Re: Child killing UI (was Re: Reliability of RPC services)

[Marcus Brinkmann]

At Fri, 28 Apr 2006 01:24:07 +0200,
Pierre THIERRY <address@hidden> wrote:
> 
> Scribit Marcus Brinkmann dies 28/04/2006 hora 00:59:
> > I don't see a problem with providing users a process tree and the
> > option attempt to kill parts of the tree, including leaf-nodes.
> 
> When something goes wrong in the browser, how does the global process
> tree dialog fires up? In KDE or Windows XP, the user has to have prior
> knwoledge on how to do this (respectively Ctrl+Esc and Ctrl+Alt+Del,
> IIRC).

Yes, the user has to have prior knowledge in order to use his computer
optimally.  This will always be the case.  However, in KDE or Windows
the machine can get into a state where the user _can't_ fire up the
dialog anymore.  This is what we are working on preventing.

> > A similar issue arises with authorising access to files via the
> > powerbox: A browser could advise the user that access to a file is for
> > a certain plugin only.
> 
> I'm not sure if the powerbox shoudl allow such potentially malicious
> behaviour: if the resource is for a plugin, shouldn't the powerbox be
> able to tell the user that the plugin indeed will be the recipient of
> the capability?

It can't, because it is the powerbox of the browser.  The plugin does
not have its own powerbox.  The behaviour is not potentially malicious
if it is clearly defined what the _advisory_ information the browser
provides actually means.

> If the browser will be the recipient, the powerbox should at least warn
> the user that any reason given by the browser about the need for the
> capability carries no guarantee.

I agree.

Thanks,
Marcus