Re: [Exim] Exchange looming..

Top Page
Delete this message
Reply to this message
Author: Matthew Byng-Maddick
Date:  
To: exim-users
Subject: Re: [Exim] Exchange looming..
On Tue, Aug 27, 2002 at 10:10:25AM -0500, Keith G. Murphy wrote:
> The big question is: Can something that complex and closed-source, no
> matter how impressive, be trusted with regards to dependability and
> security? Especially when *you can't fix it*.


I've seen this argument a lot over many things, and I often suspect I'm
in a minority in just quite how much code I have read for various of the
tools I run on a daily basis.

Certainly, my experience is that people don't audit code they're running,
in general, and the corollary to the Torvalds idea of "Many eyes make bugs
shallow" is that many people assume that someone else will look for the
bugs. My experience with Exim has been that I've only really delved into
the source once or twice, and it's been fairly easy to follow, and I'm not
sure how many people here have actually dived into the source for Exim.
What does seem to happen, however, is that those who do, and those who have
read the documentation back to front are the ones who can answer the
questions on this list.

Not being an exchange admin, I have no idea what the documentation is like,
but I'm left wondering if the support group is as comprehensive for
Exchange as this list is for Exim.

However, the way I see things, it is Phil Hazel's attitude that is why I
trust Exim not to lose mail, and the documentation that means that often
I can make small changes incrementally, and when I then come to test them,
I've normally not made any mistakes, but if I have, it's because of my
mistake, rather than the documentation being wrong. There are many other
Open Source and Free Software projects where this is not the case.

What you want out of a system should, and probably will, drive what you
choose to power that system, as others have said. But I think that the
attitude of the developers (or, in the case of Exchange, the marketing
department of Microsoft) is a far bigger reason to choose something, than
"I can fix the source myself".

MBM (apologies for this wandering off-topic)

--
Matthew Byng-Maddick         <mbm@???>           http://colondot.net/