Implicit context access

Brian Rice water at tunes.org
Mon Apr 11 08:45:53 PDT 2005


On Apr 11, 2005, at 3:50 PM, Lee Salzman wrote:

> I vote for "here". Changing the semantics of () in different 
> situations is too muddy, so
> I'd rather have it consistently return Nil everywhere.

Agreed. It's in the main repository now. I'm going to try to fix a few 
more problems (in other areas) and then put out a final 0.3.3 so more 
important code can get worked on.

> Brian Rice wrote:
>
>> Hi all,
>>
>> I realized we lack a way to generically refer to the implicit 
>> context, used for implicit-context message sends, and have a couple 
>> of ideas about that.
>>
>> Motivation: This would be helpful because we often want to define 
>> utility methods on a namespace, but we also have the &in: optional 
>> keyword for load: which should in principle allow us to load any code 
>> as though any other namespace were the lobby, but we have to hardcode 
>> lobby references.
>>
>> My initial idea was that in the context of method definitions, "_@() 
>> foo []." for example dispatches to Nil, but I had thought that maybe 
>> the surrounding namespace would be a better candidate.
>>
>> After further thought, this is really a special case and should not 
>> muddy the grammar of Slate too much. So, I'd prefer if we had a 
>> simple selector that worked anywhere, like thisContext (which seems 
>> to cause parse errors right now - I'll look into that) but more about 
>> accessing the namespace object than an activation.
>>
>> "thisNamespace" seems initially pretty good, but maybe there are 
>> better suggestions. Any ideas?
--
Brian T. Rice
LOGOS Research and Development
http://tunes.org/~water/




More information about the Slate mailing list