Subject: Re: An approach for detachable interfaces.
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
Date: 11/05/1999 15:27:34
In message <19991105215848.C610@jocelyn.rhein.de>Ignatios Souvatzis writes
>On Fri, Nov 05, 1999 at 03:44:00PM -0500, Bill Sommerfeld wrote:
>Uhm, his scenario (having two similar cards, one with parameters X at work,
>one with parameters Y at home) seems useful to me.
>Although I'd rather use the same card and reconfigure it.
Me too. But if the ethernet at home is 10base2, and the ethernet at
work is 100baseT, you need two cards regardless.
And some people use port replicators at home, or work, or both; and
they often have ethernet (10baseT USB ethernet?) built in now.
And (again) I'm still looking at the initial case where we dont have
userland notification and a daemon to handle everything.
Say i have two `ep' cards, but one is cardbus and the other is pcmcia
(assuming that Bill's notion of `same' follows attachment in the bus
hierarchy, not just the name of the leaf node).. Then Bill's proposal
would give what I want. But if I had two pcmcia cards or two cardbus
cards, it'd give what Bill wants. Least-surprise violation, no?