[openi18n-im:01207] Re: user-running iiim server discussion

: Sun, 27 Feb 2005 22:53:25 -0800 (PST)
: Hideki Hiura <[email protected]>:
> > From: AShiZaWa KaZuNoRi <[email protected]>
> > > Or you are talking about further enhancement of current -user option?
> > 
> > I should have mentioned this first.  Of course, existing code of
> > UNIX domain socket will be used.  Namespace-based file I/O
> > comes, and single iiim server for clients on single X desktop is
> > also proposed.  There should be some amount of adjusting works.
> > I'm focusing this works.
> 
> Sounds good to me. So the simple answer to 
> > > Or you are talking about further enhancement of current -user option?
> is yes, right? 

Yes.

> 
> Would you elaborate what do you have in mind to be added to current -user 
> option?

Adjustments.  I have no intention to introduce new technologies.
I just want to make sure that private input method service works
fine.

> 
> > > > How about (3)?  If private iiim service which is started by
> > > > user-A is shared with user-B, then user-A can see every key
> > > > strokes by user-B.
> > > 
> > > I can read this paragraph two different ways;
> > > 1) Do you want to enable the sharing of private server? or
> > > 2) Bring up a discussion on how to verify the credential of server 
> > >    from client?
> > 
> > Neither for me.
> 
> So did you you intend
> 
>     3) Let's not aim for (3)

It's not my intention.

Date: Mon, 28 Feb 2005 13:47:15 +0900
From: AShiZaWa KaZuNoRi <[email protected]>:
> (3) Start privileged shared iiim service when the service is
>     required by users, i.e. on demand.  This needs some support
>     by the system.

(3) is `privileged shared iiim service', not `private shared
iiim service'.  Security risk of (3) is almost the same as (4),
I assume.

Regards,
-- 
kasha