<div dir="ltr"><div>On Thu, 12 Sep 2019 at 09:45, Janko Mihelić <<a href="mailto:janjko@gmail.com">janjko@gmail.com</a>> wrote:</div><div><br></div><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div class="gmail_quote"><div>One problem with the current system is that if you click one of those dwarfs in OSM, and see it's linked to an object in wikidata, you have no way of seeing if that is the whole wikidata object, or just a part of that object, unless you download the whole OSM database. Or if you are a human, and you look at</div></div></div></blockquote><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div> the wikipedia article, and see there should be a whole bunch of dwarfs. But that example doesn't seem as important.<br></div></blockquote><div><br></div><div>You're trying to solve the data problem (to the extent that it even is a problem) in a defective</div><div> way. It's defective because wikidata is not an integrated part of OSM designed to group objects.</div><div>It's outside of our control; even if we can abuse it that way, a random edit to wikidata will</div><div>lose the object grouping. The correct way to group them is with a relation. If we don't have a</div><div> suitable type of relation then propose one. Don't use wikidata as a workaround for not having a</div><div>suitable relation type or not having a part_of_a_group=yes tag. And even without any of that,</div><div>an overpass query in the general area for artwork_type=statue + other tags they have in</div><div>common will find them.</div><div><br></div><div>It's not a problem for humans either. If they're not interested in looking at the data item then</div><div>it doesn't matter if it's tagged wikidata=* or part:wikidata=* because they won't follow it. If</div><div>they are interested in looking at the data item then it doesn't matter if it's tagged wikidata=*</div><div>or part:wikidata=* because they'll reach the same data item either way and realize there</div><div>are seven dwarfs. And won't be able to find the other six easily from there. So a relation</div><div>is still the best way to do it, then apply wikidata=* to the relation.<br></div><div><br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div class="gmail_quote"><div>Currently, the second most numerous wikidata tag in OSM is <a href="https://www.wikidata.org/wiki/Q2961670" target="_blank">https://www.wikidata.org/wiki/Q2961670</a>, an item that describes all the roman roads in historic Gaul in France. All those ways, close to 500 of them, have wikidata=Q296167. That is obviously not good tagging. But how do you differentiate good wikidata tagging from bad tagging? I think this rule and part:wikidata are the way to clean this up. I would give all these roads part:wikidata=Q29616, and than that looks much closer to reality.
</div></div>
</div></blockquote><div><br></div><div>I think the only sensible solution is to delete the wikidata tags from *all* of them. That item is</div><div>for a category, not a unique object. OSM relations are not categories because we don't tag</div><div>categories (it would result in a gigantic taxonomic hierarchy of tagging). Those road should</div><div>never have been given that wikidata tag; individual roads get a wikidata tag only if that data</div><div>item applies solely to that particular road.<br></div><div><br></div><div>It seems to me that you're trying to find a way of mapping anything that has a wikidata tag.</div><div>If so, that seems like a bad idea. Use a wikidata tag to add extra information about a unique</div><div>OSM object, don't invent OSM objects and/or ways of mapping things in order to put every</div><div>wikidata item into OSM. We could put wikidata=Q2 on every object, following your line of</div><div>reasoning. Oh, sorry, part:wikidata=Q2.<br></div><div><br></div><div>-- <br></div><div>Paul</div><div><br></div></div></div>