Document: draft-ietf-xmpp-im-22.txt Reviewer: Spencer Dawkins Date: April 28, 2004 This draft is probably the highest-quality draft I've reviewed for Gen-Art. I'm not a genius of XML, and I could implement from this draft. It's very clearly written. I do wish the authors could find a more succinct form of exposition - although well-structured, this is still more than 100 pages of "in this situation, which is slightly different from the previous situation, do almost the same things you did in the previous situation, watching closely for subtle differences" - anyone who played Adventure during the 1980s will find the similarities to "you are in a maze of twisty passages, all alike" to be striking. But if the WG didn't figure this out in 22 revisions, there's probably not an OBVIOUS way to restructure. Almost ready for publication as a Proposed Standard. I have one question, and some very minor editorial nits. Q: Section 5.1/page 21 - the user's server MUST NOT reply to subscription requests on behalf of the user, but MUST direct all subscription requests to the user's client - I probably spend too much time worrying about wireless networks, but I worry when I see a server that can accept requests at gig-E line speed that must be forwarded to a GPRS client at 40 Kb/s. I'm not saying this is an invitation for DDOS amplification, but the security considerations don't refer to problems like this - omission, or "not required because these protocols are carried on TCP anyway"? N: "subscribee" is probably OK, but if you can think of a real word, please feel free to substitute. N: "Happy path" is just weird-sounding, like a Chinese takeaway entree, and can't be ESL-friendly. "Success path"? But please feel free to substitute something. N: There are many cross-references using full section titles (in addition to section numbers), but the section titles aren't quoted, only capitalized. It's not always easy to pick out where the title starts and stops. N: [CPIM] is expired. N: There is not a single word of description in all of Appendix B, which was jarring, considering that the rest of the draft used narrative descriptions in place of formal specifications... Spencer