Re: MI: why?

• To: info-dylan@ai.mit.edu
• Subject: Re: MI: why?
• From: "Eric Clayberg" <clayberg@instantiations.com>
• Date: Wed, 12 Jan 2000 23:00:02 -0500 (EST)
• Organization: Instantiations, Inc.
• References: <uaIe4.12653$Ce.289242@monger.newsread.com> <387BC10E.BED9F41F@miqs.com> <20000112004057.18210.qmail@gorn.americom.com> <387C849F.BC993358@dial.pipex.com> <387CD2A9.1D7DF8A@miqs.com>
• Reply-To: "Eric Clayberg" <clayberg@instantiations.com>
• Xref: traf.lcs.mit.edu comp.lang.dylan:11450 comp.lang.smalltalk:94305

Scott Ribe <sribe@miqs.com> wrote in message
387CD2A9.1D7DF8A@miqs.com">news:387CD2A9.1D7DF8A@miqs.com...
>
> It's interesting to note that the only OO languages without MI are
> either older, early generation OO (Smalltalk, Objective-C, Object Pascal)

Not true in the case of Smalltalk. Smalltalk is designed to easily support
MI and early implementations had it. It was ultimately rejected as adding
too much complexity and not enough benefit. You can even get MI as an add-on
package for Smalltalk, if you really want it. The reason it is not included
in any of the major commercial implementations today is that very few
customers are asking for it. Dynamic typing, unlimited polymorphism, simple
proxy mechanisms and a host of other features add up to minimize the actual
usefulness of MI in the Smalltalk context. That is not to say that MI would
not be very useful in other languages with different feature sets.

-Eric

• Re: MI: why?
• From: carlg@world.std.com (Carl E Gundel)

• A question
• From: "Stephen J. Guthrie" <steve.guthrie@mantissa.com>
• Re: A question
• From: Scott Ribe <sribe@miqs.com>
• Re: A question
• From: rwk@americom.com
• MI: why?
• From: jt <gkt37@dial.pipex.com>
• Re: MI: why?
• From: Scott Ribe <sribe@miqs.com>

• Prev: DUIM and OpenGL
• Next: Re: Future of Dylan?
• Index(es):
  • Main
  • Thread