Re: user-visible goals

> Hi Keith;-)...  Some of these have undesirable social relations
> consequences. 

Of course they do!  Users don't understand the consequences of what they want!

Still, if we really want to understand what users want, we have to be honest
about that.  Tt's not helpful to prematurely filter out things that we believe
will do harm.

> >- Users want to be able to read mail without seeing a significant amount 
> >  of spam.  (to a recipient, "spam" is basically anything that the 
> >  recipient doesn't like that was sent by anyone except a person known 
> >  personally to the recpient)
> 
> This one needs careful review.  It is not universally true.  In particular,
> I have often sent mail to Keith that he detested,

I hate to break this to you Stef, but you're well below the scale of people
who have sent mail that I detested. :)  That and I've known you personally for
almost as long as you've been sending me mail.  So you're not a spammer by
the definition I gave.

> >- Users want to be able to recall messages that have been sent but
> >  not yet read.
> 
> This one creates some interesting possibilities for causing friction among
> colleagues.  If you sent something to my USPS mailbox, and then came by my
> house to retrieve it from my mailbox before I removed it, and then make an
> issue of it sometime later, I think I would have grounds to claim you stole 
> it from my mailbox.

We're not talking about USPS, we're talking about what users want for email.

> So, that a minimum, I as recipient must have a way to 
> block such theft. 

Okay, so we might add 

- Users want their mailboxes to be secure against removal of messages
  by unauthorized parties (which I suspect everyone would agree with) 

and maybe 

- Users don't want senders to be able to "take back" mail that they've
  already sent but hasn't been read.  (which I'm not sure would 
  get widespread agreement.)

> >- Users want to be able to send messages that will be disappear
> >  if not read before some interval  (expiration)
> 
> This one opens doors to some nasty tricks, like using it to find out without
> detection that someone is not in his office when expected to be there,
> without any further investigation. 

Yes indeed.  Which is why I also wrote:

> - Recipients want to be able to read mail without the senders of messages
>   knowing that they've read their mail

(though maybe this doesn't quite capture the entire privacy concern?)

There are inherent conflicts between different things that users want.   It's
better if we try to understand these and make explicit compromises between
them than if we try to resolve the conflicts without articulating them.

Keith