Delete key from keyserver

[David Shaw]

On Tue, Oct 25, 2005 at 06:22:10PM -0500, Alex Mauer wrote:
> David Shaw wrote:
> > On Mon, Oct 24, 2005 at 04:21:32PM -0500, Alex Mauer wrote:
> > 
> > 
> > I don't agree with this.  The user ID system in all OpenPGP products
> > gives a regular UTF-8 string.  Signatures simply bind that string to
> > the primary key.  The system says exactly "Alex Mauer belongs with key
> > 0x51192ff2" and "hawke at hawkesnest.net belongs with key 0x51192ff2".
> > You cannot sign a user ID without binding it to a key.
> 
> When the string "Alex Mauer <hawke at hawkesnest.net>" is bound to
> 0x51192ff2, there's also an association made between the two parts of
> the string.  The system does /not/ say "Alex Mauer belongs with key
> 0x51192ff2" and "hawke at hawkesnest.net belongs with key 0x51192ff2".  It
> says "Alex Mauer and email hawke at hawkesnest.net belong with 0x51192ff2"
> as one statement.  It's a subtle, but important, distinction.
> 
> I should be able to affirm that hawke at hawkesnest.net uses 0x51192ff2
> without being forced to also affirm that "Alex Mauer" uses 0x51192ff2,
> or vice versa.
> 
> Obviously, with these examples I'm happy to affirm both, since hey ..
> it's me.  But [hopefully] you get my point.

Some people (myself included) check both before signing.  The name via
some sort of formal ID, and the email via a mail challenge.

Still, if you don't want to bind both tokens together, just create an
user ID of <hawke at hawkesnest.net> without the name attached or a user
ID of "Alex Maurer" without the email address attached.  Some people
will not sign such a user ID though, and at least the name-only one is
of questionable usefulness in practice.

David