LAGER WG virtual interim 2015-08-24, 1400 UTC Minutes taken by Paul Hoffman Text from the presentations not given, just the discussion See the slides for the full presentations https://www.ietf.org/proceedings/interim/2015/08/24/lager/proceedings.html Blue sheet: Asmus Freytag Audric Schiltknecht Barry Lieba Francisco Arias Jiangkang Yao Kim Davies Marc Blanchet Mats Dufberg Paul Hoffman Peter Koch Scott Hollenbeck Wil Tan Yoshiro Yoneya Agenda was covered draft-ietf-lager-specification, Kim Davies Gave overview presentation Issues on the WG tracker, Kim Davies Went through each of the open issues in the WG tracker #2 - Mechanism to specify origin URI and refresh schedule Paul: Maybe just a URL, not all the data requested Asmus: Having a standard field would be good Kim: This is a link of where to fetch this file in the future Asmus: Not baked enough Marc: Put this on hold, maybe abandon #3 - Clarify domain element as it pertains to Root Kim: How is the root zone represented? Already text in next draft #4 - Review syntax of actions Wil: Likes "allocatable" and "blocked" Asmus, Yoshiro: Ditto Kim: Will change in next draft #5 - Add text about potential effects of recursive rules Asmus: The draft does not intend to allow recursive definitions Need to make sure that the wording does not allow that Paul: Add wording that say recursive is not allowed Asmus: OK with adding such wording Kim: Does this limit parser implementations? Asmus: Probably not Wil: His implementation parses entire XML, so it might be susceptible to recursion Asmus: Can be a corollary or a restatement Asmus will provide text for this issue #6 - Should there be a registry of dispositions? Paul: Don't need a registry, just have RFCs update this Marc: This is so specific now, and the registry will never change Kim: Happy to withdraw this, leave as-is Asmus: We don't know what additional disposition values might be Peter: This is preparing for versioning of protocol, not just asking for extensions Kim: How you act on dispositions is undefined Wil: Wants an registry to prevent overlap of names Maybe just a namespace, but can later create a registry Asmus: A registry seems more germane Agreement between authors of LGR documents and recipients of those LGRs Wil: Thinks a namespace is sufficient Marc: Add a section on extensibility Paul: People don't want to registered Asmus: Namespace solves collision problem, but not commonality between extensions Could live with "anything that is not prefixed with a namespace must come from an updated RFC" Kim will digest what was said and propose new wording Wil: Wants us to think about how people will extend #7 - Was already closed #8 - Just a typo #9 - Fix domain/scope usage Kim: Update all the examples #10 - Minor, will be fixed #11 - Rule on when/not-when attributes Asmus will write language to fix this #12 - Error, will be fixed #13 - 2 syntax for ranges Asmus: Was noted that there were too many uses of the element because the attributes vary Rules are regular expressions, and char class can be done in a similar way Thus not using char class for this Kim: Should we be more explicit about why we chose this? Asmus will add a note that says a char class is not a pseudo-repertoire, and will flesh this out more #14 - Typographical error, remove some spaces #15 - Clarify when/not-when use on char/range Asmus: This seems like a duplicate of #11 Focuses on range, strongly connected Asmus will fix with same clarification as #11 New issue, not yet in tracker, from Asmus If there are a forward and reverse context, and authors like to conflate these This makes writing a tool that makes sure that an LGR is symmetric and transitive Suggests to add text: don't use this type of shorthand, it makes it hard on us This was triggered by Arabic LGR Asmus has proposed language, and will write into the next draft Marc: When the -01 is out, authors should update the tracker items draft-schiltknecht-lgr-json, Audric Schiltknecht Just asking for interest from the community, not asking it to be adopted by the WG Wil: The conversion rules look really simple Looks like it is optimized for simple conversion, but this makes indexing more complicated Audric: Has some code samples to show how he processes Asmus: In the rules section, there are weak ordering requirement that items have to be defined before they are used So array seems reasonable for rules element Wil: Maybe can optimize with sub-elements Audric: Tried to stick to the XML for round-tripping Marc: Should this be a WG document? Paul: Please no Asmus: Ditto Marc: Let's keep this as an individual item draft-wyk-lager-cjk-terminology, Yoshiro Yoneya The terminology in the main draft is confusing, so they wrote down their understanding Paul: Hopes that the definitions that are already in draft-ietf-lager-specification are not being changed Asmus: The term "script" is used in many ways for the root LGR There was a lot of misunderstanding from the panels The definition of "script" should be somewhere This definition is also the shared understanding Marc: Concerned about this being out of scope of IETF WG Not sure what we will do with this document with respect to the WG Next virtual intermim meeting Maybe September 21 But will depend on how the -01 document looks