Gary Winiger wrote:
>> On Wed, 2008-01-16 at 10:07 +0000, Darren J Moffat wrote:
>>> The architecture I have no issues with. The terminology I have serious
>>> issues with.
>> I share Darren's concerns here. "Console User" seems like a better
>> name.
>
> I'm out of the office (and mostly on holiday) til next Wed.
> I've not carefully gone over the comments on names, and
> possibly won't til I return. My point about "owner" was two
> fold:
>
> Legacy defines
> console-owner a user who owns the
> system console device
> node
> (and this meant stat("/dev/console") and look
> at st_uid. I don't have an issue not following legacy
> other than the next point.
>
> Second, the term "user" seems too generic in terms of adding
> Rights automatically. Particular in when Virtual Consoles
> is configured. I like Joerg's comment "besitzer" and haven't
> found a good English subsititute ;-)
> Workstation/Console controller/possessor doesn't seen to flow ;-(
>
"Console User" seems to be a better term to me too in terms of virtual
console, in which there can be multiple console users at the same time
and this Infrastructure applies to them all, as we discussed in the
pre-review.
BTW, how about multiple users on multi-seat, which has been available
on SPARC for a long time?
And how about providing a library interface, something like
is_console_user()? It could have a Committed level, though the
implementation/definition may vary in different configurations?
It would be also beneficial to other projects/cases, like this one,
that expect to determine if a user is "Console User".
Regards,
Riny
