Path: utzoo!news-server.csri.toronto.edu!rutgers!cs.utexas.edu!usc!zaphod.mps.ohio-state.edu!van-bc!ubc-cs!uw-beaver!rice!bbc
From: bbc@rice.edu (Benjamin Chase)
Newsgroups: comp.lang.misc
Subject: Re: Dynamic typing -- To Have and Have Not (was Runti
Message-ID: <BBC.91Mar11103121@sicilia.rice.edu>
Date: 11 Mar 91 10:31:21 GMT
References: <593@optima.cs.arizona.edu>
	<PCG.91Mar10181749@aberdb.cs.aber.ac.uk>
Sender: news@rice.edu (News)
Reply-To: Benjamin Chase <bbc@rice.edu>
Organization: Center for Research on Parallel Computations
Lines: 32
In-Reply-To: pcg@cs.aber.ac.uk's message of 10 Mar 91 18:17:49 GMT

>I would propose the following terminology:

><type>			The equivalence class of all <values> to which
>			a given set of operations can be meanigfully
>			applied.

Make certain that you don't proscribe certain views of the world
through too narrow a definition.  Consider the language Russell, where
types are viewed not as data, but rather as collection of operations
on data.  You may consider this to be an argument over a glass of
water being "half-empty" vs. "half-full", but I found the Russell
outlook on the question of "what are types, really?" to be very
pleasant.

><latent typing>	When the <value> denoted by a <reference> has a
>			<type> that cannot be determined at compiletime.

><manifest typing>	Being able to do <typing> for the <values>
>			denoted by <references> (i.e. variables) at
>			compiletime.

Of course, the obvious name for manifest typing is really blatant
typing, the obvious meaning being that the types of things are obvious
enough that even a compiler can deduce them :-).  That way, we have
"latent" vs. "blatant", a very pleasant sounding pair.  Of course, the
invitation to incorrectly spell them either (or both?!?)  as blatent
(bleah!) or latant (fooey!) is too tempting for some...
--
"it is perfectly safe to eat the yogurt under the green fuzz,
 and I do it fairly often" - brian reid

	Ben Chase <bbc@rice.edu>, Rice University, Houston, Texas