Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 17 Aug 1998 15:32:48 +0200
From:      Eivind Eklund <eivind@yes.no>
To:        Marius Bendiksen <Marius.Bendiksen@scancall.no>, Brett Glass <brett@lariat.org>
Cc:        freebsd-chat@FreeBSD.ORG
Subject:   Re: 64-bit time_t
Message-ID:  <19980817153248.61372@follo.net>
In-Reply-To: <3.0.5.32.19980817141851.00a51710@mail.scancall.no>; from Marius Bendiksen on Mon, Aug 17, 1998 at 02:18:51PM %2B0200
References:  <Pine.SGI.3.96.980817075447.5095P-100000@animaniacs.itribe. <199808141756.LAA24900@lariat.lariat.org> <199808171200.GAA21209@lariat.lariat.org> <3.0.5.32.19980817141851.00a51710@mail.scancall.no>

next in thread | previous in thread | raw e-mail | index | archive | help
(Moved to -chat where it belongs)

On Mon, Aug 17, 1998 at 02:18:51PM +0200, Marius Bendiksen wrote:
> >Nope, we're talking about safety measures in the language and compiler. Of
> >which we are consumers, unless we choose to reinvent both (which generally
> >is not the project at hand).
> 
> Regardless of the claimed secureness of a language, the designer/programmer
> of a product should always take it upon himself to add an extra layer of it
> by making sure his code doesn't have any potential flaws. I agree with what
> was said earlier, taking a compiler on trust is a bad move.  Of course, you
> should get a car with an airbag, but that's no excuse to drive too fast.
> 
> Besides which, the lack of such measures in C / C++ empowers the programmer
> to a great extent.

I regularly hear this claim.  Do you have any reasonable argument in
favour of it?  The ability to be close to the machine mapping is _not_
a reasonable argument - this is just an argument for why it is easy to
write compilers making reasonably decent code (while it makes it hard
to write compilers that make great code).

> Heuristics have not yet progressed far enough to second-
> guess a programmer as well as he himself can;  when such is the case, we'll
> all be outdated. (with the exception of heuristics programmers, of course.)

If this statement is intended to apply to the speed of generated code:
Can you give any references or data sets indicating this claim?  As
far as I know, this is just not true anymore - compilers can usually
generate better code than humans can, unless the human use an
unreasonable amount of time to write the code (analyzing exact
cache/pipe effects).

Eivind.

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-chat" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980817153248.61372>