On Wed, 2 Jun 2004, Eric Hildum wrote:
> As for the protocol itself; actually, the telephone based protocols are
> usually much better than protocols from data people. Since telecom
> people are used to designing life critical systems they tend to
> actually consider error cases when designing protocols.
I find that often telco protocols deal with errors in a too-complex
way, making it harder for the programer to handle them, not easier. And
as well, the protocols tend to be far too feature-rich, making them
more difficult to implement, and making interoperation harder and more
error-prone. ISO/OSI versus the TCP/IP suite is a perfect example.
> How many posters here have ever actually checked the status value
> returned from say, printf, in their code?
Never. I just let my generalized exception handler deal with exceptions
where I'm not too worried about it, and write specific exception
handlers to deal with the cases where I need them. Of course, I'm not
using tools like the C language, which seem to be designed to maximize
errors in software. (Consider: almost all of the buffer overflow attacks
we've been seeing over the past decade could not have happened--because
it's not possible to write code that would let that happen--if the
software had been written in Lisp, Scheme, Smalltalk, Java, or a whole
host of other languages.)
cjs
--
Curt Sampson <cjs_at_cynic.net> +81 90 7737 2974 http://www.NetBSD.org
Don't you know, in this new Dark Age, we're all light. --XTC
Received on Thu Jun 3 04:25:08 2004