Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!rutgers!sunybcs!bingvaxu!marge.math.binghamton.edu!sullivan
From: sullivan@marge.math.binghamton.edu (fred sullivan)
Newsgroups: comp.lang.modula2
Subject: Imports in definition and implementation modules
Message-ID: <733@bingvaxu.cc.binghamton.edu>
Date: Wed, 7-Oct-87 09:37:22 EDT
Article-I.D.: bingvaxu.733
Posted: Wed Oct  7 09:37:22 1987
Date-Received: Sat, 10-Oct-87 07:51:15 EDT
Sender: news@bingvaxu.cc.binghamton.edu
Reply-To: sullivan@marge.math.binghamton.edu (fred sullivan)
Distribution: world
Organization: Department of Mathematical Sciences, SUNY-Binghamton
Lines: 20

As I promised some time ago, I wrote to N. Wirth to ask whether
identifiers imported into the definition module should be known in the
implementation module, or should be reimported.

His answer follows:

"Concerning your question about imports in definition modules, I have
been of the opinion that the corresponding implementation module adopts
those identifiers (and associated data) that are declared in the
definition module itself, but no more.  In this case, it is indeed
necessary to import client modules again in the implementation module,
if identifiers declared in it are referenced."

I yield.

Fred Sullivan
Department of Mathematical Sciences
State University of New York at Binghamton
Binghamton, New York  13903
Email: sullivan@marge.math.binghamton.edu