Adaptors, was Re: [Reality] getting started?

Jasper Phillips reality@twistedmatrix.com
Wed, 27 Aug 2003 23:02:38 -0700 (PDT)


On Thu, 28 Aug 2003, Glyph Lefkowitz wrote:

> 
> On Wednesday, August 27, 2003, at 06:21 PM, Jasper Phillips wrote:
> 
> > How do you deal with such methods?  Do you place them on what needs them
> > and call them via reflecteion, or do you extend the adaptors to keep
> > track of all such complex interactions (if so what about code dealing
> > with interactions between 2 adaptors)?
> 
> In theory, extend the adapters.  I'm not sure how far this approach will
> scale, but I think it's safe to say it will scale more than any previously
> attempted approach :).

Heh, fair enough.  I certianly haven't seen a better way to handle such a
hard problem.

This is pure conjecture, but I wonder if some vague sort of "meta" adaptor
arrangement might help.  Something to formalize interaction between adaptors,
e.g. adaptors themselves registering with other adaptors, style rules for
composite adaptors, etc.  This sort of problem feels like a prime source
of complexity, and worth considering.

I have only a fuzzy notion of what this might be like of course.

> After all, I don't think it's possible to automate an interaction like 
> "how does burning interact with the function of a rope vs. the function 
> of a loudspeaker vs. the function of a book?"

Fortunately one probably doesn't need to go so far to have an interesting
environment. ;-)


On a side note, do you consider Twisted Reality to be a pet project, or
part of some long term goal?  I.e is it just for amusement and introspection,
or do you have plans to use it seriously?

-Jasper