From: Andrzej Filinski <andrzej@daimi.aau.dk>
To: categories@mta.ca
Subject: Re: co-exponential question
Date: Thu, 22 Jul 1999 22:28:12 +0200 (MET DST) [thread overview]
Message-ID: <14231.32476.50008.946207@harald.daimi.au.dk> (raw)
Paul Levy writes:
> I don't know if this is relevant to your question, but there is an
> example in programming semantics where the dual of a cartesian closed
> category has independent significance, due to Lafont, Streicher, Reus
> and Hofmann. (Some further work was done by Selinger.) It is found
> in the following paper: [Streicher & Reus 1998]
Indeed, co-exponentials are closely tied the semantics of call/cc-like
control operators. This connection is perhaps expressed more directly in
the following (regrettably somewhat unpolished) work:
@MastersThesis{Filinski:89a,
author = "Andrzej Filinski",
title = "Declarative Continuations and Categorical Duality",
school = "Computer Science Department, University of Copenhagen",
year = 1989,
month = Aug,
note = "DIKU Report 89/11",
URL = "http://www.brics.dk/~andrzej/papers/DCaCD.ps.gz"
}
@InProceedings{Filinski:89b,
author = "Andrzej Filinski",
title = "Declarative Continuations: An Investigation of
Duality in Programming Language Semantics",
booktitle = "Category Theory and Computer Science",
series = LNCS,
number = 389,
address = "Manchester, UK",
month = Sep,
pages = "224-249"
}
Specifically, in the category induced by the types and terms of a
call-by-value language with first-class continuations, the coproduct
functor - + A actually has a left adjoint - x (A -> 0), where 0 is the
empty type. The operational intuition behind this construction is that
a function f : X -> Y + A returning either a "normal" (Y) or an
"exceptional" (A) result corresponds to a function f' : X x (A -> 0) ->
Y, returning only the normal result, but passing any exceptional results
to an additional, non-returning-function parameter.
(In a purely functional language, the type A -> 0 would of course
contain at most one value; but in the presence of control effects, such
as exceptions, there can be many distinct "functions" of this type. And
with a sufficiently powerful control operator, it actually becomes
possible to define co-application and co-currying with equational
properties exactly mirroring those of CCC exponentials.)
Still, as Paul notes,
> K is certainly an important category, but I wouldn't say that the fact
> that it has coexponentials is significant.
constructing a categorical semantics of a language with control operators
_directly_ in terms of coexponentials is somewhat awkward, and the
formulation used by Streicher and Reus is almost certainly nicer to work
with in practice.
-- Andrzej
next reply other threads:[~1999-07-22 20:28 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
1999-07-22 20:28 Andrzej Filinski [this message]
1999-07-23 19:40 ` Peter Selinger
-- strict thread matches above, loose matches on Subject: below --
1999-07-16 19:55 Bill Halchin
1999-07-20 20:57 ` William James
1999-07-20 17:41 ` Paul Levy
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=14231.32476.50008.946207@harald.daimi.au.dk \
--to=andrzej@daimi.aau.dk \
--cc=categories@mta.ca \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).