Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: $Revision: 1.6.2.16 $; site ada-uts.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!think!ada-uts!richw
From: richw@ada-uts.UUCP
Newsgroups: net.lang
Subject: Re: Type-System Questions
Message-ID: <15100009@ada-uts.UUCP>
Date: Thu, 31-Oct-85 11:06:00 EST
Article-I.D.: ada-uts.15100009
Posted: Thu Oct 31 11:06:00 1985
Date-Received: Sun, 3-Nov-85 09:45:32 EST
References: <15100005@ada-uts.UUCP>
Lines: 43
Nf-ID: #R:ada-uts:15100005:ada-uts:15100009:000:1585
Nf-From: ada-uts!richw    Oct 31 11:06:00 1985


First of all, thanks for all the replies; I've got quite a bit of
reading to do...

>>> 2) Includes BOTH compile-time types and run-time, "manifest" types?

I apologize for not being more specific about "manifest" types
(see below for a reference to the text from which I grabbed the term).
I realize now that unions (of C) or variants (of Pascal) or
oneofs/variants (of CLU) might be lumped in this category.
These aren't quite what I was curious about because they
are types which represent a union of a _finite_ set of other types.
I was curious about data which stored the name (or something)
of its type, and thus could be ANY one type of the "infinite"
number of types available (assuming the language provided for
user-defined types).

Since the term, "manifest", apparently isn't as universal as
I'd assumed:

> Terminology problem here:  "Manifest" types are types known to the
> compiler.  Types known only at runtime are called "latent" types.
> Roughly the same meanings are associated with "static types" (known to the
> compiler) and "dynamic types" (known at runtime; more to the point,
> changeable at runtime).

...let me quote my source. I used the term "manifest" after seeing
it in:

    "Structure and Interpretation of Computer Programs",
     by Abelson, Sussman, & Sussman

Specifically, page 132 says:

"A data object that has a type that can be recognized and tested
is said to have _manifest_ type."

The text then goes on to describe a scheme (no pun intended) for
doing generic arithmetic using tagged data.

Again, thanks for the replies.

-- Rich Wagner