Subject: Draft minutes TAB meeting 18 November 2016

=============================================================

Minutes TAB call (November 18, 2016)

=============================================================

Info

Time: 2pmET

Dial:

Host confcall: OASIS

US Toll Free: +1 641 715-3822

Chat room: http://webconf.soaphub.org/conf/room/tab

Open action items

AI: Patrick - merge proposed Conformance Clause edits with current working draft & send to Ashok

AI: Jacques - make an editorial pass on nimble review process to work in some of the results of our discussions.

AI: Chet - draft a cover memo to the Process Cmte to accompany the draft nimble review process document.

Agenda

1) Roll call

2) Approval of agenda

3) Approval of minutes November 4th:
https://lists.oasis-open.org/archives/tab/201611/msg00009.html

4) Status of public reviews

5) Status of open action items

6) Update from Robin Cover on use of GitHub for TCs

7) Review final changes to Conformance Clause document provided by Patrick

8) Discuss plan of work for the next couple of months

9) AOB

Minutes

1) Roll call

Attending: Ashok, Chet, Kevin, Jacques, Patrick, Robin.

2) Approval of agenda

No discussion of agenda. No objections to unanimous approval. Agenda as amended approved.

Note though that during call, I noticed that one item was missing: discussing Jacques' rolling review proposal. Group approved inserting it as new item 8.

3) Approval of minutes of November 4
https://lists.oasis-open.org/archives/tab/201611/msg00009.html

No discussion of minutes. No objections to unanimous approval. Agenda as amended approved.

4) Status of public reviews

- WS-Biometric Devices Version 1.0 - closes today. Patrick / Jacques will provide feedback.

- XLIFF Version 2.1 - closes 11/25. Agree not much value in a review as this is an update to a long standing spec.

- 6 COEL (Classification of Everyday Living) Committee Specifications - closes 12/9. Brand new, agree that review would be useful here.

5) Status of open action items

- AI: Patrick - merge proposed Conformance Clause edits with current working draft & send to Ashok

- AI: AI to Jacques and Chet to draft nimble review proposal for review at next meeting.

6) Update from Robin Cover on use of GitHub for TCs

Robin thanks the TAB for input that helped with the initial release of experimental Githubs for TCs, particularly Kevin and Patrick for concrete tests and suggestions.

Notes that next year he may as TAB to take a look at documentation or tooling or taxonomies for labels use of the facility takes off.

7) Review final changes to Conformance Clause document provided by Patrick

Tabled to next meeting.

8) Review of Jacques' rolling Public Review proposal

Latest version is at https://www.oasis-open.org/apps/org/workgroup/tab/download.php/59379/latest.

Consensus in the group that we should stick with the current, if not longer, review periods (30 days first / 15 days subsequent).

Discussion of the review turnaround time led to discussion of whether time is the problem.
Chet noted that the only time people express an interest in timelines is during the sales cycle, before an effort comes to OASIS. Once here, people never complain about the length of public reviews. Consensus that the real value proposition being offered here is *not* shorter public reviews but rather a mechanism to enable a TC to address all issues in one dynamic, extensible PR that would remove the need for a 2nd.

Consensus that only changes made to draft would be to address comments. New material could not be inserted.

Noted technical issues that would need to be solved. Number scheme (e.g. csprd 1.1, 1.2, etc)? Where updates would be stored (Kavi with a pointer doc in docs.oasis-open.org; in GitHub using Markdown)? Mechanisms to ensure that someone who provided comments early would know if changes have been made that affect their input in a subsequent draft. Requirement that updates would have to be red-lined diff files?

Jacques notes that one area of controversy may be that editors are allowed to decide when to advance a proposed new draft into the review process. No TC vote would be required until the end when the final document is ready.

Editors would have responsibility for communicating with comment providers to let them know resolution was reached and confirm their satisfaction or lack thereof. Ashok suggests a final call from the editors to commenters to see if they are satisfied with the final draft. Jacques suggests a criteria for ending a PR is that it doesn't close until (a) all comments have resolved and (b) all commenters have been notified and have had the chance to register their happy or unhappy status.

Agreed to following next step is to provide to the Process Committee with clear indication of its draft nature so that they can have a preliminary look.

AI to Jacques: Make an editorial pass to the proposal to work in some of the results of our discussions.

AI to Chet: Draft a cover memo to the Process Cmte to accompany the draft.

9) AOB

No other business was raised. Meeting was adjourned.

Next meeting Friday, 02 December 2016, 2PM Eastern time.

Minutes respectfully submitted by Chet Ensign on 21 November 2016.

Chat log

Chet: Today's agenda:

1) Roll call

2) Approval of agenda

3) Approval of minutes November 4th:
https://lists.oasis-open.org/archives/tab/201611/msg00009.html

4) Status of public reviews

- WS-Biometric Devices Version 1.0 - https://www.oasis-open.org/news/announcements/two-public-reviews-from-biometric-services-bioserv-tc-end-nov-18th

- XLIFF Version 2.1 - https://www.oasis-open.org/news/announcements/30-day-public-review-for-xliff-v2-1-ends-nov-25th

- 6 COEL (Classification of Everyday Living) Committee Specifications

5) Status of open action items

- AI: Patrick - merge proposed Conformance Clause edits with current working draft & send to Ashok

- AI: AI to Jacques and Chet to draft nimble review proposal for review at next meeting.

6) Update from Robin Cover on use of GitHub for TCs

7) Review final changes to Conformance Clause document provided by Patrick

9) AOB

anonymous morphed into Kevin Mangold (NIST)

Kevin Mangold (NIST): dialing in in a minute... i need to update my speed dial with the new number
Kevin Mangold (NIST): (for next time) i keep dialing into the old one

anonymous morphed into Patrick

Chet: Attending: Jacques, Ashok, Patrick, Kevin, Robin and Chet
Chet: Agenda: no disc, no obj. Agenda approved
Chet: Minutes: no disc, no obj. Minutes approved.
Chet: Status of public reveiws
Chet: Kevin would appreciate comments on the WS-BD spec.
Chet: 5. Status of AIs
Chet: AI: Patrick - merge proposed Conformance Clause edits with current working draft & send to Ashok
Chet: Still open
Chet: AI to Jacques and Chet to draft nimble review proposal for review at next meeting.
Chet: Closed.
Chet: 6) Robin recap on Github for TCs
Chet: Thanks for the feedback on options / alternatives for alerting Github users of the terms of submission of content. Esp Kevin and Patrick.
Chet: Jacques: any information on what the TCs intend to use the projects for?
Chet: R: this is about Github as an atlernative to SVN - isn't intended to be used as an open source project
Chet: R: quesiton there is whether they'll want the license from the GitHub for TCs if they use it for code or prefer to go to open source
Chet: R: we're a little behind W3C and IETF. I'm not expecting a lot of uptake except for TCs that want to develop code on the side
Chet: J: do we have info on in W3C they want a more robust set of versioning toos?
Chet: R: the social media aspect and the tight integration between authoring process and issues. Plus they have a whole tool chain that makes it somewhat easier for editors to manage the process
Chet: R: maybe next year TAB can take a look at documentation or tooling or taxonomies for labels if this takes off
Chet: 7) Review final changes to Conformance Clause document provided by Patrick
Chet: Table for next meeting
Chet: New agenda item: Discuss nimble review proposal
Chet: https://www.oasis-open.org/apps/org/workgroup/tab/download.php/59379/latest
Chet: Jacques gives a walk through of the document
Chet: Start with the 2 scenarios at the end of the doc
Chet: J: suggest reduction from 30 day minimum review to 3 weeks
Chet: J: one controversial change is that updated docs during review would not require votes to approve
Chet: J: would depend on the comment logs and drafts for managing the flow of the work
Chet: J: after 7 days w/ no comment, the review ends
Chet: J: TC can then use that final draft as their work for approval
Chet: as a new CD
Chet: Then there would be a vote
Chet: All based on consensus management through the comments
Chet: Kevin: changing from 30 days to 3 weeks would not be helpful. He's had expressions of concern at NIST that even 30 days is too short. It can take ISO several months just to get docs out. Think it should stay 30 days or even make it longer.
Chet: J: don't have a strong opinion on that parameter of the process
Chet: K: there are people at the security group at NIST who are not happy that the Trust-El TC work is out for only 15 days
Chet: @Kevin - I'll contact you about that separately and help get that back to the TC
Chet: Ashok: +1 on the duration question.
Chet: A: 2nd is that the spec will change out from under you as the review goes forward. You put it down after and then come back to it a week later and oh my look there is another update and another and you have to follow up.
Chet: J: true.
Chet: J: should be rules on how often the update could be made. E.g. must be at least a week in between updates.
Chet: J: e.g. comment comes in week 1 and the spec is updated. then week 3 another comment comes in that contradicts the first and the spec is changed back - but the first reviewer doesn't know that
Chet: J: editors though should be in the best position to follow the changes and communicate with reviewers
Chet: J: the editors have to reach out to the comment issuer and keep them informed and work for consensus
Chet: J: so at the end of the public review, there has to be notification from the editors to the commenters that identifies what was finally done with that comments.
Chet: J: maybe there is a last call for comments or something like that
Chet: J: maybe protest from a commenter triggers a second pr
Chet: Chet: +1 time frame
Chet: Chet: also would need to limit updates to responses to comments - no putting in brand new material
Chet: C: also technincal issues to work out - what would the numbering convention be? where would it physically reside? what would be the notification mechanism when the TC produces an update
Chet: J: agree. there would need to be some mechanism for highlighting the revisions / changes and the comments they respond to. TC could pick and choose what comments would be addressed in real time versus later on
Chet: Patrick: back to the inadequacy of time period of review. There has always been a time to market push at OASIS but is it a real need?

Patrick: No changes may be made to the public review draft during a review.

Chet: P: so the WD could change anytime during the PR - because the published draft isn't changing.
Chet: P: the real issue is has there been an adequate public review? we may be straining at something that's not an issue for the TCs themselves. Rather than making it a time issue, why not make it an issue around getting to quality. needs to be a greater focus on the quality of what's produced.
Chet: P: time to approval is not the real issue.
Chet: C: +1 - I think you have a point. People want to know about the times when we're in the OASIS sales cycle - but it never comes up once they are working.
Chet: J: The real issue is not the time of the first - but rather that TCs do not want to have to go to a second public review. I've seen TCs do whatever they can to avoid having to jump through that second PR.
Chet: J: speaking of quality - TCs will push back quality issues to avoid the 2nd review
Chet: Kevin: about where does the interim document live - maybe the document actually lives on GitHub, maybe as Markdown.
Chet: K: that's how some of our NIST docs are being put out here - on Github as Markdown - 80063 from the security group is done that way
Chet: P: so maybe the focus should be on the administrative burden to invoking public review. Maybe if we made the mechanisms for starting a public review more lightweight, maybe we could sneak in a longer PR with better comment handling.
Chet: J: indeed - the administrative friction is a big deal. The TC should loop back to the commenters to find out if they happy is important. So having a single PR that is extensible to get the whole process done.
Chet: J: so, per Ashok's comment, there has to be a call from the editors to the commenters at the end to see if they are satisfied by the final draft. Some notification mechanism, some loop back, to make the commenters aware of the final PR draft.
Chet: J: the friction could be improved indeed between PR 1 and PR 2
Chet: J: maybe the criteria for ending a PR is that it doesn't stop until (a) all comments have resolved and (b) all commenters have been notified and have had the chance to register their happy or unhappy status

Kevin Mangold (NIST): NIST 800-63 on GitHub: https://github.com/usnistgov/800-63-3

Chet: J: a fusing of PR1 and PR2
Chet: Thanks Kevin
Chet: Chet: so what's next?
Chet: K: as long as the Process Cmte knows this is a draft it would be good to get their initial feedback
Chet: J: +1 - these comments will take some time to consider
Chet: J: but the gist can still be communicated with the caveat that we are still working through some things.
Chet: J: but I would still like to make a few modifications based on what we have discussed
Chet: A / P: we can do that. There'll be a lot of bakc and forth

jacques (Fujitsu): New process = PR1 fused with PR2 (i.e. without the interim administrative overhead), with a clause to terminate the whole thing early on if no remaining comemnts/issues

Sent transcript to: chet.ensign@oasis-open.org

20161118 (last edited 2016-11-21 21:18:36 by chet.ensign)