Re: Thoughts on an OpenPGP header?
The trouble with a finger URI spec of "finger:user@host" is that
there's no support of the /W modifier specified by the finger RFC
and included in many implementations. /W specifies a higher level of
detail in the results returned.
For any reason an implementation of the finger server might not return
a PGP key *unless* the /W modifier has been given. This suggests that a
finger client implementation designed to support PGP key lookup should
always submit the /W modifier. (/W should be ignored if not implemented,
but that's only specification, and does not describe the real world.)
However, finger does not specify that a /W query result is a literal
superset of a query result without /W, so really it should submit both
ways and merge the results.
For all this, I suggest that finger://user@host/[W] or
finger://host/user[/W] is a more suitable specification of a finger
URI. At any rate, it seems a can of worms not to go with something more
complex than finger:user@host.
Not that this is the right forum, necessarily, for criticizing draft
finger: URI specs, but that's really all I can think of to say about it.
--
-D. dgc@xxxxxxxxxxxx NSIT::ENSS