search

dret / ASDL

Activity Stream Documentation Language (ASDL)
2 stars 1 forks source link

complete core vocabulary ASDL #1

Open dret opened 9 years ago

dret commented 9 years ago

currently only a very short version for some first experiments. should be completed according to the definition in http://activitystrea.ms/specs/json/1.0/

smnmyr commented 9 years ago

The "object" object type in https://github.com/dret/ASDL/blob/master/0.1/core-asdl.md is weird - http://activitystrea.ms/specs/json/1.0/#object defines this implicitly in the serialization section, but contains no further information about it... the core-asdl.md reflects this by duplicating the text from the activity object type, which is incorrect since the "object" is not an AS activity.

dret commented 9 years ago

thanks for noticing. first of all the text is definitely incorrect. but that's simply a problem of the code extracting the definitions from the spec. but more generally speaking, it seems that this is some abstract/generic object type that implicitly is the supertype of any other type. what would be the best way to reflect this, since ASDL (so far) has no model of a type hierarchy? leave it out and document the "object object type" as the implicit supertype of all types?

smnmyr commented 9 years ago

imo we can't do anything sensible at the moment, since it only "seems" that object is the generic supertype... is an "activity" object actually an "object"? ;-)

On Wed, Apr 8, 2015 at 10:24 AM, Erik Wilde notifications@github.com wrote:

thanks for noticing. first of all the text is definitely incorrect. but that's simply a problem of the code extracting the definitions from the spec. but more generally speaking, it seems that this is some abstract/generic object type that implicitly is the supertype of any other type. what would be the best way to reflect this, since ASDL (so far) has no model of a type hierarchy? leave it out and document the "object object type" as the implicit supertype of all types?

— Reply to this email directly or view it on GitHub https://github.com/dret/ASDL/issues/1#issuecomment-90980459.

dret commented 9 years ago

well, AS1 does not define or even have a clear or clean model, so this is a hard question to answer. but yes, an activity like any other object inside an activity is an object, at least that's how read it. it's a lot of contrived and unnecessarily complicated exegesis, but that's just because it's a badly written spec. but that's how it is. we cannot redefine AS1 and shouldn't attempt to do so. all we need for ASDL is a clear definition what kinds of things need to be documented, and how they fit into the AS1 framework. having an ASDL for AS1 itself, and in particular for the core, doesn't make a whole lot of sense to begin with, and it's something i mostly did to play around with it a bit and see where it works and where it doesn't. for this "possible implicit supertype" issue it probably doesn't work.

smnmyr commented 9 years ago

agreed :-) - so we ignore this issue...?

On Wed, Apr 8, 2015 at 10:59 AM, Erik Wilde notifications@github.com wrote:

well, AS1 does not define or even have a clear or clean model, so this is a hard question to answer. but yes, an activity like any other object inside an activity is an object, at least that's how read it. it's a lot of contrived and unnecessarily complicated exegesis, but that's just because it's a badly written spec. but that's how it is. we cannot redefine AS1 and shouldn't attempt to do so. all we need for ASDL is a clear definition what kinds of things need to be documented, and how they fit into the AS1 framework. having an ASDL for AS1 itself, and in particular for the core, doesn't make a whole lot of sense to begin with, and it's something i mostly did to play around with it a bit and see where it works and where it doesn't. for this "possible implicit supertype" issue it probably doesn't work.

— Reply to this email directly or view it on GitHub https://github.com/dret/ASDL/issues/1#issuecomment-90988914.

dret commented 9 years ago

my hope would be that more mature versions of ASDL would be more disciplined and explicit about this, for now 0.1 really is just a set of starting points to play around with. the "object" object type definitely is strange and https://github.com/dret/ASDL/commit/78e234ad36bfaa4395a0cd0e2b9b161a01cf2d8e removes it. maybe the other things from the core spec still make sense, or maybe just as an early experiment to be removed/changed later on.