Re: [Usability]Keeping the Quit menu item

[Havoc Pennington <hp redhat com>]

On Wed, Mar 19, 2003 at 02:53:13PM -0700, Joshua Adam Ginsberg wrote:
> 2) I do online banking with my browser... I pay my credit card bills
> online... I accumulate session cookies at secure websites that advise me
> to close my browser to ensure nobody can go back to those sites as me...
> if Galeon doesn't "vacate itself from memory", my session cookies don't
> expire... my HTTP authentication to my email doesn't expire... imagine
> this scenario at a kiosk or a public terminal... this would be *bad*...
> users are used to quitting the program giving them a clean slate when
> they start back up, not simply having the application cease to be
> visible...

Assuming users understand cookies/http-auth/etc., and those are
important security issues, then there should be a "forget my secure
information" button of some kind. You shouldn't have to make the
rather complex and in no way obvious judgment that exiting the process
will do that as a side effect, as it requires understanding not only
processes, but also cookies, http auth, and the difference between
memory and hard drive and what the program will keep where.  And the
user model doesn't include any of that.
 
> Maybe what's needed here is a good study of different scenarios
> involving different applications uses of windows, tabs, etc. and the
> various pieces of language used (close, quit, close all, quit all, etc)
> and maybe find some sort of consistent settlement. I think getting rid
> of quit is dangerous.

There's no (new) study required to know that users do not understand
processes, or the difference between memory and hard drive, or
cookies, or http auth. They just don't, this is a fact. You can go run
a new study if you want to waste your time. ;-)

Havoc