On Wed, Mar 17, 2004, [EMAIL PROTECTED] wrote:
> I don't know about good etiquette, but if the product is open source,
> you still advise the user there is a bug, and advise them you will be
> releasing the security problem to the developers web site for them to
> recitify the problem. The key issue is that both the developers and
> the client needs to know about the problem. So they both can make an
> "educated" decission on their course of action.

I'm not involved in security procedures anywhere, but my understanding
is that when advising the clients/users might also risk advising
potential attackers (as it would with most open source projects), you
need to weigh up the gain of giving users early warning against giving
attackers early warning. There's not many cases where you can warn only
the 'good' users and not the bad unless you have a very tight
relationship with a small customer base.

In the case where you actually have a patch that fixes the problem, the
users can apply it themselves if the developers don't. However, if your
advisory is along the lines of "the entire design of your project is
riddled with code that assumes $X and $X is incredibly vulnerable" then
exploits will be developed quickly once the information is known, but
fixes slowly.

In that case, I would prefer as a user the situation where developers
are advised without knowledge and I'm advised when fixes are available,
to the alternative where I know straight away, so do attackers, and
fixes aren't available for weeks.

This all changes if exploits get out of course, and also changes in the
case of uncooperative developers.

-Mary
-- 
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Reply via email to