To:
ietf-provreg@cafax.se
From:
Andrew Sullivan <andrew@ca.afilias.info>
Date:
Fri, 12 Aug 2005 10:21:59 -0400
Content-Disposition:
inline
In-Reply-To:
<046F43A8D79C794FA4733814869CDF07C927A9@dul1wnexmb01.vcorp.ad.vrsn.com>
Mail-Followup-To:
Andrew Sullivan <andrew@ca.afilias.info>,ietf-provreg@cafax.se
Reply-To:
Andrew Sullivan <andrew@ca.afilias.info>
Sender:
owner-ietf-provreg@cafax.se
User-Agent:
Mutt/1.5.9i
Subject:
Re: [ietf-provreg] Services messages in RFC 3730
Hi Scott, Thanks for your comments. On Fri, Aug 12, 2005 at 07:06:00AM -0400, Hollenbeck, Scott wrote: > > I don't see how you can interpret the queuing system as optional. The > text is completely clear about the need to create and enqueue the > messages. "dequeue and deliver messages" does NOT mean "route to > /dev/null". It can mean that messages be delivered via email, made > available for retrieval via ftp, etc. > > "dequeue and deliver". Not "throw away blindly". You're right that I'm probably overstating the case. But consider cases where the client just doesn't do anything in time for the notification to have been any use. For instance, in the ICANN world, we have 5 days in which a losing registrar may refuse a transfer (subject to a bunch of policy rules). If a transfer request message hasn't been picked up after five days, it's worthless. Worse, it's clogging the message queue, and delaying the delivery of other, potentially useful messages. The only sane thing to do with such a message is to throw it away. As a result of this sort of thing, as near as I can tell, at least some clients are completely ignoring the poll queue. We therefore have to support other kinds of notification mechanisms in parallel to the poll queue, which is a pointless duplication of effort. And as near as I can tell, this is because people find the poll queue less useful than alternative mechanisms. I was hoping to make it more useful. > What defines "explicitly required", though? I thought Jim's definition > was pretty reasonable: if you're a party to a transaction, but not the > recipient of a response, you get a message. We can't just limit > ourselves to transactions that cost money in the ICANN TLD world. No, I totally agree; and I thought Jim's definition was a good one, too; I support it. I absolutely do not want to propose anything that is restricted to the ICANN TLD world, because I think EPP has lots of potential outside that world. In fact, my view is that there are still some things in the protocol and mapping documents that are too ICANN-TLD-centric, and turn out to be more policy shaped than protocol shaped. (Just so I'm clear: I think that's an accident, and it's just the result of the collective experience of those who were working on the original protocol definition at the time the WG was active.) > On the priority-based multiple queue thing: priority is relative. It > typically either gets ignored or everything becomes "high" priority in Hrm. This makes me think that "priority" is an infelicitous word choice. What about "facility"? I'm basically thinking of something like syslog on UNIX here. You have some arbitrary tags that tell you what is generating the message; and that tag allows you to apply additional rules, depending on the tag. It's sure useful to sysadmins; I sort of think it might be useful to clients of a repository as well. > someone's eyes. I don't get a good feeling about adding complexity to > the mechanism if making it simpler is a viable option. What about the notion that you could give everything the same priority (or facility, or whatever we want to call it) value, so that you'd have something just like the current mechanism? For that matter, the attribute could be completely OPTIONAL, and its absense an indication that the client wants the next message in the queue regardless of facility. In that case, those who have implemented the current mechanism and don't want to do any more work wouldn't have any work they need to do. I could be, of course, completely wrong about the utility of a mechanism like this; I just know that people seem not to be using the poll queue as I'd like, and I'm trying to figure out how to persuade them to use it (so I can stop sending reams of email). A -- ---- Andrew Sullivan 204-4141 Yonge Street Afilias Canada Toronto, Ontario Canada <andrew@ca.afilias.info> M2P 2A8 +1 416 646 3304 x4110