Le 2013-05-06 00:09, Jona Christopher Sahnwaldt a écrit :
On 5 May 2013 20:48, Mathieu Stumpf <[email protected]>
wrote:
Le dimanche 05 mai 2013 à 16:28 +0200, Jona Christopher Sahnwaldt a
The ontology is maintained by a community that everyone can join at
http://mappings.dbpedia.org/ . An overview of the current class
hierarchy is here:
http://mappings.dbpedia.org/server/ontology/classes/ . You're more
than welcome to help! I think talk pages are not used enough on the
mappings wiki, so if you have ideas, misgivings or questions about
the
DBpedia ontology, the place to go is probably the mailing list:
https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
Do you maintain several "ontologies" in parallel? Otherwise, how do
you
plane to avoid a "cultural bias", and how do you think it may impact
the
other projects? I mean, if you try to establish "one semantic
hierarchy
to rule them all", couldn't it arise cultural diversity concerns?
We maintain only one version of the ontology. We have a pretty
diverse
community, so I hope the editors will take care of that. So far, the
ontology does have a Western bias though, more or less like the
English Wikipedia or the current list of Wikidata properties.
JC
I can't see how your community could take care of it when they have no
choice but not contribute at all or contribute to one ontology whose
structure already defined main axes. To my mind, it's a structural bias,
you can't go out of it without going out of the structure. As far as I
understand, the current "ontology"[1] you are using is a tree with a
central root, and not a DAG or any other graph. In my humble opinion, if
you need a central element/leaf, it should be precisely
"ontology"/representation, under which one may build several world
representation networks, and even more relations between this networks
which would represent how one may links concepts of different cultures.
To my mind the problem is much more important than with a local
Wikipedia (or other Wikimedia projects). Because each project can expose
subjects through the collective representation of this local community.
But with wikidata central role, isn't there a risk of "short-circuit"
this local expressions?
Also, what is your metric to measure a community diversity? I don't
want to be pessimist, nor to look like I blame the current wikidata
community, but it doesn't seems evident to me that it currently
represent human diversity. I think that there are probably a lot of
economical/social/educational/etc barriers that may seems like nothing
to anyone already involved in the wikidata community, but which are
gigantic for those non-part-of-the-community people.
Now to give my own opinion of the representation/ontology you are
building, I would say that it's based on exactly the opposite premisses
I would use. Wikidata Q1 is universe, then you have earth, life, death
and human, and it seems to me that the ontology you are building have
the same anthropocentrist bias of the universe. To my mind, should I
peak a central concept to begin with, I would not take universe, but
perception, because perceptions are what is given to you before you even
have a concept for it. Even within solipsism you can't deny perceptions
(at least as long as the solipcist pretend to exist, but if she doesn't,
who care about the opinion of a non-existing person :P). Well I wouldn't
want to flood this list with epistemological concerns, but it just to
say that even for a someone like me that you may probably categorise as
western-minded, this "ontology" looks like the opposite of my personal
opinion on the matter. I don't say that I am right and the rest of the
community is wrong. I say that I doubt that you can build an ontology
which would fit every cultural represantions into a tree of concepts.
But maybe it's not your goal in the first place, so you may explain me
what is your goal then.
[1] I use quotes because it's seems to me that what most IT people call
an ontology, is what I would call a representation.
_______________________________________________
Wikidata-l mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikidata-l
