Comments on draft-ietf-provreg-epp-beep-02.txt

[Darren New <dnew@san.rr.com>]

I looked over draft-ietf-provreg-epp-beep-02.txt briefly. I noticed a few
minor items I thought would be worth addressing...

2.
  "Fully processes?" Note that some sort of
  answer needs to be returned for every BEEP
  message. You might want to make this clear,
  or perhaps it's already implicit in the 
  description of <logout> in the other documents.

2.1.2 
  Consider service "epp_beep" since EPP runs over
  other TCP-based transports as well.

  In the example, the second L: should be I:

2.1.4 
  Text "EPP" is duplicated. "Sent over the EPP EPP session"

  I'm not sure you should specify that piggyback data must
  be piggybacked. This is more an optimization than a semantic
  difference. Perhaps "MAY be piggybacked" is better. Note that
  some toolkits don't even tell you whether the data was
  piggybacked or not.

2.1.7
  You might want to mention what the semantics of a single EPP 
  channel being closed without a <logout> is supposed to do.
  Something like "this is equivalent to a <logout> with no
  attributes or child elements." You might also want to mention
  explicitly that it's OK to close the BEEP channel after the
  <logout> has been fully processed, in order that resources
  could be reclaimed.

Thanks for your attention!

-- 
Darren New 
San Diego, CA, USA (PST). Cryptokeys on demand.