Subject: Re: modular statically typed multimethods
From: Craig Chambers (chambers@cs.washington.edu)
Date: Fri Jan 07 2000 - 11:25:48 PST
On Fri, 07 Jan 2000 17:30:42 +0530, mnaik@hss.hns.com wrote:
>
>
>
> Hello
>
> As mentioned in the paper "Modular Statically Typed Multimethods",
> the open objects relieve the tension between ease of adding
> operations to existing classes and ease of adding subclasses.
>
> What I understand by "open objects" is the prototype-based model.
No, openness has nothing to do with prototypes vs. classes. "Open
classes" work just the same way in the class-based world as "open
objects" do in the prototype-based world. The key feature is being
able to add new generic functions (a.k.a. methods) to existing
objects/classes from the outside of the original declaration of the
object/class.
> While I appreciate the expressive power of this model, I would
> like to bring to your notice Robert Martin's "Acyclic Visitor"
> (also called RTTI visitor) design pattern
>
> (http://www.objectmentor.com/publications/acv.pdf)
>
> which allows both, adding operations to existing classes and adding
> subclasses, in the class-based model. It is more powerful than the
> visitor design pattern which allows only the former.
>
> Regards
>
> -- Mayur
>
>
Thanks for the pointer. I'll take a look.
-- Craig Chambers
This archive was generated by hypermail 2b25 : Tue Oct 03 2000 - 15:21:19 PDT