Property talk:P291

qualifiers too?

This is a great idea. And that graph is fantastic! I wonder if we should add something similar "qualifiers for this property". I would for example add named as (P27) and source (Q1117) as qualifiers typically added under collects (P10). Thoughts @Podehaye: @Hanz0mon:? -- Alexbfree (talk) 15:47, 11 February 2020 (UTC)

@Alexbfree: yes, we should seek to extend this technique to have the expressivity you are seeking with the page Project:Vocabulary/GDPR.

discussion about source (Q1117)

@Alexbfree: one comment though: you use source (Q1117) as a qualifier, but I think you mean a property, no? (also on Project:Vocabulary/GDPR). Podehaye (talk) 16:24, 11 February 2020 (UTC)

@Podehaye: erm, well, it is showing the source of this specific collects (P10)-snak, so I think that makes it a qualifier, right? A property, that is being used as a qualifier? I think some properties (such as comment (Q1048)) could apply either as a qualifier on a snak or as a property on an Item - whereas some properties like source (Q1117) would apply typically only as a qualifier upon a snak. Right? If I am using concepts incorrectly, please let me know - it's certainly possible I have the wrong mental model! Alexbfree (talk) 17:00, 11 February 2020 (UTC)

@Alexbfree: you got the mental model right, I think, but didn't notice that the comment (Q1048) has a Q-number. It's an item, not a property!

@Podehaye: Let me check I got this right:

• Snak = Property + Value (where value could be an Item, Property, text or something else).
• A snak directly upon an Item could be considered a property of that Item (a 'property snak', let's say)
• A qualifier would be a snak applied to a 'property snak'. A snak of an Item property.

Is this right? Alexbfree (talk) 17:00, 11 February 2020 (UTC)

@Alexbfree: yes, except some sharpening and in bold the important bit: what you call a 'property snak' applies to an item to form together a 'statement'. What you call a 'qualifier' should for consistency be called a 'qualifier snak'. But that qualifier snak is not applied to a property snak, despite appearing so visually. To be more precise a qualifier snak is applied to the full statement.

comments related to querying/graphing and the ultimate merging into a common model

@Podehaye: I was just thinking about this some more. The ability to list and graph usages/related concepts is very powerful. I think something we could do soon is generate a query that lists all the objects of collects (P10) snaks on ALL instances of data controller (Q96). If possible gathering both the direct object text of those collects (P10) properties - but also the named as (P27) object text of all collects (P10) properties as well. With this list, once Merge is fixed (maybe it already is?) we should be able to do some serious merging of related concepts (though we should be careful to add named as (P27) qualifiers where they are absent so that no generated GDPR letter gets its output changed by the merging... Alexbfree (talk) 17:05, 11 February 2020 (UTC)

@Alexbfree: Yes. To generalize what you just said: we start with a class C, look at a property P we expect instances of C to have present, and look at the values V. Then we start taxonomizing the V. I think this pattern will be present quite a few times, and most useful indeed for this use case of

?item instance of (P3) data controller (Q96); 
      collects (P10) ?value. 
?value instance of (P3) ?unknown.