Presenters: Paul Hoffman, John Levine
Draft: https://datatracker.ietf.org/doc/html/draft-rswg-xml2rfcv3-implemented/
Objective: Current status, Issue Review
Toerles Eckert: Would it be useful to have in the text itself a set of semantics
John Levine: No it would not be useful.
Julian Reshke: The presentation info has been there forever. The change thqat was made was formatting according to the country code.
With respect to the RFC number the information is already there in the data tracker.
John: Then we have a dependency on datatracker.
Alexis Rossi: I think it is useful to keep the link to datatracker.
Martin Thomson: If the datatracker goes away, so does the link target.
Robert Sparks: Point is the data has been captured.
John: A bunch of this stuff is intertwined with our tooling.
Toerless Eckert: Would there be a way to say ASCII rendering of this if the unicode is not supported.
John: If we are going to do stuff for accessibility, should do it at a higher level to cover more cases.
Pete: Don't get into tooling
Paul: Can we not do the mic line on this
Jay: we are conflating a lot of bits here, there was an over-restrictive implementation of RFC 7997 that was taken to the RSAB who agreed and this is being corrected. That is quite different from Maths.
PHB: expanding Unicode is very useful; I use a lot of math, and not having Unicode forces you to change the nomenclature, which is very confusing
John: this gets you symbols; for full math you need SVG
PHB: The letters are what really matter.
Eliot: Could you put a pin in these two slides for later discussion?
Carsten Bormann: We do use contacts in the contributors section.
Presenters: Paul Hoffman
Draft: https://datatracker.ietf.org/doc/html/draft-rswg-rfc7990-updates/
Objective: Current status, Issue Review
Discussion: deferred
Presenters: Martin Thomson
Draft: https://datatracker.ietf.org/doc/html/draft-thomson-rswg-syntax-change/
Objective: Discuss potential work item
Jay Daley: Key thing is that an RFC must preserve the original intent when it was first written. Can enhance the XML, change it because we are not changing the original intent.
We do publish documents we know are wrong. Errata are published. Should be able to incorporate errata.
The immutability we have should be examined.
Pete Resnick: Desire was to preserve the semantics in the XML, not the intent of the WG.
Martin: Intention was to keep the subset. Wanted to see if we could make minor changes
Paul Hoffman: Like what Jay is saying, not how he is saying. Issue is canonical, is getting in our way. Should think in terms of what people are trying to achieve rather than define canonical
PHB: Versions! Adding in errata can produce RFC9999;3. Update to XML to meet new schema
Pete Resnick: The pin is getting bigger.
Carsten: Point of IETF is to run various consensus. Point of the document is not intent, only consensus. You will never get the band back together tha had the consensus. Consensus built around rendering, not the source
Martin: Disagree somewhat, we are often generating consensus around markdown. Often see the RFC Editor fix line wrapping in particular ways. May have alternative ways to do that.
Eliot Lear: Why do we need canonical in the first place? Point of the series is interoperability. Talk about canonical from a readers perspective is we don't want HTML to say one thing and PDF another. We are not the UN, we don't need to have a comittee to decide which language is correct or has precedence. Key thing really is the rendered format.
Alexis Rossi: Would be useful to look at canonical and archival separately. Don't need to make it a gating factor here.
Robert Sparks: Publishing the rendered formats, if we had to change the fonts, say. Being able to re-render backwards could be a use case.
On the eratta, inline rendering experiment on the eratta, the erattas content is not always in a form that can simply make a change that fixes the semantics. Particularly if it is changing ABNF. Not written in a before/after format.
Pete Resnick: If we start looking at errata we end up looking at the whole way they are created.
Jay Daley: Canonical is problematic, nobody really looks at the XML source. The point is that it is the source from which the presentations are generated. Sometimes there is a mismatch.
On consensus: We actually have two formats, an authoring formt and a publication format. We are probably not ever getting past the boundary of where the consensus lies. Auth48 is a process for dealing with that boundary.
Paul Hoffman: Likes what Jay said about canonical. Its not the eratta, its the updates. RFCx changes a SHOULD to a MUST in RFCy...
Pete Resnick: So you would like to set asside the question of errata and updates for now.
Martin: I am less concerned about the use of the term canonical, still a useful concept to have.
Eliot Lear: Probably worth scoping a bit in terms of what is allowed in terms of semantics. Can probably be a bit more liberal in terms of what we can allow.
Alexis Rossi: Consider hornets nest kicked. Would want to think long and hard about
PHB: all the drafts I wrote with math squashed out, now I could rewrite with math (I don't need integrals). groups should not waste time working out how to thread it through the publication process. [anecdote about mathml.]
Pete R: for that kind of issue, which is about adding functionality, should come to the list as a separate proposal. it is independent of the deeper question being asked here
Martin: Alexis, talking about the metadata outside the document, don't want to touch that. There is also metadata in the document as well. Would like to fix some of that across the series.
Presenters: Pete Resnick
Would like to see if we can come to agreement. Martin, can your draft be made into something solid in the next few weeks?
Should see if we can get to consensus on what we can do quick.
Paul: Suggest we make Martin's doc a working group document now so it becomes a target people take notice of
Robert Sparks: Can we do an interim?
Pete: Would like to entertain idea.
Alexis: Won't move forward with Paul's draft and will with Martin's.
Paul: So we start with Martin's then move to mine. Because there is still more to fix.
Pete: Online interim was my thinking.
Pete: This is nobody's day job so progress is slow to get moving. Will have to push and push to get things to happen. Online interims are likely to be necessary.
Doing it in person puts us against technical things. Perhaps we could find a non conflict time during in person meets. Do we have enough people in the room. May just be stuck on that.
Paul: This is a WG where there is a lot more discussion on the list.
Eliot: Suggest you plan two interims so can handle time zones.
PHB: squash the idea meeting off-schedule. some of the people who are most vocal are not attending IETF meetings
Pete: I'm hearing the push-back
Alexis Rossi: Conversation in RSAB, message was when coming up with new ideas, would be good to bring them to RSAB first so can say if there is already policy.
Eliot Lear: RSAB vested much power in the RPC, for example rules on using UNICODE, If the RSWG wants to do stuff, their output is via RFCs. If there is any concern about RPC going to far can take it to RSAB. [Note taker is completely confused at this point.]
Robert Sparks: Return to geting things done topic. Would be useful to have a roadmap with milestones and deadlines.
Jay Daley: One of the things that was considered was setting up a repo for language changes, this is probably worth revisiting.
1125-1130: AOB
Robert Sparks: Will ship tool that allows Unicode pretty much anywhere.
Carsten Borman: Happy to lift block but hope people don't write stuff to try patience of RFC tooling.
Paul Hoffman: Without disagreeing with idea unicode should be anywhere, might not be consistent with policy. Tooling is not necessarily authoratative, RFC editor uses judgement on e.g. spelling. Tool is not there to warn on things.
Pete: Tools are there to be helpful when relevant.
Jay: Not sure how it got so complex. Tool allows unicode anywhere but the style guide says don't...
PHB: Formal Methods group can write recomendations on how to process their stuff.
Carsten: RFC 7990 was really just written for IETF. IRTF is not concerned with interop.
Eliot: Don't see problem with lifting restrictions.