[horde] What to do when removing user

[Eric Rostetter]

Quoting Mike Bydalek <mbydalek at mobilemini.com>:

> Eric:
>
> To reply to all your comments: I totally agree with you, but the fact of the
> matter is, people aren't perfect.  I'm was just trying to play the Devil's
> advocate by providing some worst case scenarios.

Hmmm...  That's usually my job (to argue as Devil's Advocate).

> I guess my main point is to
> just keep the user informed.  If all that should be done is after clicking
> to delete "jsmith" a warning saying "All data will be lost, etc. etc., are
> you sure?" then so be it.

Yes, I think it should have a confirmation pop-up or other confirmation.
Horde usually confirms all deletions of any relevance.  I'd even say maybe
the message should be more explicit ("All data, including any shared data,
in all available modules, will be lost, etc. etc. etc.").

> To me, Horde is great because of all the different options it presents, so
> presenting several options when deleting a user wouldn't be out of line.  If

I'm not against multiple UI implementations for delete, but someone has to
define a workable, well defined alternative to "remove everything they own"
and so far I haven't seen any such submission.

> This I'm not sure I get what you are saying.  I didn't say "Don't ever
> delete user data," just to have an option to or not.  If you don't opt to
> delete their data, then at least get rid of 100% certain non-shared data
> while keeping potentially shared data in tact with of course a list of it so
> you can review it later.

I'll allow two options of delete everything, or delete no data.  But anything
in-between needs a workable solution presented, and I've yet to see one that
meets my requirements, and I doubt the others have either (though I shouldn't
presume to speak for them).

In your statement, we need a workable (and low overhead) way to determine
the "100% certain non-shared data" to be removed, a way to remove it, and
a way to keep the other data along with some way to "list" it for 
consideration.
I've no answer to how to determine which are in which category, or how best
to build this "list" of data to be reviewed.

> Regards,
> Mike

--
Eric Rostetter
The Department of Physics
The University of Texas at Austin

Why get even? Get odd!