Re: result of an Action

On Tue, Nov 18, 2014 at 7:37 AM, ☮ elf Pavlik ☮ <
perpetual-tripper@wwelves.org> wrote:

> Hello,
>
> I noticed updates to RsvpAction in sdo-venkman announcement:
> http://sdo-venkman.appspot.com/RsvpAction
>
> Earlier I suggested reconsidering what information we reference directly
> from Action and what from it's result:
> http://lists.w3.org/Archives/Public/public-vocabs/2014Oct/0137.html
>
> Currently we can find couple of IMO clear examples where an Action
> yields certain result:
>
> Action                 result
>
> PhotographAction       Photograph
> WriteAction            Article
> FilmAction             Movie
> PaintAction            Painting
> ReviewAction           Review
> CommentAction          Comment
>
> In case of RsvpAction, we add two properties directly rsvpResponse and
> additionalNumberOfGuests. At the same time leaving Reservation and
> Ticket not related to RsvpAction in any way. Maybe we could use
> Reservation as a result of RsvpAction?


I think these (i.e. having RsvpAction.rsvpResponse and
RsvpAction.additionalNumberOfGuests VERSUS RsvpAction.result = Reservation)
are not mutually exclusive, but rather they are equally valid.

I'm thinking of rsvpResponse and additionalNumberOfGuests as "adverbs",
that is, modifiers of the manner in which you are rsvp-ing. Hence, why they
started as a property of RsvpAction.

Similarly, you could also possibly/optionally have a Reservation to be
created as a result of RsvpAction, which you would use RsvpAction.result
for.

I don't think Rsvp-ing necessarily implies the creation of a Reservation or
a Ticket.

Am I making sense?


> Can we maybe write an example of
> scenario where RSVP YES to an event requires getting a ticket? We also
> introduce in latest update expectsAcceptanceOf which could help here.
>
> CommentAction gives another example where I see current design rather
> unclear. Currently included example:
>
> {
>   "@context": "http://schema.org",
>   "@type": "CommentAction",
>   "agent": {
>     "@type": "Person",
>     "name": "John"
>   },
>   "object": {
>     "@type": "UserComment",
>     "name": "That's cool!"
>   },
>   "about": {
>     "@type": "ScholarlyArticle",
>     "name": "We found that P = NP!"
>   }
> }
>
> I would write instead:
>
> {
>   "@context": "http://schema.org",
>   "@type": "CommentAction",
>   "@id": "http://example.org/people/jogn/log/12342342",
>   "agent": {
>     "@type": "Person",
>     "@id": "http://example.org/people/john",
>     "name": "John"
>   },
>   "object": {
>     "@type": "ScholarlyArticle",
>     "@id": "http://example.org/articles/p-equal-np",
>     "name": "We found that P = NP!"
>   },
>   "result": {
>     "@type": "Comment",
>     "@id": "http://example.org/articles/p-equal-np/comments/1"
>     "about": "http://example.org/articles/p-equal-np",
>     "author": "http://example.org/people/john",
>     "text": "That's cool!"
>   }
> }
>

Yep, I agree with you.

Your example is a better representation than what we have.

Git pull request?

On a related note, deprecreating the entire UserInteraction sub-tree was in
one of the early proposals for Actions but I think it got lost somehow.


>
> This way an Article can simply include comments (which can also link to
> the action they result from):
>
> {
>   "@context": "http://schema.org",
>   "@type": "ScholarlyArticle",
>   "@id": "http://example.org/articles/p-equal-np",
>   "name": "We found that P = NP!"
>   },
>   "comment": [
>     {
>       "@type": "Comment",
>       "@id": "http://example.org/articles/p-equal-np/comments/1"
>       "about": "http://example.org/articles/p-equal-np",
>       "author": "http://example.org/people/john",
>       "text": "That's cool!",
>       "@reverse": {
>         "result": "http://example.org/people/jogn/log/12342342"
>       }
>     }
>   ]
> }
>
> I also notice that in many cases agent in an Action becomes author in
> this action's result, similar object in an action becomes about in result.
>
> Once we clarify general convention on using results of Actions, I could
> help with writing more concrete examples which would provide clear
> reference.
>
> Cheers!
>
>