Newsgroups: comp.lang.dylan
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!news.sprintlink.net!demon!uknet!newsfeed.ed.ac.uk!dcs.ed.ac.uk!scgk
From: scgk@dcs.ed.ac.uk (Simon Kinahan)
Subject: Re: Sealing off a whole system & getting static type safety
Message-ID: <scgk.802207640@dcs.ed.ac.uk>
Sender: cnews@dcs.ed.ac.uk (UseNet News Admin)
Organization: Department of Computer Science, Edinburgh University
References: <DERWAY.95Jun1151036@alumni.ndc.com> <3qlphe$at9@cantaloupe.srv.cs.cmu.edu>
Date: Sat, 3 Jun 1995 19:27:20 GMT
Lines: 21

>In Dylan, as in Common Lisp, programmers can declare types to be as
>general (<object>) or as specific as they like.  If the programmer
>declares everything tightly, a compiler compiling a program for
>delivery should in principle be able to do as much type-inference,
>type-checking, and optimization as in the most strict of statically
>typed languages.

I am not sure about this at all. The most strict static languages
that do not force users to put a type on absolutely everything
are those that use type inference. Although the situation is 
changing fairly rapidly, these systems have problems with 
such things as subtyping and inheritance. I would be pretty
cautious about saying that even a sealed Dylan program can
be statically analysed to this extent.


-- 

Simon Kinahan            "Only in our dreams are we truly free,
scgk@dcs.ed.ac.uk         the rest of the time we need wages. "
scgk@tardis.ed.ac.uk               -- Terry Pratchett, Wyrd Sisters
