RE: [AS2] Deprecate "url" in favor of "id" and "self" link relations

On 25 Aug 2014 at 19:47, James M Snell wrote:
> On Mon, Aug 25, 2014 at 6:05 AM, Markus Lanthaler wrote:
>> On 21 Aug 2014 at 18:16, James M Snell wrote:
>>> In an activity streams document, the "url" property is typically used
>>> to point to a displayable (typically HTML) representation of an
>>> object. The "id" property is typically used to provide just an
>>> identifier. With AS2's direct incorporation of IANA Link Relations, it
>>> makes more sense to deprecate the use of "url" in favor of link
>>> relations such as "self" or "alternate". Doing so simplifies things a
>>> bit more and gives us better alignment with things like JSON-LD.
>> 
>> Why do you think the tokens "self" or "alternate" give "better
>> alignment with things like JSON-LD"? 
> 
> My apologies, I was not clear enough on this point. This had to do with
> how Link Value's were defined, not with the "self" or "alternate" link
> relations in general. Previously, there was ambiguity in the way Link
> Value's used either url or id. That ambiguity has been cleared up.
>
>> It is interesting to note that Schema.org (likely the most widely used
>> vocabulary) has a "url" [1] property which matches quite well with what
>> "url" in AS1 was supposed to do. In fact, I think it is the better match
>> than IANA's "self". 
> 
> Based on the feedback, I've taken another pass on this that does not
> deprecate url, but goes through an specifically defines it's
> relationship to "id", "self" and "alternate". While it's just a first
> pass and should still be considered a draft, I believe the new
> language works better:
> 
> See: http://jasnell.github.io/w3c-socialwg-activitystreams/activitystreams2.html#object-links

Hmm... so we now have

   id = @id
   duplicates (points to ids)
   self (dereferenceable IRIs)

   url ("preferred" alternate)
   alternate (more alternates)

   multiple "embedded links"

Sorry, but this sounds overly complex to me. Could you please explain briefly what use cases this design is trying to address?


Thanks,
Markus



--
Markus Lanthaler
@markuslanthaler

Received on Thursday, 4 September 2014 10:40:49 UTC