Re: [apps-discuss] draft-yevstifeyev-pop-wrong-state

[Chris Newman <[e-mail]>]

--On August 23, 2011 13:00:29 -0400 Barry Leiba <[e-mail]> 
wrote:
>> 1. Is there significant interest in implementing this POP response code
>> extension? I ask because, if there is, it might be more appropriate to
>> bring this into some Applications working group.
>
> I can't imagine that there'll be a lot of interest.  POP clients and
> servers have been interoperating well for a good many years without
> this response code, and I don't see folks lining up to rev their code
> for something that basically amounts to servers helping new clients
> debug their bad implementations.

POP clients do not interoperate well when using TLS client certificate 
authentication on port 995. This might have been useful if the discussion 
around draft-melnikov-pop3-over-tls-01.txt had a different outcome, but I 
do not view it as particularly useful presently.

>> 2. If there isn't widespread interest, does anyone see any harm in this
>> being published as Experimental in the Independent stream? It seems to
>> me a very simple extension with little significant impact.
>
> It appears to me to be fairly useless, but totally harmless.  I think
> it's a waste of time, but I otherwise have no objection to it.

I'd call it "mostly harmless" :-). If a server implementation tries to be 
helpful prior to authentication, that might increase the code attack 
surface albeit by very little. But the draft does not require that 
behavior. Non-standards track RFCs do end up in RFPs sometimes and this one 
might fall in the category of "easier to implement than to try to correct a 
bad RFP". So it could end up wasting implementer time, albeit not very much.

I have no objections to going ahead with this. If someone is motivated to 
do specification work I do not like to discourage them with a hard 
rejection and it's helpful for them to learn the hard way that publishing 
an RFC does not result in useful deployment ;-).

		- Chris