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

Re: MzScheme and SML




If you care about speed, compare SML with Chez Scheme. 

You will quickly find that static polymorphic typing `a la ML with modules
is as difficult to compile into fast code as compiling Scheme, if not much 
worse. Numerous places still have dissertations on this topic. A naive way
of compiling ML is guaranteed to be worse than Scheme. 

-- Matthias

   X-Authentication-Warning: fast.cs.utah.edu: majordom set sender to owner-plt-scheme@flux.cs.utah.edu using -f
   From: "Ji-Yong D. Chung" <virtualcyber@erols.com>
   Cc: <plt-scheme@fast.cs.utah.edu>
   Date: Sat, 17 Feb 2001 14:46:21 -0500
   Content-Type: text/plain;
	   charset="iso-8859-1"
   X-Priority: 3
   X-MSMail-Priority: Normal
   X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300
   Sender: owner-plt-scheme@fast.cs.utah.edu
   Precedence: bulk

   Hi, you wrote:

   > For a user who doesn't care about all these details, your remark

   >>     Is SML/NJ just a faster Scheme system
   >> with different syntax and compile-time type checking?

   > is a sufficient summary.

       Actually, I do care.  

       After I joined this mailing list,  I was
   so inspired messages here, that I simply
   had to implement my own Scheme interpreter in 
   c++ (with define-syntax expander in c++).  

       I started looking at MzScheme for insights into 
   my interpreter's performance bottlenecks.  Unfortunately,
   it seems that the real speed issues seem to be connected to
   the languages' underlying design.

       Had I known about ML, perhaps I would have 
   implemented that instead, as I am just a bit
   more concerned with efficiency than what Scheme's
   language design philosophy provides for.