Topic on User talk:Bouzinac

Hi Bouzinac,

I can see why you want to split Babylon in two entities (even though that introduces other problems and I would not have done it this way). But, given that most Wikipedia articles are probably about the site rather than the country (political entity/city-state/whatever you want to call it), and given that the site is much more well known, I would suggest to make Babylon (Q5684) the item about the site and the new one Babylon (Q100329356) about the political entity. So exactly the other way around than what you did. What do you think?

Hi there ; Q number should not be a reference per se (merge, move, etc) ; but Wikidata must be storing the singlest concept. So indeed a Q for the site (when was it discovered, where, what time can we visit it, etc) and for the old country (who did rule it, whom country did it went war with, etc). What are the problem exactly ? If you find a wiki speaking mostly about the site rather than the old country, you can move the wikipage to the relevant Q

I know it isn't a reference, but for (just one) example, is about the site, whereas , which is First Babylonian Empire (Q733897), is about one of the states that emerged from Babylon. So actually you created an entity that already exists, and that is not the singlest concept at all. The concept of Babylon is not so simple as just a site and a state. It is much more complex than that. Keeping it all in one item is not a good solution, but creating a new item without doing any cleanup isn't any better either - or at least that's what I think. So again, I would leave the information about the site at Babylon (Q5684) as that better reflects the existing links from Wikipedia, and the historical contents of Q5684 as well. Best,

? Babylon as strictly archealogical site didn't exist? I'm no specialist of babylonian history but https://w.wiki/3enC looks OK. As always, WD needs some cleaning, with Q16630263 for instance.

Well, it sort of existed at Babylon (Q5684), until you split it off. So again, rather than creating a new item, it would have been better to split off the information on the political states from Q5684 to other already existing entities, such as First Babylonian Empire (Q733897). So my suggestion is: merge your item back into Q5684 and then update some of the other relevant items. Your new item complicates matters more than that it solves issues, I believe. Best,

Hi ChristianKl (talk • contribs • logs), what do you think about Zoeperkoe remarks ?

Just to be clear: I think that Babylon (Q5684) should be about Babylon as a city; it would have (for example) P31=archaeological site and P31=human settlement (with start-end dates) and P31=capital (if that exists, with, start-end dates) and it would have the UNESCO World Heritage site info. It would not have the values P31=historical country and P31=sovereign state. These would be moved to, for example, First Babylonian Empire (Q733897) and other entities about Babylonian empires (leaving aside the question that a clear-cut distinction between capital and state works for modern states, but not ancient ones like Babylon).

The reason that I think Q5684 should be about the site is that most Wikipedia articles linking to it are about the site, as well as the external identifiers and a quite a few other properties.

This would also mean that all information that is now in Babylon (Q100329356) will be moved back to Q5684, just as it was before Q100329356 was created.

By the way, the same goes for Ugarit (Q191369).

Babylon is a city (or human settlement) that existed 1000 AD. No archeological site for Babylon existed at that time. If there's an entity that exists today that's an archaeological site and part of the UNESCO World Heritage that's a different entity then the city of Babylon and thus deserves a different item.

Wikipedia is irrelevant to the question of whether to have one or more Wikidata items.

I don't know enough about Bablyon to say whether it makes sense to seperate a city of Babylon from a state of Babylon

@ChristianKl You say: "If there's an entity that exists today that's an archaeological site and part of the UNESCO World Heritage that's a different entity then the city of Babylon and thus deserves a different item." Just to get this clear: why? Why do you think that splitting them is a better solution than keeping the city and archaeological site in one item and solving this with start/end dates on the property instance of=city?

Because Wikidata is about the singlest concept. If you mix them, you would have both "this is a ruin" + "this has been demolished at year YYYY" AND "this was a city, populated with XXX people, capital of state XXX" Obviously, you cannot pretend a ruin was a capital at same time and that the ruins were demolished at year YYYY.

That depends on how you model it. So, just to be clear: are you saying that every item about every archaeological site on Wikidata needs to be split in two because they cannot be an archaeological site and a human settlement (or city, or temple, or whatever) "at the same time"?

Yes : it is clearly stated there https://www.wikidata.org/wiki/Help:Items You just don't mix carrot and cabbages in the same field.

So if I understand you correctly, you argue that every item about every archaeological site should be split. Would you mind if I post this question, whether tens of thousands of items about archaeological sites should be split, on the project chat to get some more feedback on this?

Yes, in fact. There is just ~200 archeological sites having dissolved, abolished or demolished date (P576) https://w.wiki/3f4P And only 24 sites being P31 both archeological site + old city https://w.wiki/3f4U Isn't it weird to see statements that that arch. site X was demolished XXXX years ago ? Isn't it weird to see statements that that arch. site X was founded XXXX years ago ?

Sure, that might be weird. But the problem there lies in those statements, and not in the fact that there's only one item about an archaeological site. If those statements were used, for example, as qualifiers to the statement instance of=city, then that would be much better already. So instead of splitting items off, it would be better to change those statements because many of them are just wrong in the way that they are implemented now.

I've taken the liberty of posting the general question on whether archaeological sites should be split, like you suggest, here: Wikidata:Project chat#Modelling archaeological sites

The city and the archeological site existed at different points of time. Splitting enitites in multiple items is standard procedure for Wikidata. In Wikipedia it frequently makes sense to mix multiple entities together.

In Wikidata it's very useful if items share formatting. Your suggestion means that it gets much harder to run a query for settlements that were dissolved before year X.

I would suggest that the problem is not in splitting city/archaeological site, but that the problem lies in the use of the statement "dissolved" (and the likes) to describe what happens at an archaeological site... But I would prefer to keep this discussion in one place, and I would suggest to make that place Wikidata:Project chat#Modelling archaeological sites instead of here. Thanks!