Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!think!mit-eddie!genrad!decvax!dartvax!uvm-gen!poole
From: poole@uvm-gen.UUCP (Damon "nomad" Poole)
Newsgroups: net.lang.mod2
Subject: m2 working group
Message-ID: <444@uvm-gen.UUCP>
Date: Wed, 29-Oct-86 17:57:21 EST
Article-I.D.: uvm-gen.444
Posted: Wed Oct 29 17:57:21 1986
Date-Received: Tue, 4-Nov-86 00:03:12 EST
Organization: EMBA Computer Facility, Univ. of Vermont, Burlington.
Lines: 46


      I would like to make the following comments on the report released by
the Modula II working group:

WG084:  Good thinking, there are cases where you want to define something but
        not export it.  The programmer should have TOTAL control over exporting.

WG102:  typename(expression) should not be referred to as coercing.

WG115:  Type transfer and type conversion should remain separate.

WG117:  How about a second parameter for HIGH to indicate the desired array
        index?

WG120:  Could some one explain the need for this??

      Other comments:

      If strings are defined as arrays of char, then assignment of strings of
unequal size should not be allowed.  Such assignment would be inconsistent
with normal array usage.  How about partial assignment of one array to another?
No, only arrays of equal size should be allowed.

      'StringTerminator' is an EXCELLENT idea, but since the current terminator
is 'nul' and 'nul' is used so frequently in string handling, a shorter name
should be adopted.

      Adding a new 'string' type disregards the handling of arrays already
available.  Why complicate the job of the compiler with a new standard type?
One of the biggests idea behind Modula II is simplification of compilation.
That's why all of the I/O is in separately compiled modules, aside from the
portability considerations.  Just look at the grammar for M2: 97 lines com-
pared with how many hundred for Pascal??  Let's not come out with "Modula II,
the common man's Ada!"  YUCK!

      I totally agree with Richard Pattis' comments about opaque types.  Re-
quiring an opaque type which is actually of type INTEGER to be a pointer to
an integer would be a mistake.  Incidentally: "Long live the ADT!"

                        Damon "nomad" Poole

P.S. - I am a senior at the University of Vermont, and am trying to get
Modula II introduced into some of the courses here.  Could anyone suggest
an alternative to Logitech's M2 for the PC?  I am very pleased with their
product, but it doesn't fit on two disks conveniently and would complicate
the life of PC users who don't have Hard Disks.  Thank-You in advance.