> How close is pam_kcoda related to the official coda-project? Are they using 
> current coda-sources for their development, how fast will changes be merged 
> into it?

pam_kcoda has essentially nothing in common with Coda sources, except
the knowledge about clog semantics.

It relied on clog using cached Kerberos credentials, it just runs clog
with some arguments and passes no other data to it.
It should be rewritten to pipe a password.
It might become a "general purpose ''run a binary with some flags
and optionally give it the pass''" module.
I didn't found one like that when I searched.

Though, I believe Coda tokens can be a suitable layer for implementing
global identities (see the list archive) and using them for more than file
access (say, host login authentication). Then it might be worth linking
a corresponding module with Coda libraries...

Best regards,
--
Ivan