On Fri, Jun 20, 2003 at 12:55:58AM -0500, David Phillips wrote:
> Munir Nassar writes:
> > i bet you ten bucks that somebody will find a security hole. it is not
> > that i doubt your coding skills but it is a fact that security holes
> > are a fact of life.
> 
> Only ten?  Care to make it interesting?

Yes: I duble-dog-dare you.

>                                          Security holes are not a fact of
> life.  Security holes come from being ignorant, having poor / sloppy coding
> skills and not being mindful of security. 

... as opposed to the regular software bugs which just magically
appear...

>                                            It is not difficult to write
> secure code in a scripting language.  If I was not certain that I could
> write secure applications, then I would be looking for a new line of work.
> 
> > consider this:
> > the openbsd hackers pride themselves in secure code... they code audit
> > everything before it can be used. one could argue that they are
> > security experts.
> 
> Consider this:  qmail, one of the most widely deployed MTAs, has never had a
> security hole.  It was first released in January of 1996.

qmail has not had a discovered and publicized security hole. That does
not say _anything_. The fact that no security hole has been discovered
in 7 years says something, but don't bet your life on that.

And what good is qmail's security when the box is rooted through
bind/apache/whatever? Before you declare your application secure, you
better audit the whole stack - from TCP/IP to your database. Good
luck.

Go and read the excelent "Secrets and Lies" by Bruce Schneier.

florin

-- 

"NT is to UNIX what a doughnut is to a particle accelerator."
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://shadowknight.real-time.com/pipermail/tclug-list/attachments/20030620/819ef324/attachment.pgp