Re: What method do *you* use for signing Usenet posts?

[Josh Huber <huber@alum.wpi.edu>]

Andreas Fuchs <asf@void.at> writes:

> I wonder - why does the user have to decide such things? Would it
> not make more sense to just specify the method and let gnus itself
> figure out whether it should use mime or not (i.e. if there is a
> mime part, use detached signatures, if not, use plain pgp)?

I really like this idea.  Way back in the day (okay, only about a year
ago, I think... :), I put in some logic which scans though the message
and decides whether or not to automatically insert a multipart signed
tag, or a part signed tag.

Prior to that, if you selected "sign" or "encrypt" it would only get
the first part in the case of a multi-part message. This could be
quite bad, and almost certainly was not what the user was expecting to
happen.

We could have some other tag, such as "pgpauto", which selected plain
PGP when there were no other attachments, but used pgp/mime if there
were.

The problem which arises from this (IMHO) is one of keybindings.  What
do we use?  I already don't like the fact that there are separate
keybindings for pgp and pgp/mime -- personally, I think there should
be one keyboard interface for signing/encrypting, and there should be
another variable which selects what style those commands use.

-- 
Josh Huber