Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site houxk.UUCP
Path: utzoo!linus!decvax!harpo!eagle!mhuxi!houxm!houxk!Richard Trauben
From: Richard@houxk.UUCP
Newsgroups: net.lang
Subject: Language Comparison, Stacks, etc
Message-ID: <103@houxk.UUCP>
Date: Tue, 11-Oct-83 21:24:04 EDT
Article-I.D.: houxk.103
Posted: Tue Oct 11 21:24:04 1983
Date-Received: Thu, 13-Oct-83 04:15:00 EDT
Sender: rdt@houxk.UUCP
Organization: Bell Labs, Holmdel NJ
Lines: 39

This is a open request to this newgroup for information, opinions and
examples. I am a casual user of Pascal and C.

Recently,  I read a Consultants report done by BBN for Lawrence Livermore Labs
comparing and contrasting 4 different high level languages: Ada, Praxis, 
Pascal and C. The document is available as a DoD report.

There were a few questions I had regarding recommendations of the report. 

1)	BBN said that the lack of a variant record structure in C was a 
	serious weakness. I didn't feel that way based on my frequency of 
	it's use in Pascal. Am I mistaken? Are there serious examples of 
	places where it is ABSOLUTELY needed?
2)	What is overloading? Ada is claimed to have too complicated a 
	scope rule due to overloading. Could someone tell me in plain english
	what is overloading AND is it a frill?
3)	What is the Pascal mechanism used by the compiler to implement 
	the display mechanism? As I understand it, the display is a way
	for a called procedure to determine the values of a caller procedures
	local variables. It seems to be a secondary stack whose contents
	are those stack addresses at which the starting addresses
	of each procedure frame resides. Is this correct. If so, do compilers
	allocate this secondary stack pointer (display stack pointer) as
	a general purpose register (e.g. %r3 of VAX-11/180) or in a fixed
	memory address?
4)	One unrelated but interesting question: Is a negative growing
	stack frame discipline better or worse than positive growing
	stack frame discipline? Is a ngs discipline re-traceable 
	or reconstructable if the stack frame gets broken during run
	time? Is there a public domain recovery procedure for ngs short
	of killing the process?

Regards,
(No Flames please, just illumination)
Richard Trauben 
AT&T Bell Labs
(201) 949-0088