/27/19 12:34 PM, Aidan Hogan wrote:
Hey all,
Andra recently mentioned about finding laureates in Wikidata, and it
reminded me that some weeks ago I was trying to come up with a SPARQL
query to find all Nobel Prize Winners in Wikidata.
What I ended up with was:
SELECT ?winner
WHERE {
?winner wdt:P166 ?prize .
?prize (wdt:P361|wdt:P31|wdt:P279) wd:Q7191 .
}
More specifically, looking into the data I found:
Nobel Peace Prize (Q35637)
part of (P361)
Nobel Prize (Q7191) .
Nobel Prize in Literature (Q37922)
subclass of (P279)
Nobel Prize (Q7191) .
Nobel Prize in Economics (Q47170)
instance of (P31)
Nobel Prize (Q7191) ;
part of (P361)
Nobel Prize (Q7191) .
Nobel Prize in Chemistry (Q44585)
instance of (P31)
Nobel Prize (Q7191) ;
part of (P361)
Nobel Prize (Q7191) .
Nobel Prize in Physics (Q38104)
subclass of (P31)
Nobel Prize (Q7191) ;
part of (P361)
Nobel Prize (Q7191) .
In summary, of the six types of Nobel prizes, three different
properties are used in five different combinations to state that they
"are", in fact, Nobel prizes. :)
Now while it would be interesting to discuss the relative merits of
P31 vs. P279 vs. P361 vs. some combination thereof in this case and
similar such cases, I guess I am more interested in the general
problem of the lack of consensus that such a case exhibits.
What processes (be they social, technical, or some combination
thereof) are currently in place to reach consensus in these cases in
Wikidata?
What could be put in place in future to highlight and reach consensus?
Or is the idea more to leave the burden of "integrating" different
viewpoints to the consumer (e.g., to the person writing the query)?
(Of course these are all "million dollar questions" that have been
with the Semantic Web since the beginning, but I am curious about what
is being done or can be done in the specific context of Wikidata to
foster consensus and reduce heterogeneity in such cases.)
Best,
Aidan
_______________________________________________
Wikidata mailing list
Wikidata@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata
_______________________________________________
Wikidata mailing list
Wikidata@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata
