Re: Implicit vs explicit object types

Amy: excellent analysis. Great to see.

elf, Amy's post is quite good and I appreciate the enthusiasm but I'm
not sure there's really anything in Amy's post that needs to be
discussed on the call. We have lots of issues already that need to be
worked through. I would encourage everyone to read what Amy has put
together as well as take the time on their own to do similar analysis.
That would certainly help matters.


On Mon, Apr 13, 2015 at 1:39 AM, ☮ elf Pavlik ☮
<perpetual-tripper@wwelves.org> wrote:
> On 04/12/2015 10:55 PM, Amy G wrote:
>> Hi all,
> Hi Amy,
>
>>
>> I've been mucking about with the data model for my blog (which is a weird
>> linked data / indieweb hybrid) on the way to eventually emitting AS2, and
>> as a result have started to understand (I think) the implicit vs explicit
>> types debate. I wrote a post:
>> http://rhiaro.co.uk/2015/04/post-and-activity-types
> Thank you for dedicating your time to such solid analysis of this issue!
> I just added it to the agenda for our next meeting.
>
>>
>> tl;dr - I think each side of the debate about explicit vs implicit types
>> are misaligned, due to conflation of Objects and Activities (as elf noted
>> <https://github.com/jasnell/w3c-socialwg-activitystreams/issues/94>). I
>> think Objects (like blogposts) needn't have explicit types, but Activities
>> (like the act of creating a post) should do. Rationale in-post.
> I can agree that Objects (like BlogPostings) may not need explicit
> specific types, still generic one like mf:h-entry or sioc:Item wouldn't
> hurt. But other Objects (like Person, Event, YogaWorkshop, Place,
> Toilet, Device, Vehicle) possibly better can have explicit type. IMO it
> may look little different when we describe *online content* and we
> describe *physical things*.
>
> I would also like to bring to our attention that every Thing can have
> *multiple types*. So sometimes it might make sense to use
> { "@type": ["mf:h-entry", "sioc:MailMessage"] }
> To my understanding types state inclusion in some particular set of
> things and come pretty useful when we want to query / filter dataset we
> work with. Even more for those who use reasoning. So when asking for all
> { "@type": "Fruit" }, we also get in result things which have only
> explicit { "@type": "Apple" }
> :Apple rdfs:subClassOf :Fruit .
>
> Use case: analyze your diet (fruit vs. sweets, juice vs. coffee)
> * http://rhiaro.co.uk/llog/
> * http://openfoodfacts.org/
>
> Also since we work with *graphs* we don't need to decide on single path,
> and each node can also have in a way multiple "@id" (reconciled with
> owl:sameAs) e.g.
> * https://example.net/photos/139819319
> * https://example.net/notes/juggling-in-portland
> both can *denote* exactly the same resource.
> { "@type": ["schema:Photograph", "sioc:Post"] }
>
> Example - not exactly such case but in many ways similar with single
> sioc:MailMessage in two sioc:MailingList archives.
> https://www.w3.org/mid/55278428.8090109@wwelves.org
>
> Side note: I consider publishing a page for each message I've posted to
> public mailing lists and link to copies available in each list archive
> via rev="canonical" (*reverse* not rel).
>
>
>>
>> Accepting responses by email, webmention or IRC :)
> Maybe in some near future inline via annotator like one enabled on
> http://www.w3.org/TR/annotation-model/
> ?
>
> I will copy few excerpts from your blog into email and reply 'inline' in
> next days.
>
> Once again, your posting comes very helpful to discuss this issue in
> depth. Thank you for dedicating your time to such solid research and
> write up!
>
> rhiaro++ :)
>
>

Received on Monday, 13 April 2015 15:09:50 UTC