Ralf Hemmecke <[EMAIL PROTECTED]> writes:
| Forwarded with the permission of Stephen Watt.
|
| On 05/13/2008 05:30 PM, Stephen Watt wrote:
| > Hi Ralf,
| >
| > For does not introduce a new scope for its body for a few reasons:
| >
| > 1. In languages like C any compound statement is a new scope so we
| > have new scopes for loop bodies, branches of if-s etc.
Note that in C++ (and some other Algol-like languages), an
if-statement introduces a scope for each branch.
Unrealted to this (and way before I left yesterday), I was planning to
send another message on that topic. Currently, Spad does not
introduce a scope for each branch of an if-statement, and that can
lead to surprises. There is a `semantic error' building the algebra
on that speicifc point.
| >
| > This does not interact well with implicitly local variables.
| >
| > If a new variable is by default local to the smallest enclosing
| > scope, then new variables would be visible only in overly small
| > regions.
Yes. When I do computations in for loops, I want the local variables
to stay local to the loop. I don't want them to pollute the enclosing
scope.
| >
| > E.g. suppose compound statements were scopes, then
| >
| > if foo? then a := 1 else { f(); a := 2 }
| >
| > would give unexpected results.
Not in OpenAxiom if written as
a: Integer
if foo? then a := 1 else { f(); a := 2 }
The rule in OpenAxiom is:
if there is a visible name, then it is taken unless it is a
definition (e.g. variable + typespecification + := + value)
. Otherwise, it is implicit declaration local to the enclosing scope.
| > 2. It is convenient to be able to test the last value that a for loop
| > index reached.
| >
| > 3. Since the language allows goto, it is useful to be able to reduce
| > structured control flow to the goto case.
Spad does not have goto.
| > I admit that none of these are absolutely compelling. So I should add
| > that we tried it the other way, and this way made coding the library
| > easier.
-- Gaby
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
open-axiom-devel mailing list
open-axiom-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/open-axiom-devel
