[standards-jig] auth & reg redux

Evan Prodromou evan at prodromou.san-francisco.ca.us
Fri May 30 17:10:07 UTC 2003


>>>>> "PS" == Peter Saint-Andre <stpeter at jabber.org> writes:

    PS> Hmm, further questions arise.  1. Should these JEPs remain
    PS> standards-track? We made them standards-track so that we could
    PS> include them in our protocol compliance testing suite
    PS> (JEP-0073). Is that reasoning still valid? I think so, but I'm
    PS> not 100% sure now.

Peter,

It seems like the life of j:i:a and j:i:r depends strongly on JEP
73. I think it'd take an extremist neophile to argue that they weren't
important for the transition period between pre-1.0 and 1.0 Jabber. (I
think that an in-band registration mechanism is pretty key for public
servers even in post-transition time -- others may disagree.)

Anyways, I guess I'm saying that we might want to use a mark-and-sweep
garbage collection algorithm for these two protocols. As long as JEP
73 keeps a reference to them, they stay active. When we feel
comfortable enough with the state of Jabber to remove them from Jabber
IM Basic, they lose their reference and are subject to collection. :-)

    PS> 2. It seems odd to advance a protocol to Draft when we know
    PS> that it will be deprecated and then made obsolete on a fairly
    PS> well-known schedule.

Yes. However, I think this is due to the circumstances -- they're two
ubiquitous protocols which antedate the JEP process and which are
crucial to interoperability for the near term.

    PS> My feeling is that we make these standards-track, with a
    PS> review every six months to determine if they should be
    PS> deprecated. But I'm open to argument.

My argument is this: as long as we think they're necessary for The
Transition in JIMB0 (OK, I just made that up, and it's probably
already used), they stay alive. When JIMB 1.1 or 2.0 or 3.0 comes out
without a requirement for j:i:a or j:i:r, they go on the chopping
block.

~ESP

-- 
Evan Prodromou
evan at prodromou.san-francisco.ca.us






More information about the Standards mailing list