[Top] [All Lists]

Re: The armour issue

1997-11-26 13:44:11
Ian Brown wrote:
Ascii armor allows PGP binary data to be held or transmitted in
any environment that may not be safe to raw binary data.

Can you give us an example of this apart from mail? Web-based keyservers
only use armour because they are based on mail servers. HTTP allows
binary data to be transferred directly.

( Armour also allows cut and paste.  Presumably MIME would as well, I
don't know. )

I can see very little reason why people would want to store PGP data
armoured in a system that supports 8-bit data.

I can't think of a reason either.  What I do want to do is to store
clear text  signatures.  That I need.  I don't know whether MIME can do
that, or whether it is applicable ...

The specific instance is the contract.  The cleartext sig must be
embedded, obvious, and small (a la PRZ bit-culling).  I would see S/MIME
as a non-starter here as people talk about multiple K sigs, but maybe a
MIME encoding of a PGP sig is ok.  I won't know until I see one and look
at how scriptable it is.

I don't understand enough about the MIME vs Armour debate to know
whether clear text sigs are effected.  Is this Armour or something
else?  It's certainly PGP.

There are other applications that use PGP technology
for authentication, enveloping, etc. and use armor.

Is this because they transfer data by mail?

If assuming the above contract is included in the notion of Amour, then
the answer is "not only mail, they have to be readable by humans in a
wide variety of contexts."

To forbid armor is to say that there is no conceivable use for it.

No-one has suggested this.

I suggest you be boring and repeat (every second post) what it is that
is being suggested.  One of the problems I had with this debate was
working out what the proposal was (that's partly why I kept away from
it, waiting for someone to state what).

Armor itself is
still a preferred way of transmitting OP objects that are not messages.

Could you give us a case where it is vital to the understanding of an
object that it is armoured - i.e. just interpreting the underlying
binary packet(s) will not give you the full meaning of the object.

The contract above must be clear, not binary.  The reasons for this are
complex and messy.  I'll fill you in if we can decide if a clear text
sig is applicable to this debate.


FP: 1189 4417 F202 5DBD  5DF3 4FCD 3685 FDDE on

<Prev in Thread] Current Thread [Next in Thread>