[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: low tech MI

>    Date: Mon, 22 Apr 96 13:27:52 BST
>    From: Jeff Dalton <jeff@aiai.ed.ac.uk>
>    > MI is controversial [in the world of object-oriented programming]
>    > chiefly because it introduces a lot of complexity (both semantic and
>    > implementational) into method lookup.

>    I'm not sure method lookup is the greatest problem.  A number of MI
>    schemes have general MI method lookup but do something special, with
>    restrictions, for slots....
> True.  I thought about mentioning that when I composed my previous message,
> but I decided that since that was primarily a -performance- problem, and I
> couldn't imagine anybody arguing that accessing the slots of a condition
> had to be single-memory-reference fast, I could get away with not
> mentioning it. 

I don't think it is primarily a performance problem, though that's
certainly a factor and sometimes the most important one.  But
questions encapsulation and inheritance semantics are often more

For whatever reasons, the disputes seem less when it comes to methods.

> Law number one of electronic mail design discussions must
> be:

Perhaps the answer is to never reply to anyone unless the reply
is directed solely to what was said in the message.

BTW, I suspect you are right about the implementors meeting.
But if they produce a revised language, perhaps it would
reduce friction if they called it something other than "Scheme"
(though the name might include "Scheme").

In any case, some people may find it fun to invent names the
implementors could use.  The best example along those lines
that I know occurred in the Franz Lisp mailing list back when
the T-style names now used in Scheme were more controversial.
When someone wrote that T was so beautiful that it brought
tears to your eyes, there came a reply that T should have
been called ONION for "ONION is Not Its Original Name".

-- jd