Re: do types influence behavior?

On 2015-02-11 01:45 PM, Erik Wilde wrote:
> hello henry.
>
> On 2015-02-11 9:28 , henry.story@bblfish.net wrote:
>> My guess is that this is what used to be grouped under "Developer 
>> Stories".
>> It is something along the lines:
>> 1. Conchita develops an Android client
>> 2. During development she notices that people sometimes miss out 
>> data, so to give a better user experience she builds very light 
>> weight RDFS inferencing into her client to fill in the missing data
>
> unfortunately we cannot push this into the client. you as somebody 
> consuming AS from some AS ecosystem want to know what will be 
> delivered to you if you ask for "respond" activities. there are 
> different ways how we can answer this question (including "all kinds 
> of things can happen, don't count on the behavior or your AS ecosystem 
> to be consistent or predictable"), but we have to answer it.
pump.io, for example, keeps a collection of people who have "liked" an 
object. Activity Streams 1.0 doesn't have a way to describe hierarchies 
of activity types, so it only looks for actions with the "like" verb (or 
its synonym, "favorite"). Actions with other kinds of verbs are passed 
along through the system, but users who "superlike" an object don't get 
into the "likers" list.

I think this is a case where the advantage to the creator ("You can make 
up your own kind of 'like'!") is outweighed by the disadvantage of all 
the downstream consumers who'll have to figure out whether a 'floop' is 
really a 'like' or something else.

I doubt I would bother with this. I'd probably just be really irritated 
with the 'superlike' developers, if it ever came up.

-Evan

Received on Wednesday, 11 February 2015 18:54:24 UTC