USB drive unmount

[Gavin McCullagh]

Hi,

On Tue, 07 Aug 2007, Scott Balneaves wrote:

> ltspfs mounts the media with no cacheing, and unmounts in the background
> promptly on idle usage.  I've yet to receive a bug report from anyone
> of corrupted media because of an early eject.
> 
> It just makes sense: I'm done with the usb key, so I wait for the light
> to stop flashing, yank it, and stick it in my pocket.

I agree that in isolation the design is good (actually I think the best
design would be one which locks the USB key in until the system has
unmounted, but that's not something we can change).  However, that's not
the issue I was bringing up.  The issue is whether you should teach people
your (albeit better) design when 99.9% of the world use a different one
which does involve interaction.

Designing an expensive car which stopped itself from crashing would be
great, as long as drivers of it knew that the great majority of other cars
they would drive didn't do this.  If they don't know that, it might be
safer to just tell them that they had to avoid walls, cars, etc. while
allowing the car to still protect them should the moment arise.

As you say, there are technical barriers to an unmount menu entry just now
so I don't expect this to be fixed, but I still think ideally it might be
better to have the menu entries and even for it to complain when users
unplug at random.

Gavin