Hoi,
The big point of having data that is consistently "different" is that it
can be queried and using query it can be changed. There is a problem in
having so much data but the problem in not having data is even bigger. A
good example is in "professor".. it has its own structures (multiple) and
in essence they are a "position". Like a president, a professor retains
this title after the period of active University involvement. Some people
insist that it is a "profession", I am of the opposite point of view. Now
someone who holds the profession of professor must be or must have been in
the employment of a university. So it is possible to find problematic
professors.
So yes it is a scaffold and yes things may change because there is all too
much that is severely problematic. This is not a problem when the data has
some kind of structure.
Thanks,
GerardM
On 17 May 2017 at 18:30, Joakim Soderberg <[email protected]>
wrote:
> Aubrey,
> Thanks for the explanation and enlightening examples.
> Here are my two cents’ worth.
>
> Having an open community defining semantics to cross domain items is
> powerful and challenging.
> To mention one problem, I have come across several statements where
> ‘instance of’ has been confused with ‘about’, which is harmful when you
> want to query for a type.
> Perhaps manual curation or bots can help out more?
>
> The challenge is that as wikidata complexity grows, it becomes more
> difficult for the community to chose the right properties and classes.
> A somewhat related example from Wikipedia; the article about Amy_Brand was
> written using the inbox template “adult biography” instead of “person”,
> which had the consequence of the entity being classified as “adult actor”
> in dbpedia.
>
> Best regards
> Joakim
>
> On May 17, 2017, at 8:17 AM, Andrea Zanni <[email protected]>
> wrote:
>
> [Sorry for cross-posting. I just sent this mail to the Wikicite mailing
> list, but I thought it could be of interest also here. Moreover, the
> community here could provide some important insight to better explain
> wikidata to "newbies" and set the right path for a meaningful conference.]
>
>
> Dear all,
> sorry for the messy email, but I'd like to to express a small concern I
> have regarding the awesome Wikicite conference we'll have in few days.
>
> My main point is that Wikidata is *complex*.
> It's not just the data model (which is not easy, per se), but the whole
> idea behind wikidata, the policies, the community, the workflow, the tools,
> and the whole helo of vagueness that sorrounds it.
>
> My favourite example about this is this little story:
> https://en.wikipedia.org/wiki/Blind_men_and_an_elephant
> (which was actually a very good metaphor used by my professors introducing
> the Information Science course...)
>
> Different people with different background work with data in different
> manners: the word "data" (as "information") means anything and its
> contrary, so we must be careful. You know better than me that everyone
> projects her own dreams and delusions upon Wikidata, so we must work
> towards a good understanding at the beginning, to avoid the painful and
> time-consuming job of making people un-learn things they think they know.
> This was quite evident especially last year, when a lot of librarians and
> wikimedians were in the same room, and everyone knew many things about
> metadata and their metadata model: librarians had difficulties grasping
> things about Wikibase, Wikidata, policies and communities, and wikimedians
> about bibliographic models and complexities.
>
> We spent at least 30 minutes explaining Wikidata from the beginning, also
> adding some "color" about strategies, policies and community.
>
> I'll try to make examples.
>
> * we need to explain that in Wikidata (and Wikibase) everything is an
> *item*. Everything. Every items has properties, values, qualifiers. What's
> not possible is to have sets or clusters of items, and to give properties
> to such sets and clusters. This is somehow related to the book topic.
>
> * we need to explain that there are at least 2 different possible
> strategies: create few general properties and many specific items,
> or create many specific properties and less general items. Wikidata chose
> the former.
>
> * we need to explain the "scaffolding principle", meaning that we don't
> need to put *all the info in all the items*. We need to create and organize
> items that are *queriable*, in such a way that I can make a query that get
> all the data and details I need, scattered among different items. If the
> items in questions are built "on top of each other", this is doable. It is
> actually very important to understand this, because people get confused
> about how many things to say inside one item. This principle (and the
> former) was explained to me by Tobias1984, and helped me a lot in my
> understanding of Wikidata.
>
> I think that this kind of insight is crucial for working with Wikidata in
> a meaningful way, because wikidata offers *one product*, with *one data
> model*, and it simply impossible to adapt and stretch the Complexity Of The
> World to Wikidata without loss of information.
> I'm sure many of you know this perfectly, but other people maybe don't, or
> at least they will struggle with it.
> We all come from different backgrounds and are emotionally very attached
> to our models and our crucial pieces of information we don't want to lose:
> in this sensse, Wikidata is much more a "negotiation" than Wikipedia,
> because Wikipedia is not structured and llows for much more messiness.
>
> Every model and decision we will make will be a trade-off, and I think
> it's worth trying to save time and effort trying to establish these
> boundaries at the beginning.
> Of course, there are many insights about Wikidata I don't have, but those
> are kinda the things we want to understand first in these kind of
> conferences.
> Hope this makes sense.
>
> Aubrey
>
>
>
>
>
>
>
> _______________________________________________
> Wikidata mailing list
> [email protected]
> https://lists.wikimedia.org/mailman/listinfo/wikidata
>
>
>
> _______________________________________________
> Wikidata mailing list
> [email protected]
> https://lists.wikimedia.org/mailman/listinfo/wikidata
>
>
_______________________________________________
Wikidata mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikidata
