-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Thu, Jul 03, 2003 at 07:05:16PM -0700, Hal Finney wrote:
Jon Callas writes:
I have a request for an algorithm number for bz2 compression. The
implementer in question has promised on a stack of holy books only
to use it along with compression prefs. Anyone object strongly?
I don't see a need to add another compression algorithm unless there is
something wrong with the ones we already have. Adding a new one can only
hurt interoperability in the long run. What is the reason for adding it?
I don't have strong feelings for or against adding bz2, but your
comment about interoperability raises a related issue. In theory, the
preference system would prevent the use of bz2 except when it can be
properly handled by the recipient so there should be no
interoperability issues.
Of course, that's theory. The preference system works quite well on
paper, but unfortunately fails in the case where a key is generated in
an implementation that can use bz2 and the public key is distributed.
Later, the user changes their implementation to one that cannot use
bz2. Anyone sending a message to that public key has the belief that
bz2 can be safely used, and may well use it, causing a problem since
the user's new implementation cannot handle bz2 (even though their key
claims they can).
I have already seen a few examples of this problem (a PGP-generated
key with an IDEA pref being used on GnuPG, and a GnuPG-generated key
with a ZLIB pref being used on PGP).
I don't think the answer here is to restrict the use of new
algorithms. 2440 has this to say, which pretty much eliminates the
problem in the design:
Since a self-signature contains important information about the
key's use, an implementation SHOULD allow the user to rewrite the
self-signature, and important information in it, such as
preferences and key expiration.
I don't advocate making any severe changes in the preference system,
but perhaps the language here could be made a bit stronger? Something
like "Note that without the ability to rewrite a self-signature,
interoperability issues may occur when the same key is used in more
than one implementation." would be great.
David
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3rc1 (GNU/Linux)
Comment: Key available at http://www.jabberwocky.com/david/keys.asc
iD8DBQE/BQDe4mZch0nhy8kRAsLRAJsF+Zc8fD85cjGV4JIT8Kv7QJLg5wCffisr
U+65IozEBIVm+SznfIwniDk=
=9xSE
-----END PGP SIGNATURE-----