From Wikidata
Jump to navigation Jump to search

@Bovlb: regarding change see discussion self depicting or not it is used in queries like tweet when we generate #IIIF frm Wikidata.... - Salgo60 (talk) 06:12, 14 March 2019 (UTC)

@Salgo60: Thanks for the link to the discussion, but I have to agree with VIGNERON (talkcontribslogs) here that these items are generally not self-depicting, and that it is the wrong representation to confuse, say, a statue with an arbitrary photograph of the statue. If we want to say what a (Commons) image depicts then that is information about the image. It is meaningless to say that the statue depicts itself or nearby buildings, or that these depictions can be located in a 2-d co-ordinate system. Cheers, Bovlb (talk) 16:38, 14 March 2019 (UTC)
@Bovlb: I agree but we need to get structured data in commons.... After doing a training this monday using IIIF I feel we need to redesign more things to support WD objects with more pictures and more annotations on one picture with different purpose... maybe a WikiAnnotation layer is needed - Salgo60 (talk) 11:56, 15 March 2019 (UTC)
@Salgo60: That all sounds like a good idea, but in the meantime we need to avoid adding false claims to Wikidata just because there is no representation available for the facts we are trying to express. Cheers, Bovlb (talk) 15:28, 15 March 2019 (UTC)
@Bovlb: if it false or not is if you relate depicts (P180) to the image (P18) of the WD object or not. If you have a statue and that statue has a picture then if you find a depict is that depict related to the picture of the statue or just about he object itself? Having a unmature Open platform as Wikidata you run into things like this... by using relative position within image (P2677) then it relates to a specific picture... i.e. it must be the objects image (P18)?!?!? - Salgo60 (talk) 15:47, 15 March 2019 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── @Salgo60: depicts (P180) should be about what the object itself depicts, not (one of) the current image (P18)(s). It does not make sense to say that a statue depicts itself, nor to say that it depict the building next to it; these are facts about a specific image. And it doesn't make sense to give 2-D co-ordinates of depictions within an object unless that object is inherently two-dimensional (and orientable), and we can therefore construct a canonical image. What's going to happen if someone replaces the image, or adds a second? Bovlb (talk) 16:42, 15 March 2019 (UTC)

@Bovlb: The replacing problem you always have with relative position and therefore we need to wait for structured Commons is my guess - Salgo60 (talk) 17:09, 15 March 2019 (UTC)