[10:52:17] --- julian.reschke has joined
[10:52:23] --- julian.reschke has left
[11:14:50] --- Cary has joined
[11:28:49] --- julian.reschke has joined
[11:44:59] --- jlcjohn has joined
[11:50:02] --- Yangwoo Ko has joined
[11:50:02] --- Yangwoo Ko has left: Lost connection
[11:50:02] --- Yangwoo Ko has joined
[11:56:14] --- resnick has joined
[12:01:39] --- Yangwoo Ko has left: Computer went to sleep
[12:02:40] <julian.reschke> So what's the new location (...with respect to the audio broadcast)?
[12:03:25] <jlcjohn> The audio is dreadful: did Ted say apparea is moving to Spinnaker?
[12:04:56] <julian.reschke> Yes, seems to br Spinnaker.
[12:05:10] --- Yangwoo Ko has joined
[12:05:13] <julian.reschke> http://videolab.uoregon.edu/events/ietf/ietf673.m3u
[12:05:41] <jlcjohn> Audio much better there...
[12:08:53] --- pguenther has joined
[12:09:14] * pguenther resigns to being scribe
[12:09:23] <jlcjohn> thx
[12:09:41] --- memo57 has joined
[12:09:50] --- bernard.desruisseaux has joined
[12:10:12] <pguenther> spinnaker doesn't have a mic in the back, so questions may need to be repeated for the audio
[12:10:43] --- Lisa has joined
[12:11:02] <pguenther> kurt Z talks about LDAP standardization bits:
[12:11:12] --- shinta has joined
[12:11:20] <pguenther> LDAPbis finished revised base specs
[12:11:37] <pguenther> followup: interop reports, moving to to draft standard
[12:11:37] --- fujiwara has joined
[12:11:52] --- cnewman@jabber.org has joined
[12:11:58] <pguenther> RFCs 4510-4521 are the new RFCs
[12:12:14] <pguenther> unclear whether there will be a need to revise for draft
[12:12:20] <pguenther> ongoing work:
[12:12:30] --- patrik has joined
[12:12:31] --- tlyu@jis.mit.edu has joined
[12:12:34] <pguenther> - relaxRules control (don't enforce data model)
[12:12:58] <pguenther> - LDAP session tracking control
[12:13:10] <pguenther> - LDAP transactions
[12:13:29] --- hotta has joined
[12:13:38] --- fujiwara2 has joined
[12:14:14] <pguenther> discussions on the above and other extensions are still taking place on the ldapext mailing list
[12:14:52] <pguenther> related work:
[12:15:27] <pguenther> xml enabled directory (uses XML encoding instead of ASN.1 BER) draft-legg-xed-roadmap
[12:15:48] <pguenther> dhcp option for ldap directory server location
[12:15:57] <pguenther> loose ends:
[12:16:15] --- klensin has joined
[12:16:18] <pguenther> - distributed authentication and operations
[12:16:37] <pguenther> mapping of X.500 admin model into LDAP (to take advantage of ITU work in that area)
[12:17:30] --- klensin has left: Disconnected
[12:17:49] <pguenther> Bob Morgan: schema registry. identityschema.org has a start for ID related bits
[12:18:25] <pguenther> registries for atrribute names _are_ open and being updated
[12:18:34] <pguenther> BCP 64 covers the process
[12:19:18] --- Simon Josefsson has joined
[12:19:25] --- ryu.inada has joined
[12:19:36] --- ryu.inada has left
[12:20:01] --- harald has joined
[12:20:08] --- Glenn Parsons has joined
[12:21:03] <pguenther> John Klensin: review of IDNA
[12:21:22] <pguenther> proposal from small design team:
[12:21:31] <harald> the spinnaker room is where v6ops was to have been.
[12:22:07] <pguenther> - implement 'inclusion' model
[12:22:17] --- Lisa has left: Replaced by new connection
[12:22:33] --- fujiwara has left: Logged out
[12:22:35] <pguenther> - update to new unicode, avoid lots of issues with unicode canonicalization
[12:23:18] <pguenther> separate procedures for putting stuff in and those for taking them out (generate vs accept)
[12:23:23] --- Lisa has joined
[12:23:44] <pguenther> trying to avoid changes in interpretation: valid current will still be valid
[12:23:53] <pguenther> _or_ be invalid
[12:23:58] <pguenther> *not* a different valid string
[12:24:17] <pguenther> old invalid might be valid going forward
[12:24:23] <pguenther> read drafts for details
[12:24:51] <pguenther> major issues with ligatures and digraphes
[12:25:14] <pguenther> implausible ideas: put any word in any lang into DNS
[12:25:25] <pguenther> right-to-left problems
[12:25:48] <pguenther> currently prohibit some RTL usage
[12:26:53] <pguenther> particular examples: dhivehi, language of the Maldives
[12:26:55] <pguenther> Yiddishj
[12:26:59] <pguenther> -j
[12:27:22] <pguenther> one possible solution: allow for neutral direction at end of string
[12:27:33] <pguenther> other challenges
[12:28:08] <pguenther> invariants, etc
[12:29:10] <pguenther> most normalization & compat are treated as localization/OS/UI issue
[12:29:47] <pguenther> unicode versioning problem
[12:30:01] --- kdz has joined
[12:30:28] <pguenther> need procedures that are stable with version, as the app may not know or have chance to choose
[12:31:17] <pguenther> what's next?
[12:31:30] <pguenther> new stringprep definition
[12:31:58] <pguenther> needs stable NFKC (see doc)
[12:32:22] <pguenther> try to resolve differences with Unicore
[12:33:00] <kdz> Re XML Enabled Directories, additional information is available at <http://www.xmled.info>, list <mailto:xeddev@eb2bcom.com>, subscribe at <http://eb2bcom.com/mailman/listinfo/xeddev_eb2bcom.com/>.
[12:33:14] <pguenther> kdz: thanks
[12:33:30] <pguenther> reading list draft-*-idnabis-*
[12:34:07] <pguenther> http://stupid.domain.name/idnabis/draft-faltstrom-idnabis-tables-00.htm
[12:34:09] <pguenther> l
[12:34:21] <pguenther> (that has the .html suffix)
[12:34:33] <pguenther> idna-update@alvestrand.no
[12:34:42] <patrik> http://stupid.domain.name/idnabis/draft-faltstrom-idnabis-tables-00.html
[12:35:19] <pguenther> Q: is it guaranteed that all old labels with be valid new labels? Answer: no
[12:36:01] --- Aki Niemi has joined
[12:36:33] <pguenther> ICANN rules have always banned a number of names that IDNA permits, so they could appear at deeper levels
[12:38:08] <pguenther> PF: "should people that register domain names be worried?" A: only if they do really weird stuff
[12:38:49] --- Lisa has left: Logged out
[12:39:07] <pguenther> Tony: concerns about registrars enforcing rules
[12:40:00] <pguenther> A: 1) changes to IDNA technically take care of that
[12:40:34] <pguenther> 2) previously discussed rules for only using one script or one language per label are only enforcable at the registration level
[12:40:53] <pguenther> bad implementations are not preventable
[12:42:22] --- fenton has joined
[12:42:44] <pguenther> Ted: while we _could_ change the IDNA prefix, that would set everything back many years, so that definitely not a choice to make lightly
[12:43:43] <pguenther> Harald: EAI
[12:43:45] --- klensin has joined
[12:44:45] --- Lisa has joined
[12:44:46] <pguenther> some people find ascii hard to deal with, so let's specify an *experimental* set of changes to allow UTF-8 in email addresses and headers
[12:45:14] <pguenther> changes to SMTP, message format, POP3, IMAP, mailing list, etc.
[12:45:38] <pguenther> gateway spec for between SMTP and the "new world"
[12:47:04] --- psavola has joined
[12:47:52] <pguenther> mechanism: label for work UTF8SMTP
[12:48:08] <pguenther> use an SMTP extension to keep extended messages inside UTF8SMTP-aware context
[12:48:17] <pguenther> downgrade when leaving that environment
[12:49:16] <pguenther> specify address handling for downgrade: do _not_ specify an address conversion algorithm (ala punycode); instead, send along an "alternate address" that is ASCII
[12:49:34] --- barryleiba has joined
[12:49:43] <pguenther> framework doc nearly finished
[12:50:00] <pguenther> drafts for SMTP, headers, downgrade, IMAP, POP3...
[12:50:14] <pguenther> some early impl work, no landmines found
[12:50:51] <pguenther> Dave Crocker: submit not on the list of drafts?
[12:51:06] <pguenther> JohnK: authors of submit watching EAI closely
[12:51:33] <pguenther> Dave Crocker: downgrade _usually_ means loss of information. Is that true here?
[12:51:35] --- hardie@jabber.qualcomm.com has joined
[12:51:40] <pguenther> HA: "maybe"
[12:51:45] <pguenther> currently under discussion
[12:51:58] <pguenther> can we add enough information to recover original content
[12:52:19] <pguenther> controversial bits:
[12:52:33] <pguenther> * alt-address vs 'encoded' address
[12:52:54] <pguenther> * encapsulation vs conversion at downgrade - decided on conversion
[12:53:24] <pguenther> * exact format for alt-address representation unsettled
[12:53:45] <pguenther> current favorite: <utf8@utf8 <ascii@punycode>>
[12:54:10] <pguenther> timeline: now aiming for before March
[12:54:20] <pguenther> need more experimenting
[12:54:57] <pguenther> EAI meets on Thursday afternoon
[12:55:53] --- Simon Josefsson has left
[12:55:54] <pguenther> alexey melnikov: new stuff in email (except DKIM or EAI(
[12:55:55] <pguenther> )
[12:56:50] <pguenther> two WGs doing IMAP stuff: IMAPEXT and LEMONADE
[12:57:41] <pguenther> IMAPEXT: already did ACLs, finishing up sorting, threads, annotations, i18n
[12:58:17] <pguenther> LEMONADE: enhancements for clients with slow or long links
[12:58:25] <pguenther> - forward-without-download
[12:58:32] <pguenther> - pushing of notifications
[12:58:39] <pguenther> - content conversion
[12:59:14] <pguenther> Lots of stuff done in IMAPEXT
[12:59:21] --- kdz has left
[12:59:40] <pguenther> ACL, UIDPLUS revision, CONDSTORE, forward-without-download "trio"
[13:00:03] <pguenther> <picture of forward-without-download>
[13:00:28] --- pigdog has joined
[13:01:11] <pguenther> forwarding can apply to just part of a message, particular MIME parts/attachments, etc
[13:02:01] <pguenther> demo at interop event: 64MB of messages forwarded in 12 seconds
[13:03:14] <pguenther> imap: URL doc being refreshed, ugly bits include relative URLs and how they work in corner cases
[13:05:46] <pguenther> continuing work: i18n and localization
[13:06:07] <hardie@jabber.qualcomm.com> We're half way through here. Will folks be asleep by the end?
[13:06:27] <hardie@jabber.qualcomm.com> 11/21 slides, that is
[13:07:00] <pguenther> TLS compression support is poor, so application protocol-level compression has been defined for IMAP
[13:07:08] <pguenther> does that also need to be done in SMTP?
[13:08:12] <pguenther> quick reconnect handling for IMAP
[13:08:32] --- hardie@jabber.qualcomm.com has left: Replaced by new connection.
[13:08:51] --- Antoin has joined
[13:10:26] --- mankin has joined
[13:12:03] <pguenther> discussions for persistent searches for stuff like views
[13:13:06] <pguenther> in-band notifications
[13:15:55] --- rlbob has joined
[13:17:36] <pguenther> open problem: IMAP is MIME aware, but has no ability to do S/MIME processing on the server
[13:18:41] <pguenther> i.e, can't do verification of signatures on the server or decrypt a part on the server to walk into the nested MIME structure
[13:18:55] <pguenther> makes forward-without-dowload not work with S/MIME
[13:19:45] <pguenther> <brief summary of sieve>
[13:19:57] <pguenther> Barry Lieba status of DKIM
[13:21:42] <pguenther> Lisa: what are the chokepoints in the current email development work
[13:22:35] <pguenther> comparators was one, now what?
[13:23:08] --- pigdog has left
[13:23:10] <pguenther> johnK: utf8text doc may be one
[13:23:26] <pguenther> JohnK: perhaps a small email directorate?
[13:23:59] <pguenther> Eliot Lear: IANA and port assignments
[13:24:30] <pguenther> current procedures are crutfy
[13:24:57] <mankin> could you report the hands raised?
[13:25:16] <pguenther> 10? 14?
[13:25:27] <pguenther> (I was facing the wrong direction)
[13:25:30] <patrik> I think most people (including myself) was ironic when rising hand.
[13:26:18] <pguenther> "well known" port concept is the wrong questions and pointless as is
[13:26:48] <mankin> comment - this is like a number of old registries
[13:26:58] <pguenther> current registry entries are out of date
[13:27:26] <pguenther> goals: we want port usage documented in terms of what is running on that port
[13:27:41] <pguenther> we want broadly deployed protocols to avoid conflicts
[13:27:50] <pguenther> where approrpriate, recommend/use SRV records
[13:28:36] <pguenther> SRV records: similar to MX but not tied to SMTP and with extra 'weight' field
[13:29:07] <pguenther> "SRV good": no port allocation required, allows prioritization
[13:29:54] <pguenther> "SRV bad": DNS dependency, could be additional service delay
[13:30:00] <mankin> use mic or repeat question
[13:30:07] <pguenther> DaveC: firewall difficulties with SRV?
[13:30:16] <pguenther> Eliot: not really
[13:31:37] <pguenther> comment from someone (can't see): firewalls not an issue for the server
[13:31:56] <pguenther> Thomas Narten: underestimating problems with firewalling SRV
[13:32:53] <pguenther> Fenton: need registry for service *names*? yes
[13:33:31] <pguenther> Peter ?: need specification of the SRV record and usage
[13:33:57] <pguenther> LisaD: sounds like a discussion for the discuss@apps.ietf.org list
[13:34:12] <pguenther> eliot resumes...
[13:34:28] <Lisa> I wanted to get somebody on the spot to start that discussion on the list; any takers?
[13:34:34] --- dcrocker has joined
[13:34:48] <pguenther> service documentation: would be good to have a pointer to something describing the bits
[13:35:06] <pguenther> RFC 3205 is still valid: better to use port numbers rather than HTTP
[13:35:17] <pguenther> draft-lear-iana-no-more-well-known-ports-02.txt
[13:36:18] <mankin> please repeat the question
[13:36:31] <pguenther> me: where is template for SRV usage in specs? EL: 2782
[13:36:33] <mankin> please repeat the question
[13:36:54] <mankin> please repeat the question
[13:36:55] <pguenther> chris newman: SASL/GSS has registry for service names,
[13:37:23] <pguenther> should we be merging those?
[13:37:26] <mankin> thanks
[13:37:31] <pguenther> (but KurtZ shakes his head 'no')
[13:38:21] <pguenther> ----
[13:38:26] <pguenther> tv: URIs
[13:38:35] <pguenther> drafpt-keesmaat-tvreg-00
[13:38:50] <pguenther> identifying TV broadcasts
[13:39:11] --- memo57 has left
[13:39:18] <pguenther> use cases: messaging avout TV "Watch CNN now"
[13:39:34] <pguenther> buddy watching "Bob is watching ARD1"
[13:39:44] <pguenther> (presence in relation with TV watching)
[13:40:11] <pguenther> other forms of communcation related to TV broadcasts: scheduling and webpage references
[13:40:23] <pguenther> DaveC asks...l
[13:40:56] --- m_ersue has joined
[13:41:21] <pguenther> are first two cases the same?
[13:41:30] <pguenther> same purpose for use, but different generation
[13:41:42] <pguenther> (human vs programatic)
[13:41:49] <pguenther> requirements:
[13:42:08] <pguenther> - identier -> broadcast mapping must be unique
[13:42:32] <pguenther> - few ids per broadcast (broadcast -> id mapping SHOULD be unique)
[13:43:00] <pguenther> - ids should cover all forms of TV broadcast, not just cable + airvwaves, etc
[13:43:13] <pguenther> - usable in internet communication (IM, calendar, etc)
[13:43:27] <pguenther> - easy to use by humans
[13:44:10] <pguenther> (i.e., _not_ the existing VCR plus or the european similar)
[13:44:34] <pguenther> kurtZ: first and last requirements are conflicting
[13:45:02] <pguenther> johnK: goes back to the use case slide
[13:45:34] <pguenther> johnK: if we say "watch CNN now!", that conflicts with the lagging of programs between areas
[13:45:45] <pguenther> fixes for that conflict with easy-of-use constraints
[13:46:18] <pguenther> DaveC: really cool effort, but!
[13:46:34] <pguenther> CNN in hong kong != CNN USA
[13:46:40] <pguenther> references are too generic
[13:46:56] <pguenther> simplisitic forms vs elaborated forms?
[13:47:42] <pguenther> local references vs global references;
[13:48:06] <pguenther> patrickF: how does this relate to 2838?
[13:48:21] --- barryleiba has left
[13:48:40] <pguenther> not required to be resolvable
[13:49:19] --- mankin has left
[13:49:30] <pguenther> no way to figure out which ones are valid
[13:51:09] <pguenther> or what is the meaning of a given URI (?)
[13:51:18] <pguenther> requirements for registry:
[13:51:28] <pguenther> atteaches identification info to URIs
[13:51:38] <pguenther> can be searched/browsed by humans
[13:51:54] <pguenther> should be implementable even without involvement of TV broadcast parties
[13:52:30] <pguenther> (nice to have): mappable to local TV broadcast frequencies, channels, IP address, etc
[13:55:41] <pguenther> issues:
[13:55:55] <pguenther> structure of the registry, form of records
[13:56:03] <pguenther> - process: who can register what
[13:56:15] <pguenther> - (business) organization of the registry
[13:56:43] <pguenther> - relation with the DNS?
[13:57:01] --- patrik has left
[13:57:03] <pguenther> - mappability to location info
[13:57:57] <pguenther> Paul Hoffman: proposal 3: don't do it in the IETF
[13:58:37] <pguenther> there are companies to whom this would be worth tens of millions of dollars
[13:59:12] <pguenther> if they haven't do it, it's because it's impossible or had problems...
[13:59:19] <pguenther> s/had/has/
[14:00:00] <pguenther> DaveC: again, is the IETF the right place? feels like a yes, but it's not clear who would consume the output, so don't know what the form of the output should be
[14:00:04] --- kdz has joined
[14:00:21] <pguenther> No one is saying "I need this tomorrow", so it's just a matter of thrashing around
[14:00:36] <pguenther> until someone shows up who wants to build on the output
[14:01:15] <pguenther> tedH: TV anytime forum registered a URN namespace for program guide bits
[14:02:23] <pguenther> tedH: different enough that this is a blank slate work
[14:04:54] <pguenther> discussions of this has been on the apps area list
[14:05:45] <pguenther> open mike, starting with: ORCHIDs
[14:06:04] <pguenther> look like IPv6 addresses, but aren't really
[14:06:09] <pguenther> for HIP experiementation
[14:06:51] <pguenther> a /28 allocation for use in shims
[14:07:59] <pguenther> meta-experimentation: doesn't name what they are for, but given one you can't tell what a given one is for
[14:08:32] <pguenther> putting these in an SDP offer or URI would cause confusion
[14:08:41] <pguenther> is this a good way to approach this experiment?
[14:08:47] <pguenther> IETF last call is ongoing
[14:09:50] <pguenther> DaveC: when this first came up, pre-SHIM6, this was defining a private address-space
[14:10:06] <pguenther> where it was private to the machine itself
[14:10:54] <kdz> Regarding my comment, I may have been confused. I assumed that author was asking for the IETF to establish a registry (at IANA) and hence my comment that this seemed problematic as the registry serves end-users not protocol engineers.
[14:11:33] --- m_ersue has left
[14:11:48] <pguenther> johnK disagrees with TedH's characterization for his position
[14:12:10] <psavola> What I'd be interested in hearing is that if this is not a good way to approach this experiment, what would be a better way? In the comments on the mailing lists I haven't seen any suggestion.
[14:13:00] <Yangwoo Ko> http://www.ietf.org/internet-drafts/draft-laganier-ipv6-khi-05.txt
[14:13:24] <klensin> I think failures are fine, if they are _very_ predictable. DNS lookup failures are a reasonable example. Failures that are dependent on the properties of particular local machine configuration are unpredictable across users and time, and are, IMO, really bad news.
[14:13:45] --- Lisa has left
[14:14:56] <pguenther> DaveC: email architecture doc is at -05, will be a -06
[14:15:13] --- Antoin has left
[14:15:41] --- dcrocker has left
[14:15:59] <pguenther> adjourned
[14:16:19] --- pguenther has left
[14:16:28] --- bernard.desruisseaux has left
[14:16:33] --- julian.reschke has left
[14:16:58] --- Yangwoo Ko has left: Computer went to sleep
[14:17:14] --- shinta has left
[14:17:32] --- klensin has left
[14:17:40] --- harald has left
[14:17:44] --- Aki Niemi has left
[14:17:47] --- cnewman@jabber.org has left
[14:18:07] --- kdz has left
[14:18:20] --- resnick has left
[14:21:25] --- rlbob has left
[14:22:12] --- Cary has left
[14:22:53] --- hotta has left
[14:29:51] --- fenton has left
[14:31:17] --- tlyu@jis.mit.edu has left
[14:31:26] --- Glenn Parsons has left
[14:41:05] --- psavola has left
[15:51:53] --- harald has joined
[15:51:53] --- harald has left
[15:53:27] --- fujiwara2 has left
[16:15:15] --- jlcjohn has left