IETF
precis@jabber.ietf.org
Friday, November 14, 2014< ^ >
yone has set the subject to: précis wg :) | http://tools.ietf.org/wg/precis/ | IETF 90 | slides at https://datatracker.ietf.org/meeting/90/materials.html#precis | audio at http://ietf90streaming.dnsalias.net/ietf/ietf905.m3u
Room Configuration
Room Occupants

GMT+0
[18:58:02] lminiero joins the room
[19:07:26] Yoshiro Yoneya joins the room
[19:11:58] Yoshiro Yoneya has set the subject to: precis wg :) | http://tools.ietf.org/wg/precis/ | IETF 91 | slides at https://datatracker.ietf.org/meeting/91/materials.html#precis | audio at http://ietf91streaming.dnsalias.net/ietf/ietf916.m3u | meetecho at http://www.meetecho.com/ietf91/precis
[19:17:43] Sam S joins the room
[19:18:19] Simon Romano joins the room
[19:18:20] Sam S leaves the room
[19:31:43] ben joins the room
[19:51:44] Simon Romano leaves the room
[20:03:24] ben leaves the room
[20:03:41] ben joins the room
[20:05:28] marc.blanchet.qc joins the room
[20:05:37] ben leaves the room
[20:05:57] Barry Leiba joins the room
[20:12:11] Andrew Sullivan joins the room
[20:13:46] ben joins the room
[20:14:39] Meetecho Feed joins the room
[20:15:15] stpeter joins the room
[20:15:33] <stpeter> we start in 15 minutes, right?
[20:15:50] <Yoshiro Yoneya> @stpeter, yes
[20:17:59] Lorenzo Miniero joins the room
[20:18:50] <lminiero> stpeter: you there for a quick test?
[20:18:52] resnick joins the room
[20:20:31] <stpeter> lminiero: yes
[20:20:32] Chris Newman joins the room
[20:20:43] <lminiero> did we send yuou the link already?
[20:21:10] <stpeter> yes, let me look at the email message
[20:21:29] <lminiero> just sent it privately :)
[20:21:46] Peter SaintAndre joins the room
[20:22:36] <Lorenzo Miniero> you can unmute using the Play icon next to your name, once you accept the webrtc permissions
[20:23:13] hildjj joins the room
[20:23:34] <Lorenzo Miniero> about the wild audio: that was me. sorry :)
[20:24:01] Joe Hildebrand joins the room
[20:24:40] <marc.blanchet.qc> https://tools.ietf.org/wg/precis/minutes
[20:25:03] jimsch1 joins the room
[20:25:42] <resnick> Or here: http://etherpad.tools.ietf.org:9000/p/notes-ietf-91-precis
[20:27:06] <resnick> We can hear you.
[20:27:21] Tony Hansen joins the room
[20:27:37] Jim Galvin joins the room
[20:27:52] <Peter SaintAndre> yay
[20:28:05] <Tony Hansen> wow, THREE video streams :-)
[20:28:22] <Peter SaintAndre> Lorenzo: Meetecho thinks that "-" is now allowed in a username :-)
[20:28:41] <Lorenzo Miniero> ok we'll fix that as well
[20:28:41] <stpeter> er, not
[20:29:05] hildjj is now known as -
[20:29:19] <-> I am the dash.
[20:29:27] <marc.blanchet.qc> we applied unicode normalization rules to the sound system and the output was silence.
[20:29:45] <Lorenzo Miniero> :)
[20:30:18] <Peter SaintAndre> I will mute except when I am speaking
[20:30:19] - is now known as —
[20:30:22] <Lorenzo Miniero> ok
[20:30:29] <—> In RFC's you can refer to me as M. Dash.
[20:30:35] <Barry Leiba> Speak, except when you're muting.
[20:30:40] — is now known as hildjj
[20:30:46] <Barry Leiba> And please prefix with "MIC" if you want me to speak for you.
[20:30:53] <hildjj> things have gotten quite punchy in the HNL Hilton today.
[20:31:00] <Peter SaintAndre> :)
[20:35:11] John Klensin joins the room
[20:36:32] <resnick> Is John Klensin listening in?
[20:37:06] <John Klensin> @pete: No :-(
[20:39:26] <resnick> :-/
[20:41:09] <John Klensin> (lost audio (again).   gong to need to log back in)
[20:41:19] John Klensin leaves the room
[20:41:42] John Klensin joins the room
[20:42:41] <John Klensin> MIC: The main reason I've gone silent about this subject  is that the experience with a newly-added code point in Unicode 7 that could not be captured by any rules we knew enough to set up.  Part of that was misunderstanding of the Unicode rules for assigning code points, but that doesn't make the problem any less.
[20:43:16] <John Klensin> (contd) If PRECIS wanted to address that, please propose the rule that would have identified that code point and issue.
[20:43:44] <John Klensin> (and, much as I like Meetecho, it isn't ready for prime time.
[20:45:30] john.levine joins the room
[20:45:47] alessandro amirante joins the room
[20:46:15] <Andrew Sullivan> John: did that come through?
[20:46:27] <John Klensin> Mic: It is a disaster.  We shouldn't stop work and failure at the margins is likely to occur.  But we should;t, it at all possible, end up with IDNA and PRECIS handling those edge cases separately -- or one handling them and the other not.
[20:47:11] John Doe joins the room
[20:47:33] <Andrew Sullivan> @John: do you think that the two things are going to handle them differently?
[20:47:44] Suz joins the room
[20:47:47] John Doe leaves the room
[20:48:04] <Andrew Sullivan> Because my reading is that they produce approximately the same result (modulo profile things)
[20:48:28] <John Klensin> That is more or less agreement with Andrew, iwth the qualification that I don't want to trap ourselvs by over-reliance on algorithms (or, worse, code samples).   Doing so implies the risk of separate directions that Pete asked about
[20:49:14] <Andrew Sullivan> I don't get it, though — that sounds like a reason to go back to lists of code points
[20:49:31] <John Klensin> (dno't waste Mic time, but I have no video at present and the audio seems to stop (and only sometimes spontaneously restart)
[20:50:14] <John Klensin> @Andrew: not unless we pick a version of Unicode and freeze on it.  And we know where that leads.
[20:50:39] <Andrew Sullivan> Well, right
[20:50:49] <Andrew Sullivan> Reductio, therefore algorithm.
[20:51:01] alessandro amirante leaves the room
[20:51:18] Pau Boemio joins the room
[20:51:29] alex amirante joins the room
[20:51:44] <John Klensin> PRECIS should not, IMO, wait on this, but Patrik and I agreed last week how to present the options about that problem code point and its implications to the commuhity,  I'm holding the pen; I-D probably next week.
[20:52:49] <John Klensin> The algorithm problem, stated in algorithmic terms, is that the new code point issue strongly suggests that the rules on which such an algorithm might be based, and hence the algorithm itself, are not NP-complete.
[20:53:40] <hildjj> unicode folks need to highlight these to us.  they know when they are making problems.
[20:54:24] <hildjj> as proved by the doc they wrote having to justify the character in question.
[20:55:11] <marc.blanchet.qc> « unicode folks » claim that this « issue » is also for many other code points from quite some time and therefore, it is not an issue, but just the way it works.
[20:55:18] <John Klensin> @Joe: I agree.  But, with this case as an example, they have chosen "denail" and a lot of assertions about languages instead.  The problem, at root, is not only does their spec say that they won't do that, but they prmoised us during the design of IDNA2008 that they wouldn't.
[20:56:01] <hildjj> still an issue for the liaison in my opinion.
[20:56:41] <hildjj> and if that doesn't work, we need to escalate.
[20:57:07] <John Klensin> @Marc:  Not quite "many".  Apparently order a large handful.  But, yes, the other version of tghe story is "there are a bunch of them already and, even though some of us told you earlier ther ethere were so such cases, we can now use them as justification for adding more of them"
[20:57:46] Pau Boemio leaves the room
[20:58:18] <marc.blanchet.qc> what I am trying to say is that from the Unicode side, they don’t see this as an issue, therefore most likely not trying to do anything about it in the future.
[20:58:22] <John Klensin> @Joe: excalate to whom?  (Let's discuss offline -- tryng to follow what Peter is saying)
[20:58:32] <hildjj> nod
[20:58:42] <John Klensin> @Marc: certainly agree.
[20:59:20] <marc.blanchet.qc> therefore, we shall think a solution based on this assessment.
[20:59:37] <John Klensin> ... and, Joe, that is the problem with "escalation".  Regardless of the analysis of how we have gotten here, they aren't willing to acknowledge that there is a problem and that makes escalation meaningless.
[20:59:55] <Andrew Sullivan> @Marc: I think John's point is that it isn't _possible_ to have a solution if Unicode only sometimes follows its own rules
[21:00:19] <John Klensin> @Andrew: approximatelyy.  yes.
[21:01:04] <John Klensin> At least short of NFI[ETF]
[21:01:05] <hildjj> we have to get them to at least flag these cases for us when they use this BS precedent justification.
[21:01:54] <John Klensin> @Joe: again, let's take this up later.  The problem, as Marc and I seem to agree, is a little harder than that.  Perhaps a lot harder.
[21:03:04] <John Klensin> I'm told of a rather heated discussion from which the conclusion by one of the parties was that the other one just did "get" the idea of not having language information when one used or interpreted a character/ code point
[21:16:31] <John Klensin> MIC: I thihk the IETF Last Call was symptomatic of a very deep set of issues with i18n reviews in the IETF.  The odds that you will get significantly more informaiton are low, but that puts a _lot_ of burden on this WG (and maybe an appeal to the semi-dead IAB i18n program) to do a careful review that addresses the issues carefully.  If someone, ideally Peter and/or the Chairs, could put together a list of issues that people should be especially checking, that would  make me, and probably the IESG, etc., feel a lot more complicated.
[21:16:47] <John Klensin> s/complicated/ comfortable/  :-(
[21:19:22] Florian Zeitz joins the room
[21:19:54] <John Klensin> @Andrew" just a symptom of zombie-dom.
[21:20:56] alex amirante leaves the room
[21:21:13] ale amirante joins the room
[21:29:23] <hildjj> (btw, this was one of the reasons i wanted NFD instead of NFC.  Even though there's an algorithm for checking if a string is NFC, it takes a lot more power than doing the same check for NFD)
[21:29:38] <hildjj> (not trying to re-open that consensus!)
[21:29:48] <resnick> We lost you Peter.
[21:29:50] <Andrew Sullivan> audio seems to have failed, Peter
[21:30:09] <Peter SaintAndre> ok
[21:30:23] <Peter SaintAndre> I said we don't need to change saslprepbis
[21:30:24] Meetecho Feed leaves the room
[21:30:26] jimsch1 leaves the room
[21:30:39] <Peter SaintAndre> but we can discuss in the HTTPAUTH WG
[21:30:45] <resnick> ack
[21:31:23] <stpeter> I think Meetecho cuts you off when the WG session is scheduled to end :-)
[21:31:36] Barry Leiba leaves the room
[21:31:36] ben leaves the room
[21:31:47] Peter SaintAndre leaves the room
[21:31:51] <hildjj> peter, can you come down for lunch first week of december, and we spend the afternoon geeking out on this stuff?
[21:31:51] Suz leaves the room
[21:32:05] <marc.blanchet.qc> thanks Peter?
[21:32:14] <marc.blanchet.qc> really appreciated.
[21:32:23] <stpeter> we are done??
[21:32:27] <hildjj> yes.
[21:32:30] <stpeter> ok!
[21:32:32] <Yoshiro Yoneya> done
[21:32:43] Jim Galvin leaves the room
[21:32:53] <stpeter> I will see you all in HTTPAUTH after your break :-)
[21:32:56] <John Klensin> better be, because my audio just joined by video and slide view in never-never-land
[21:33:01] ale amirante leaves the room
[21:33:11] John Klensin leaves the room
[21:33:19] Chris Newman leaves the room
[21:33:27] <stpeter> Joe Hildebrand: sounds plausible
[21:33:32] Andrew Sullivan leaves the room
[21:33:40] Lorenzo Miniero leaves the room
[21:33:52] Yoshiro Yoneya leaves the room
[21:34:27] lminiero leaves the room
[21:34:54] Joe Hildebrand leaves the room
[21:36:17] marc.blanchet.qc leaves the room
[21:36:22] john.levine leaves the room
[21:36:39] Tony Hansen leaves the room
[21:36:58] resnick leaves the room
[21:39:09] jimsch1 joins the room
[21:39:22] jimsch1 leaves the room
[21:46:19] hildjj leaves the room
[21:57:29] Florian Zeitz leaves the room: offline
[22:50:13] stpeter leaves the room
[23:33:59] marc.blanchet.qc joins the room
[23:50:54] Suz joins the room
[23:54:20] Suz leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!