Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!aplcen!uakari.primate.wisc.edu!unmvax!sandia!jgmicon
From: jgmicon@sandia.UUCP (Jeff G. Micono )
Newsgroups: comp.lang.c
Subject: Re: Crash a RISC machine from user-mode code:
Summary: Utopia
Keywords: RISC
Message-ID: <292@sandia.UUCP>
Date: 14 Aug 90 16:48:23 GMT
References: <1826@mountn.dec.com> <49041@seismo.CSS.GOV>
Organization: Sandia Natl Labs, Div. 9224
Lines: 60

In article <49041@seismo.CSS.GOV>, stead@beno.CSS.GOV (Richard Stead) writes:
> Pretty Silly.
> 
> Do VAX-CISC programmers spend their days branching to random data?  I

no, but programmers DO make mistakes, and some people are malicious...

> thought programmers were paid to create software that did real things.
> I would hope that I never write a code that branches to random data.  Or
> if I ever do, I would fix it pretty damn quick.  (My definition of a
> code that branches to random data is "broken").  

my definition of broken is a system in which a user mode program (however
generated) can crash the system.  what if your "broken" program corrupted 
the OS, how could you fix it without rebooting?

> Who could possibly care
> that a random instruction sequence crashes a risc box?  That's WHY we
> have compilers - so we don't generate such things.  

and compilers NEVER generate wrong code...

> In any case, VAX-CISC
> and 68020's have been around for a long time - all the random instruction

                                                 ^^^^^^^^^^^^^^^^^^^^^^^^^^
really?  i would estimate this to be a very large number.  very large.

.> sequences that could crash those have probably been identified and fixed
> in software.  The oldest popular risc architecture is sparc, and that's only

fixed in software?  do you know anything about computers? 

> been around a short time.  I don't need this kind of ROBUSTNESS to protect

i sure need the architecture to provide a foundation so the OS can give me
"robustness" to protect me from idiots like you!

> me from bad programming - I do that well enough myself, thank you.
> Let the bad programmers pay through the butt for vaxen and deal with

vaxen == vaxeln?

> slow, archaic architecture (and, most likely, VMS - a regular Gulag
> of operating systems compared to Unix).

obviously, you've never had the opportunity to work with VMS on a 
project which was large and written for regular people to use,and 
not experts (a la Unix applications)

Unix -- just what is the print command these days?

I can't wait for Mach to come out and lay this VMS/Unix war to rest
once and for all...

> 
> Richard Stead
> stead@seismo.css.gov

Jeff Micono