Newsgroups: comp.lang.misc,comp.object,comp.lang.scheme,comp.lang.functional,comp.lang.dylan
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!oitnews.harvard.edu!purdue!lerc.nasa.gov!magnus.acs.ohio-state.edu!math.ohio-state.edu!howland.reston.ans.net!ix.netcom.com!netcom.com!NewsWatcher!user
From: hbaker@netcom.com (Henry Baker)
Subject: Re: (debunk-myth '(debunk-myth '(<-efficient gc malloc/free)))
Message-ID: <hbaker-0411951040540001@10.0.2.15>
Sender: hbaker@netcom22.netcom.com
Organization: nil organization
References: <QOBI.95Oct22061418@qobi.ai> <46dm26$if4@jive.cs.utexas.edu> <QOBI.95Oct22124417@qobi.ai> <DH0GMA.A9w@Cadence.COM> <46n129$ssg@goanna.cs.rmit.EDU.AU> <rward-0411952019270001@ou048030.otago.ac.nz>
Date: Sat, 4 Nov 1995 18:40:54 GMT
Lines: 33
Xref: glinda.oz.cs.cmu.edu comp.lang.misc:23654 comp.object:40214 comp.lang.scheme:14249 comp.lang.functional:6634 comp.lang.dylan:5812

In article <rward-0411952019270001@ou048030.otago.ac.nz>,
rward@random.otago.ac.nz (Roy Ward) wrote:

> In article <46n129$ssg@goanna.cs.rmit.EDU.AU>, ok@goanna.cs.rmit.EDU.AU
> (Richard A. O'Keefe) wrote:
> >     There is a language called Clean; version 1.0 is available for the Mac.
> >     It is a very nice lazy functional language similar to Haskell and Gofer
> >     BUT they have "uniqueness types" and an I/O system built on top of that
> >     concept.  If for example I write
> [SNIP]
> >     The use of uniqueness types permits descructive update, and in
particular
> >     there is a unique "world" which I/O updates.  Also, when a
reference to a
> >     unique value dies, the value can be reclaimed.
> 
> There is also a fledgeling language for the Mac (ReWrite) that _only_ has
> unique types. This means that no GC is required at all. I am biased (being
> the author) but have found that performance is not too bad (considering
> that the current version is purely dynamically typed) _except_ that
> I have to be careful not to have more than one reference to a big data
> structure unless I want multiple copies (Later versions may allow
> 'non-unique' types but I would prefer to avoid GC for as long as possible).

You might also look at the following paper on linear/unique types:

ftp://ftp.netcom.com/pub/hb/hbaker/Use1Var.html (also .ps.Z)

This paper references a number of other papers on the Web on the same subject.

-- 
www/ftp directory:
ftp://ftp.netcom.com/pub/hb/hbaker/home.html
