#social 2016-12-31

2016-12-31 UTC
bengo joined the channel
#
csarven bengo
#
bengo hello
#
bengo csarven
#
csarven Thanks for testing! Two things: 1) the inbox tooling (Solid) is having problems with JSON-LD serialisation. I'll use mayktso for the inboxes (or at least tests) soon. Going to swap that.. 2) dokieli doesn't yet understand as:Create.. it supports some stuff but not all Activities.
#
csarven So, if you get an error when you try to retrieve the notification from a Solid inbox (Wher eyou've provided JSON-LD) take that as temporary. The dokieli bit is something else and I can update that separately.
#
bengo Good to hear about #2. I was discussing that with rhiaro last night.
#
bengo #1 shouldn't be a problem as what I'm currently working on won't try to retrieve notifications from inbox, should send them
#
csarven re: http://json-ld.org/playground/#startTab=tab-nquads&json-ld=%7B%22%40context%22%3A%22https%3A%2F%2Fwww.w3.org%2Fns%2Factivitystreams%22%2C%22type%22%3A%22Create%22%2C%22object%22%3A%7B%22%40context%22%3A%22https%3A%2F%2Fwww.w3.org%2Fns%2Factivitystreams%22%2C%22type%22%3A%22Note%22%2C%22content%22%3A%22Test%22%2C%22cc%22%3A%5B%22https%3A%2F%2Fwww.w3.org%2Fns%2Factivitystreams%23Public%22%2C%22https%3A%2F%2Flinkedresearch.org%2Fldn%2Ftests%2F%22%5D%2C%
#
csarven 22uuid%22%3A%224d72c129-38c6-437c-9166-2a6a0fb72162%22%2C%22replies%22%3A%22%2Factivities%2F4d72c129-38c6-437c-9166-2a6a0fb72162%2Freplies%22%7D
#
csarven 22inReplyTo%22%3A%22https%3A%2F%2Flinkedresearch.org%2Fldn%2Ftests%2F%22%7D%2C%22cc%22%3A%5B%22https%3A%2F%2Fwww.w3.org%2Fns%2Factivitystreams%23Public%22%2C%22https%3A%2F%2Flinkedresearch.org%2Fldn%2Ftests%2F%22%5D%2C%22id%22%3A%22urn%3Auuid%3A4d72c129-38c6-437c-9166-2a6a0fb72162%22%2C%22published%22%3A%222016-12-31T12%3A10%3A21.987Z%22%2C%22inbox%22%3A%22%2Factivitypub%2Finbox%22%2C%22url%22%3A%22%2Factivities%2F4d72c129-38c6-437c-9166-2a6a0fb72162%22%2C%
#
csarven You might want to use @id for the as:object, otherwise it is a blank node.
#
csarven plus, it helps to know that resource exists/resolveable
#
bengo Good idea, adding to TODOs
#
csarven You don't need to assign a uuid for the as:Create subject either. Use @id: "" to have it refer to the base URL.
#
csarven But maybe you can say more about what the notification is intended to be.. I see some cc's.. I'm guessing what it does but I coul dbe missing your actual intentions
#
bengo It's basically "Ben created a note over here that's inReplyTo your resource"
#
bengo I'm intending to implement this https://w3c.github.io/activitypub/#create-activity-inbox
#
csarven Would you mind creating https://github.com/linkeddata/dokieli/issues/ mentioning 1) what the notification would contain 2) what data to expect inside the external resources
#
bengo No I wouldn't mind
#
bengo FWIW I found another minor thing while testing just now - https://github.com/csarven/ldn-tests/issues/15
#
csarven thanks
#
csarven Yea, it'd be great, then I can implement that on dokieli. rhiaro also wants some of her notifications/articles to show up properly.. I need actual examples.
#
bengo Makes sense. FWIW this is an as:Collection of my examples http://distbin.com/public
#
bengo But I will file issue too
#
csarven I think the outer double quote is giving you the issue.. see my comment on issue 15
#
bengo ohno
#
bengo ok
#
bengo Ugh you're right. I had copypasted from the other place I forgot that quote
#
bengo sry
#
csarven Happens.
#
csarven BTW, I don't think there is a uuid property in AS2
#
csarven if you just want to use that as an identifier of some sort consider dcterms:identifier, bibo:uri or someting
#
csarven dcterms:identifier is probably more appropriate since the value is not really a URI any way
#
csarven could also get away with rdf:value
#
csarven personally I wouldn't bother with it. the urn:uuid that you have is good enough (if you don't want to or need to give a resolveable identifier)
#
bengo csarven Good to know about dcterms:identifier - I put it in there because I got sick of parsing it out of the id URI. It was a lazy move. Will consider moving or at least identifirying as dcterms:identifier
#
csarven then why not use the whole urn:uuid as the identifier not just the uuid.v1 or whatever
#
csarven the string that's generated from the uuid is not particular interesting or unique (or clear). having the scheme adds a tiny bit more value IMO
#
bengo Just grepped and I only use it in one place that I can do a diff way
#
bengo So will remove
#
bengo the issue you requested, FWIW https://github.com/linkeddata/dokieli/issues/166
#
csarven BTW, dokieli has a similar UC to distbin such that if the document it is on has a reference to an oa:annotationService, and if the user is signed-in, what happens when they annotate the article is that, thee body of the annotation gets to be saved at their personal storage (being the canonical copy), and another copy is sent to the annotationService.
#
csarven Thanks! Reading.
#
csarven Will implement.
#
bengo oooh that's pretty cool
#
bengo What do you mean by "UC"?
#
csarven See the screencast at https://dokie.li/#annotate
#
csarven Use Case
#
bengo ah got it
#
bengo Was thinking today about how I should impl Web Annotation Protocol next/soon
#
bengo csarven Do you know if/how there is any equivalence between sioc:has_reply and oa:replying ?
#
bengo like an inference rule... x hasMotifivation Replying & x target parent -> x sioc:reply_of parent
#
bengo Eh I will ask in #annotation
#
csarven I don't think there is a clear rule. oa:replying is a concept, and used in context of a qualified relation i.e., it makes sense as a whole.
#
csarven Annotations are in a way capture the annotation activity by connecting x and y, with all the reasons and precision as to what/how it was done.
#
csarven sioc:reply_of captures only that x points to y
#
csarven For example, a 3rd party can create an annotation which describes the relationship that's made between the parts of x and y.
#
csarven While the annotation approach is verbose, it gives a lot of granularity and allows 3rd parties to create new knowledge.
#
csarven sioc:has_reply case is useful if the owner (the subject) makes the claim.
#
csarven the inference rule you came up with is pretty close
timbl joined the channel