Workshop for Working Group Chairs D. Crocker Silicon Graphics dcrocker@sgi.com / +1 415 390 1804 Hello o Introduction - What are we doing here, at this hour? - There is only rough consensus about the process o Agenda I. IETF Structure II. Formal process III. The inner working group IV. Conflict resolution The need for working group chair training o IETF large, diverse - Process increasingly formal - No voting means (very) rough consensus o Difficulty making progress and being fair - Listen to all points of view - Keep working group focus o Chairs often unclear about limitations and authorities Documents o The Internet Standards Process (RFC 1310) o IETF Working Group Guidelines and Procedures (in progress) o ["Evolving the System" in Internet System Handbook, Lynch & Rose, eds.] o ["Making Standards the IETF Way" in ACM StandardsView, Summer 1993; reprinted in ConneXions, August 1993.] I. IETF Structure ISOC Internet Society legal cover IAB Internet Architecture Board Design cohesion, process appeals, IETF liaisons IETF Secretariat Staff support IESG Internet Engineering Steering Group IETF oversight AD Area director Oversight for specific working groups WG chair Manage a working group to a productive end Working group The people who do the work Working group roles o Chair oversees entire process, but: Facilitator Process management, things fair, focussed, on time Judge Evaluation of technical options and driver towards "right" choice Scribe Record-keeper and editor of documents o Working group is jury, providing ideas, review, consensus o Design team is primary advocate as self-selecting group with common vision, providing core effort Formal Process o Formal labels for a specification o Developmental steps o Acceptance criteria Formal labels Internet-Draft: no official standing, fluid working document Proposed Standard: stable spec, no known errors, might have implementation Draft Standard: multiple, interoperable implementations testing all functionality Internet Standard: field experience and clear community acceptance (and use) IETF acceptance criteria Competence: technically sound Constituency: providers & users Coherence: clear writing Consensus: rough but clear II. Developmental steps 0. Birds of a feather (BOF) - "Market research" to determine interest and ability to pursue topic - Optional, one-shot meeting 1. Charter Role: Public announcement & project management plan Scope: What is to be pursued Approach: How will it be pursued Product: What will be delivered Checkpoints: Milestones and dates 2. Document specification - Clarity of purpose - Clarity of writing - Clarity of solution 3. WG consensus - Clearly dominant agreement - Diversity of opinion about solution may be resolved by agreement to make some decision - Agreement about parts may permit eventual agreement about whole 4. Area director approval - Technical review - Process review - Independent review when results of wg in question 5. Submission to IESG - Via secretariat & AD 6. Last Call - Request for final feedback from IETF - Intended to detect major errors in process or content that might have slipped through cracks - Not intended as formal, full review 7. IESG Review (and approval) - May conduct independent review (7.5) IAB conflict resolution - If formal challenge not resolved by IESG 8. RFC publication - RFC Editor has publication criteria III. The inner working group o The lives of a chair o Group style o Group roles o Developmental phases (problem solving 101) o Venues o Debate o Conflict management The lives of a chair o How to keep from being sat on - Agenda & schedule - Adequate debate, but not more than that - Maintain clear focus - Rehash only if constructive and working group desires o Proactive management - Maintain pressure for forward progress - Escalate to IETF management when process stalled Working group style o Free-flowing - Cohesive group - Clear purpose o Tightly-managed - Complex topic - Group diversity - Major differences in philosophy WG management roles Facilitator: ensuring fairness and a thorough airing of views and alternatives Judge: evaluation of choices and movement toward choice Scribe: keeping track of things Design Team: Primary advocates for the core effort, when wg diverse and topic complex; must work to keep wg consensus Working group: Jury and other contributors Problem solving 101 o Problem statement o Solution exploration o Solution adoption o Specification refinement Discussion & decision venues o Email - International participation - Inefficient, but expensive - The real place for consensus - Can be run as "meetings" o Face-to-face meetings - Well-advertised ahead of time - Inherently restricted attendance - Limited time - Need for clear agenda and crisp management Email vs. Meetings (One person's perspective) o WG results must show approval based on entire working group o Meetings have limited attendance o Treat meetings as "strong indicator" prima facie basis for decisions o Ensure verification through email Debate o Can clarify purpose, implications, alternatives o Can tear the group apart o Must be tolerated and even encouraged, until resolution or impasse IV. Conflict Resolution o Preferable to solve within working group 1. Conflict types 2. Timing of objections o Often can't 3. Chain of appeal 1. Conflict Types o Technical Specific detail: minor vs. showstopper Basic philosophies: rarely resolved o Process Unfair practice: usually claim against wg chair Topic missed: oops. (showstopper?) 2. Timing of objections o Technical showstoppers welcome any time o Small details welcome only at time WG covers the subject o Philosophical debate welcome only at time WG making decisions about approach o Unfair practice complaints allowed whenever infraction felt o WG may allow topic to be re-opened if WG feels issue compelling or new alternative intriguing 3. Chain of appeal o WG chair o Area director o IESG (plenary) o IAB If you can keep your head when those around you... o Most IETF members are remarkably well-intentioned o Differences happen - Tempers often flare, but then settle down - Not all differences can be settled - When minority view clearly will not sway working group, respect opinion, but move on o Ask questions o Make it happen!