On 22/09/2019 08:48, Sebastian Hellmann wrote: ...
The formula here is quite easy: If you look at DBpedia's data in detail or a part of it, it will not shine so much since it is extracted,
Sure, but I think that this is not clear to many people who are currently using DBpedia as a dataset (even if only for testing/research purposes). Also, there would surely be value in analysing the differences more closely. I agree with you that quantitatively, Wikidata might be orders of magnitudes ahead. Yet, there can still be individual bits of information that are in DBpedia but missing from Wikidata so far.
For example, DBpedia EN has 32 people educated at the University of Leipzig, whereas Wikidata has 1217. Nevertheless, there is, for example, John Henry Wright (Q6238997), who is known to DBpedia but not to Wikidata (yet). Such cases might be worth systematic weeding out so that we can really come to the point where Wikidata is a strict superset of all (correct) data in DBpedia.
Cheers,
Markus