Re: C# is not Dylan (was: Re: C# : The new language from M$)

To: info-dylan@ai.mit.edu
Subject: Re: C# is not Dylan (was: Re: C# : The new language from M$)
From: Brian Rogoff <bpr@shell5.ba.best.com>
Date: Sun, 2 Jul 2000 16:30:02 -0400 (EDT)
Organization: Verio
References: <m3zoo7w69a.fsf@cadet.dsl.speakeasy.net> <huvjls0t6hfio3htr0r1ee3bk57b0th1dr@4ax.com> <c29g0pxbubg.fsf@nerd-xing.mit.edu> <kpamlscdjl89ihfhsdk3d4rinoa1hji78j@4ax.com> <bruce-3485B7.01374830062000@news.akl.ihug.co.nz> <m2u2ebzfyf.fsf@gododdin.internal.jasmine.org.uk> <58rolsg17q7154fdf6dq5l038rbbnl10jt@4ax.com> <3171354053463029@naggum.no> <j7175.9050$Zr5.877270@typhoon.ne.mediaone.net> <3171366678290484@naggum.no> <395D0225.4FFEE339@rpal.rockwell.com> <wHo75.10794$Zr5.957339@typhoon.ne.mediaone.net>
Xref: traf.lcs.mit.edu comp.lang.dylan:12336

comp.lang.lisp removed, as this is irrelevant there
On Sat, 1 Jul 2000, Scott McKay wrote:
> OK, I'll tell you some of the things I like about Dylan, and some
> of the things I like better in Lisp.  First the likes:
> 
>  - I like Dylan's type system.  I like that the "built-in" types fix into
>    the type system in a clean and consistent way.  I like that you
>    can use it "typeless" but can then clamp down the types to get
>    better code, and more importantly, better compile-time error
>    detection.  Going back to Lisp makes me fell butt-naked on the
>    type safety front; as the static-type people have been saying
>    for years, a lot of errors really are type errors, and Dylan does
>    a good job catching these.

Interesting. I like statically typed languages too, just not the popular
ones 

>  - I am surprised to say that I have come to like the infix syntax.
>    For one, using type decls in Dylan is a lot easier than in Lisp,
>    and a lot more readable, too.  Don't delude yourself into thinking
>    Lisp has "no syntax" (as many people claim); it does, but it's
>    just harder to see.

People like what they get used to. I even know people who like Perl! I
find Dylan syntax close enough to a language I use and like (Ada 95) that 
I agree with you. If I were using Lisp or Scheme heavily I might be in the
other camp. In any case I agree with you that its not something to get
hung up on. I also agree with Frank Adrian that programming languages are 
*NOT* like spoken (or signed :-) human languages. If you read Pinker, I'm
surprised you haven't reached the same conclusion.  

>  - I like Dylan's hygienic macro system.  If we adopted Keith
>    Playford's work on procedural macros, Dylan's macros would
>    be very nearly as powerful as Lisp's, and many of Lisp's
>    pitfalls would fall away.

Could you provide a pointer to more information on this? I haven't been 
able to find documentation on it at all. Are either of the two
implementors considering adding this to their implementations? 

... snip...

>  - Believe it or not, I now prefer the simpler method combination
>    in Dylan. 

I recall reading some time ago about a slightly "better" system for 
Dylan which is pretty close to what is there now. Maybe if there is a 
new version of Dylan...

> Now some dislikes:
...snip...

My biggest gripe right now is the lack of implementations. The commercial 
Dylan is for an OS I don't care about, and the free one is a bit too
heavyweight. I'd prefer a "tractable" Dylan implementation for
experimentation, even if the code it generated were an order of magnitude
slower than C. As an example of what I mean, consider the OCaml compiler, 
which is relatively small, and compiles on every Unix out there with
minimal effort, and generates very good code! I'm not criticizing the
Gwydion developers here, its just that they have very different
priorities.

-- Brian

Follow-Ups: