Re: Good book on Dylan?

To: info-dylan@ai.mit.edu
Subject: Re: Good book on Dylan?
From: "Mark Jordan" <mark_jordan@nospam.ieee.org>
Date: Sat, 31 Mar 2001 14:15:02 -0500 (EST)
References: <GQPw6.3041$Ph.129089@stones> <r5h8ctknt7pl0s2g607qn6lqiqritgao0j@4ax.com>
Reply-To: "Mark Jordan" <mark_jordan@nospam.ieee.org>
Xref: traf.lcs.mit.edu comp.lang.dylan:13175

"Jason Trenouth" <jason@harlequin.co.uk> wrote in message
r5h8ctknt7pl0s2g607qn6lqiqritgao0j@4ax.com">news:r5h8ctknt7pl0s2g607qn6lqiqritgao0j@4ax.com...
> On Thu, 29 Mar 2001 19:00:02 -0500 (EST), "Mark Jordan"
> <mark_jordan@nospam.ieee.org> wrote:
>
> > Also if you could tell me how Dylan might cope with my
> > interface/implementation problem I'd be most grateful.
>
> Well, a simple answer might be that multiple-inheritance is easier in
> Dylan. So you don't need to use object composition to avoid it.

I wasn't avoiding muliple inheritance, I just have an existing C++
black-box component library which uses it quite a bit. I could rethink
the design though if I translated it to Dylan.

<snipped nice example, thanks>

After my original post I realised that Dylan must also have similar
levels of indirection to C++ for interfaces and probably be no
more efficient, although I don't know for sure.

Does Dylan have a keyword or some other means to tell the compiler
that it can optimise all it likes because the object behind the interface
isn't going to change at runtime?  I suppose what I'm saying is that in
the majority of cases I only need compile-time polymorphism, not
full blown run-time polymorphism.

Thanks,
Mark.

Follow-Ups:

Re: Good book on Dylan?

From: Jeffrey Siegal <jbs@quiotix.com>

References:

Good book on Dylan?

From: "Mark Jordan" <mark_jordan@nospam.ieee.org>

Re: Good book on Dylan?

From: Jason Trenouth <jason@harlequin.co.uk>