Re: Interfaces

• To: <info-dylan@ai.mit.edu>
• Subject: Re: Interfaces
• From: Rob Myers <robmyers@cwcom.net>
• Date: Sun, 02 Jul 2000 13:48:48 +0100
• In-Reply-To: <IDx75.11213$Zr5.996872@typhoon.ne.mediaone.net>
• Sender: owner-info-dylan@ai.mit.edu
• User-Agent: Microsoft-Outlook-Express-Macintosh-Edition/5.02.2022

When I mentioned protocols, I was thinking of forward-iteration-protocol,
which of course isn't any sort of runtime introspection, but works well in
some, well-defined situations where arbitary interfaces would otherwise be
used. 
You do still have a slight bootstrapping problem in finding the protocol
method in the first place, though. :-)

- Rob.

> From: "Scott McKay" <swm@mediaone.net>
> Subject: Re: Interfaces
> 
> But the reason I don't like it, is that it's not really first class.
> Dylan modules are compile-time objects, and you can't ask
> anything about them at run-time.  That is, you can ask a class
> what methods or superclasses it has.  You can ask a G.F.
> what methods it has.  But you can't ask this sort of an interface
> what GFs are in it.  I think this sort of introspection could be
> useful in a number of ways in real-worl applications.

• Re: Interfaces
• From: "Scott McKay" <swm@mediaone.net>

• Prev: Re: C# is not Dylan (was: Re: C# : The new language from M$)
• Next: Re: C# is not Dylan (was: Re: C# : The new language from M$)
• Index(es):
  • Main
  • Thread