Re: Wikidata-Links

I'm using Google translate, and quoting accordingly. My apologies if the translation was not very accurate and made me misunderstand things.

woodpeck wrote:

This is something I have not fully understood at Wikidata. With Wikipedia, one is strictly concerned that each language area has its own "culture"; A German-language article is usually not easy to translate the English or vice versa, but may well have different accents on the same subject or even a contradictory definition. This is also wanted, because no culture should be "authoritative".
With Wikidata does not go however, one must set oneself to one. And then I see problems; Keyword: in an amenity = bar in Italy there is espresso and cornetto in the morning, in an amenity = bar in Germany there is only the garbage bags and the cigarette smoke from the evening before.

I agree that different cultures (not just languages!) sometimes present slightly different description in Wikipedia. But we (OSM) use wikipedia/wikidata tags as links to Wikipedia, not as sources. If it was a source, we should use "source:wikipedia" or something similar. When a village X has a wikipedia/wikidata tag, it means "This OSM object has a corresponding page in Wikipedia, so you may want to read-up about it there". Moreover, any data consumer who may wish to show Wikipedia page to a user would ignore the language you specify - because most users would want to read Wikipedia article in their language, and they will simply use Wikidata to lookup that page in another language. Also, this sounds very theoretical. Have you actually seen any problems like this? How substantial is this problem?

woodpeck wrote:

I'm concerned that OSM is taking the reins out of our hands by people who have no idea of and no interest in OSM, which then begin to edit stuff with us to make it better for the Wikidata world.

I am concerned that you might be seriously offending people with this statement. Most people participating in OSM discussion care deeply about OSM. Please don't accuse just because you don't like the position of others, this is very inappropriate. I myself have met some of the active proponents of using Wikidata at OSM conferences.

woodpeck wrote:

In the discussion on the talk mailing list, many possible problems were pointed out, including that Wikidata basically has a different mentality, the same as in Wikipedia, that you can capture what you find in secondary sources, while at OSM capture what we see on the ground. Or on the fact that Wikidata gets a wet rubbish around database rights of third parties and cheerfully (for example) coordinates from Google.

This is true that each language in Wikipedia, as well as Wikidata, have different set of rules. Google Translate amusingly translated your expression as "gets a wet rubbish around database rights", and I take it to mean that Wikipedia simply ignores the laws. I don't think this is accurate - Wikipedia has a large number of very diligent lawyers who might have a somewhat different interpretation of the laws, especially in how they relate to US, and the community is extremely careful with the copyright law. But again, this is irrelevant. We are not copying data from Wikipedia or Wikidata. We are discussing linking to it.
If I, as a user, want to read more about a place, I will click on the link to Wikipedia. If I don't care, I ignore it. This is identical to a "url" or "website" tag. We are not discussing if the content of that site is violating database right or has other issues, right?

woodpeck wrote:

I do not want to reject everything categorically; At the beginning I was even inspired by the idea that people would finally stop to miss a name in an obscure language in a world-wide change set by means of strange third-party sources, because you can find everything in Wikidata. (Keyword "on the ground" mapping - I live in Karlsruhe, but I would not have a chance to determine by the signpost, if eg name: az = Karlsruye is correct ...) Bloss I now feel that here of a few is marching with fairly rapid steps in a direction that is not really clear to anyone. I would be more comfortable, if one would make a few thorough thoughts, also to the question: What is, if the competition project for Wikidata is about to start? We then support both - patron_saint: wikidata = Q1234, patron_saint: datawiki = # AABBCC, ...? What are the requirements for a third-party database to be an integral part of OSM?

I think you are exaggerating. This whole thing started after I added a few thousand wikidata tags to the objects that already had wikipedia tags. Same way as iD editor did it automatically for hundreds of thousands, often without any oversight. The number of wikidata tags in the past few years have been used for all sorts of things, there is no new development. You could say it has reached critical mass and enough people started noticing it. But you don't have to use them - noone is removing "brand", or "subject", or any other tags. These are additional tags - for those who need them - and again, there are a lot of data consumers who really do want them.
To answer your question - if someone creates another such project - great. But I seriously doubt it will happen. Google had a similar project called Freebase for a while, and with all their resources, they gave up - and moved all their data to Wikidata. For a very simple reason - it is very easy for a few developers to create a new project. It is extremely hard to build a large community around it, especially when something similar already exists and has a huge community and resources. But regardless - if someone creates it, and people find it useful, i see no problem for community to start adding new tags. Who am I to tell others what new information they should or shouldn't add?