Re: [GD-General] Eiffel

[Patrick M D. <pa...@wa...>]

On Sun, 23 Dec 2001, Brian Hook wrote:

> At 11:20 PM 12/23/2001 -0500, Patrick M Doane wrote:
> >Perhaps it would be interesting to come up with a few small programs or
> >algorithms and see how they are implemented in various languages.  That
> >might make learning these languages easier.  Any thoughts?
>
> Actually, that's been done before (sort of) in the Great Computer Language
> Shootout.  A pretty nifty article that covers the performance aspects.

I was thinking more from the perspective of what might be useful to game
developers.  The programs in Doug's shootout are way too small.  Here are
some examples of the scope I was thinking about:

  - MUD server
  - ray tracer
  - AI algorithms, e.g: best paths, neural networks

Some of these could obviously get complex, but there should be simple
implementations to make all of these approachable.

> The problem is that small programs don't tell you that much.

Absolutely - that's why I was thinking of slightly more complicated
examples.

> At a high level, there is C vs. C++ vs. Obj-C vs. Java vs. Eiffel, and once
> again a small program doesn't necessarily take into account.

I'm not sure I understand what you're saying here.

> Just reading about Eiffel taught me a lot (both about it and C++), and
> it wasn't until I started getting into the grittier details that some
> of the warts started to show up.  I still think it's a cool language,
> but for games specifically it may not be ideal (no real bit twiddling,
> no hex constants) without calling out to C/C++ constantly to do real
> work.  The environments and libraries also have their own set of
> issues.

If issues like bit-twiddling and hex constants are keeping you from using
Eiffel, then use a preprocessor to begin with, and if these functions
still seem very important, extending the language should be easy
enough.

Environments and libraries are a little harder to deal with of course.

> That's why listening to informed opinions is fascinating to me, and
> unfortunately real good comparisons are hard to come by.  I'm still very
> impressed by Ian Joyner's relatively objective analysis of Java vs. C++ vs.
> Eiffel in "Objects Unencapsulated".  There's another book I just ordered
> that talks about OOP in terms of C++, Obj-C, Object Pascal and, I believe,
> SmallTalk.

You also might find "A Theory of Objects" by Abadi & Cardelli to be
interesting.  The first 5 chapters review object oriented features and do
a good job of separating concepts like inheritance and subsumption.  The
authors are also very aware of the wide variety of OOP languages that
exist -- at least 20 different programming languages are explicitly
mentioned in the text.

Note that the majority of the book focuses on developing a theory of
objects that could serve as a foundation for all OOP languages.  In
effect, they are trying to establish what lambda calculus did for
procedural languages.  So, I would really only recommend reading it if you
are interested in learning more about progamming language theory.

Patrick