Re: Nautilus, metadata and extendet attributes

[y9toi7y02 sneakemail com]

On Tue, 3 Feb 2004, Manuel Amador (Rudd-O) amadorm-at-usm.edu.ec |nautilus| wrote:

> El mar, 03-02-2004 a las 05:42, Olaf FrÄ?czyk escribió:
>
> > I wouldn't click the above "Readme.txt              .scr" either :)
> > I just don't click on suspicious attachments.
> > And if a user is "not educated enough" to not open suspicious
> > attachments, I would bet that he will open binary Readme.txt.
>
> Olaf, you haven't understood yet that the computer has to work for us,
> and not the other way around.  I shouldn't be "educated" to use a
> computer.  Sure, by bonking around with the computer I might damage my
> own files, but it should be sturdy enough so that I can use it with
> confidence, both in terms of physical endurance and software robustness,
> and not damage the system or be vulnerable to external malicious
> factors, to a certain degree.  Go read the User Interface design book
> from Joel Spolsky, it's on his Web site.
>
> When you understand this, you will be able to understand the whole
> argument behind this push.  It's about the increased software robustness
> and richer featureset that a metadata system brings to the table.
>
Note that neither a perfect sniffing method, the extension method or a
perfect metadata system will fix the problem of stupid users running
programs they received from strangers, thereby giving the strangers
control.

The only way is forbidding running programs from any untrusted source
alltogether. Or perhaps you might consider giving people a pop quiz:

'You are about to do something stupid. Before you can continue, you must
answer this question correctly to prove that you know what you are doing:

Which is the icon for shell scripts?

[icon] [icon] [icon]
  ( )    ( )    ( )

     <OK>   <Forget it>

Apologies for sounding sarcastic, but this non-discussion isn't going
anywhere anyway.

Martijn Vernooij