Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!wuarchive!husc6!spdcc!ima!esegue!compilers-sender
From: jejones@microware.com (James Jones)
Newsgroups: comp.compilers
Subject: Re: Enumerated data types
Keywords: types, design
Message-ID: <9008271418.AA28641@microware.com>
Date: 27 Aug 90 14:18:52 GMT
Sender: compilers-sender@esegue.segue.boston.ma.us
Reply-To: jejones@microware.com (James Jones)
Organization: Microware Systems Corp., Des Moines, Iowa
Lines: 23
Approved: compilers@esegue.segue.boston.ma.us
In-Reply-To: <1990Aug23.134826.2865@forwiss.uni-passau.de>

In article <1990Aug23.134826.2865@forwiss.uni-passau.de> Luis Mandel writes:
>[query about enumerated types with name clashes]

If I remember rightly, Russell has enumerated types.  Since constants in
Russell are linguistically 0-ary functions, figuring out whether a given
instance of "red" would in full Russell be bike_colours$red[] or
car_colour$red[] is a particular case of the type inference problem, and
the same kind of context (i.e. how is the result used?) would be used to
decide whether succ[red] is bike_colours$succ[bike_colours$red[]] or ...
well, you get the idea. :-)

The BSD version of Russell had a kind of convoluted but usually workable
heuristic for type inference.   Hans Boehm showed that the problem of type
inference is undecidable, but I think he's recently written a paper that
appeared in SIGPLAN Notices that perhaps gives a useful subset of the
problem that's easier to figure out.  I hope that a new Russell or language
like it appears that implements such work, because Russell is, IMHO, a
very spiffy language and merits study by any would-be language designer.

	James Jones
-- 
Send compilers articles to compilers@esegue.segue.boston.ma.us
{ima | spdcc | world}!esegue.  Meta-mail to compilers-request@esegue.