From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <2c45af7b985c5fcc1e3bcf51edbbd9e4@mightycheese.com>
To: 9fans@cse.psu.edu
Subject: RE: [9fans] g++
From: "rob pike, esq." <rob@mightycheese.com>
In-Reply-To: <81132473206F3A46A72BD6116E1A06AE479C66@black.aprote.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Date: Sun, 14 Sep 2003 09:13:14 -0700
Topicbox-Message-UUID: 34d21884-eacc-11e9-9e20-41e7f4b1d025

> There appears to be some nasty law of physics at work here: the simple and
> elegant O-O languages are easy to use right, but at the same time lack the
> single biggest trump of O-O -- generics; while the languages that contain
> generics are cumbersome and ugly. Go figure.

i agree.  o-o is so easy to do by hand in the few cases it really
contributes to the design of software that it seems unimportant to
have languages centered on it, yet polymorphism is a genuine boon but
implementing it using o-o so difficult you get monstrosities like the
STL or lose all compile-time type safety, as with all the casting
through Object in Java.  don't know much about the generics in the new
Java but i've heard a lot of grumbling, so i don't hold out much hope.

but here's the bad news: within this state of affairs, we see o-o
being accepted as the foundation of good design and we educate our
students in that point of view.  o-o has its moments, but using it
well is very difficult.  i think the pattern book is the test that
proves the rule; if that's the best you can say, your apprach is
wrong.  o-o is clumsy, a big hammer that is only occasionally the
right tool.  when all you have in your hand is the o-o hammer,
everything is a thumb.

how did we get here? i despair.

-rob