DebconfOrganizationWorkingGroup

Previous discussions

• DebConf10: Debconf and Debian
• DebConf11: Debconf in Debian
• DebConf12: Debconf organization BoF
• DebConf13: DebConf governance

Documents to review

• Goals of DebConf (main points bellow)
  • Enable face-to-face interactions
  • Provide talks and video
  • Provide time to work on Debian
  • Motivate contributors
  • Motivate the local community

• Jobs
  • Description of jobs that are better done locally, others that can be done remotely but need local cooperation and jobs that do not require local cooperation

• Debconf Manual
  • A collection of wiki pages describing how things have been done in the past

DebConf14 Working Group

Picking up the pieces — Where do we stand and what do we want?

There were two 1-hour open meetings and some private conversations to gather views on how DebConf organization could be improved. Participants were invited to describe how things worked in "their" DebConf team, what worked well and what were the problems.

Meeting #1: August 24th 2014

Slides: File:DebConf workshop 1.pdf

OdyX

• Changing shape of decision bodies in the middle: new people becoming de facto part of teams. It was not always clear where decisions are to be made, there are many different bodies.
• DC13 thought the bid decision was for team and place, but then the choice of place/venue was challenged later.
• They published decisions/minutes and assumed that lack of feedback/concerns should/could be taken as approval, so the team moved on. Then the global team reacted in surprise past dc12. The problem was also that the "global team" was busy with dc12 and so didn't have time to reply.
• There was an expectation for the chairs to help make decisions in a timely manner (but that did not happen)
• The separation between local and global teams became clearer towards the end.
• They lacked clear leadership, and sometimes individuals made uncoordinated decisions.

Ana

• Problems didn't start last year. DC13 was the pinnacle of problems that were waiting to blow up.
• Silence was assumed as consensus. The 'silent approval' is a problem.

Patty and Steve:

• Tried hard to be as transparent as possible towards the remote team members. Communicating "upstream" is a challenge, especially under time pressure, and given that the local team possesses a lot more information. Bikeshedding is obviously a problem.
• Delegating responsabilities and powers to LT members worked well.
• Communicate local deadlines clearly. Asking for a decision to be made, with a clear deadline, and a default decision worked well.
• Soft deadlines didn't work well.
• The last two weeks before the conference needs a dedicated person for everyone to contact, and that is a full-time job.
• It was difficult to gather critical information in time: network and video were problems (hard to get info).
• There was less friction with LT decisions; global team aligned with local team.
• Only one chair as a single point of contact was a plus.
• Topics that were going to be controversial were more prominently communicated to global team.
• Patty and Steve were seen as clear leaders.

Judit

• During dc13 planning and after, global team as a supervising instance seemed to disappear. Decision power shifted from global to local teams.
• DebCamp decision was driven without global team involvement, as an example of this.

Tincho

• Tincho & Marga declared themselves leaders and were able to drive decisions (like venue change) and it felt comfortable.
• There was a high level of trust.

Moray

• DC7 bid competition was a first.
• Unclear decision making processes, and leadership.
• People were careful of overstepping their competencies and so many decisions were made collaboratively on IRC.
• Shorter time-frame.
• Conflicts arised about dates, and about the lack of access to the venue at night.
• Who has a "right" to make an input? How much should this input be valued?

Madduck

• DC15 being told to wait after DC14 ends, lack of bandwidth for N+1.
• The DC15 team had a lot of DC-experienced people and therefore were able to get a really early and strong start. However, what it would be like otherwise? He thinks they would have had a lot of difficulties getting guidance or the basis for making influential decisions, such as the venue negotiations. In conclusion, we ought to be prepared to manage two DebConfs at once.
• There was a long, flamey discussion about the DC15 legal entity, but no feedback on the venue terms. This seemed backwards, as the legal entity is really a background necessity that ought not need to be discussed globally, whereas the venue terms are sort of *central* to DebConf.

Meeting #2: August 26th 2014

Slides: File:DebConf workshop 1.pdf

azeem

• Bikedding coming from people not so involved in DebConf.

gunnar

• Dc6 team understood there were procedures to be followed.
• Since we moved so far, there is no sense of recovering details, but what is repetitive in many conferences: given the differences between teams, locations, etc, there will always be mismatches, things are not straight.

faw

• Sponsorship/bursaries team decisions could be greatly automated and more transparent.
• Not enough time for them to work properly.
• There is no proper definition of rules, it seems that everyone in the team makes subjective decisions based on their own preferences.
• <vorlon> bursary policies might need to be discussed in a larger circle, with dc-team and debian.
• Privacy concerns make it even more complicated.

brian

• Debconf organization might be tough, but the final result is amazing. we should not forget that we are doing something fantastic.
• In discussions, learn to state our opinions, and stop there.

tiago

• Video team functioning mode seems to be broken.
• Before the conference start, it was not clear who was in charge of making decisions.
• They are doing a nice job, but we have lost connection with what was done in the past.
• People has quit. I feel there is a less friendly environment.
• <vorlon> dc-video sounds like it has the dc problems in a microcosmos

lucas

• The DPL has no experience in debconf organization, that is why we have chairs.
• First few months of DPL mandate are quite busy, it is hard to respond to debconf time constraints.
• There is no clear description of how the DPL should be involved in the decision making process.

eric

• dc10 team was a strong team, but with no debconf experience. The global team was not availbale. The local team felt alone and did the best they could, but in the last moment people appeared and tried to interfere with previous decisions.
• Problems over handoff in video-team (dc14).
• Question: what do we do when someone quit? when there is a split, there needs to be a mechanism to the global team to take over.
• He has done more than he was able to, and that made him sick. Somehow we need to protect each other.

Odyx

• Mismatch in expectations versus actual manpower.

Comments from emails and other conversations

Martin Krafft

• We need a clear definition of roles and associated powers around DebConf
  • Bid team, DebConfN team (also, N-1 and N+1), DebConf selection committee, advisory board, sponsorship/budget team, press, talks, travel, infrastructure, network, video, accommodation
• With the proper planning, those in charge can make decisions and represent DebConf without having to seek clearance or enter further discussions
• It is reasonable to ask for feedback, but it should still be possible to make a decision at any time
• "Process": timeline-based description of the steps leading up to and following a bid decision

RichiH, member of DC15 local team

• hard to find out what the actual requirements are as documentation was less than ideal (Bid process)
• deadlines in the documentation were rough guidelines at best and they were redefined on the fly.
• found the overall process opaque and the chairs hard to reach

Giacomo Catenazzi

• The bid should be simplified
• We need a better handling of team know-how and clear responsible persons
• Important teams should have 1-4 responsible persons and most of people on that teams should serve for many DebConfs
• Instead of chairs, current and next DC teams should call meetings and push things
• Teams should self-organize and let the chairs to be more a link/control between -team and Debian

Vorlon

• the structure of the current delegation is broken, and many of the problems of the recent past were a direct result of a delegation structure that relied on consensus to get anything done.
• When two of the chairs resigned leaving a single delegated decision-maker for DebConf, this was an *improvement* over the status quo

Norman Garcia

• no clear separation of "global" and "local" team in DC12, and he advotes for this model
• they larged used manuals and past reports
• we need to define clear roles, powers and procedures
• it is a must for DebConfN+1 team join DebConfN team and learn how things work
• chairs or "elders" team is needed, not necessarily from 3 previous conferences

Ganneff

• The delegation is broken, the best would be without any such delegation, but i know that wont be possible.
• A load of repeated discussions, distrust and bad feelings.
• Give people tasks with boundaries, with freedom to organize internally, but reporting back. Document their responsibilities, and trust them.
• Keep a balance with the amount of details in which structures are defined. It makes it harder to people to join. Don't make it too "german-style".
• If remote people are not cared about, it kills the spirit of DebConf.

H01ger

• There are things where less consensus is better.
• Take decisions globally, except when needed.
• Recover trust.
• Motivation is hard to maintain. Sometimes people just lose it.
• After dc13, many people in the global team left, the lack of documentation became evident.
• The manual is outdated or incomplete, most people ignore it.
• Let the local team decide on small-medium spending.
• Preparation for DebConf takes too long. DebConf is a year-long job.
• Make a no-work-on-DCn+1 window around DCn, after the final report is finished, plus one month.

Sledge

• Involved strongly in dc7, fading then through the years.
• Involvement was stressful, but fun.
• Issue: things are not followed up.
• Issue: some people don't finish their tasks in time, and other people stand in.
• Idea: final report for DCn is done by the DCn+1 people.
• Problems: communication, bikeshedding.
• Let the local team make mistakes, just not expensive ones.
• Document how things are done, and why they are done that way.

Compiling days

There was a 2-day break to process the received comments, complaints and suggestions.

Summary of reported issues:

• Chairs were hard to reach
• Decision making process: consensus x time pressure
• Team structure: too loose? how tidy do we want it to be?
• Bid process: less than ideal documentation for requirements and deadlines, local team work alone for one year
• How to efficiently handle 2 conferences at the same time?

Tying up the ends — where do we go from here?

Meeting #3: August 28th 2014

Slides: File:DebConf workshop 3.pdf

Feedback about subteams:

gunnar: having more structure is not a necessity, very few things need to be defined. For example, how money is spent? we dont want to bother DPL all the time. We need some informality, some decisions are up to the teams.

madduck: as norman said, this is how we function already.

sven: each localteam must decide how to proceed. It always comes to some structure, but getting there can be a painfull process.

gunnnar: this varies with cultural and personall preferences.

sven: in my experience, they were very similar to each other from year to year, maybe this could be used as a starting point.

kathey: maybe what we need are guidelines, starting points, not strict rules.

madduck: experience must be kept.

tassia: feedback from sledge: "a too rigid structure may demotivate some people".

sven: I agree, just a few people wouldn't like to have those few guidelines. I got into VideoTeam because there was a need.

didier: "this will be handled by these people this by this, and this is in the middle groud (in a bootstrap wiki). Do not encode too much. Manual too long. A bootstrap wiki or meeting in person (fly the chairs over to talk to the LT)

madduck: "mentors, not necessaryly chairs". Local functioning mode is locally decided, as long as they respect some rules.

sven: it should still be 'guidelines'.

gunnar: in Bosnia there was a vist from the chairs and they have achieved important things, it served as motivation for locals. It is not always needed, but we should always consider that.

didier: in Switzerland it was too late.

Feedback about permanent teams, timeline, documentation

madduck: excelent suggestion, we already do that in the project. we shouldn't document that too extensivelly, we might allow experiments. stay openminded.

moray: even if you made the decisions 5 years ago, you might not remember why you made those decisions.

tiago: agree with cate, that's how dc works, what is missing is delegation. holger couldn't manage, and there is no delegation to do his work, nobody was feeling the authority to made decisions. noone filled the gap in the right time.

kathey: you probably have the people that could have made the decision (a problem of legitimacy).

madduck: in addition of having a backup leader, there should also have a supervisory position.

gunnar: I know that many members of that team already knew the work, the problem was: who coordinates.

didier: even a random person is better than 10 people.

eric: communication problem. who is the point of contact. who is in charge?

Feedback about team structure and leadership

kathey: I like the ideas, but if it becomes too concrete, you might loose what makes this community special. there might be a lot of dealing with ego issues.

gunnar: I dont like patty's comments ("full-time job for the last 2 weeks"), this is what is expected. if it was only 2 weeks, it was highly expected.

didier: decisions chalenged, communication overhead.

kathey: we need to encourage people to not be indispensable.

faw: we should force rollover. "you're doing this for 7 years now, give the work to someone else".

sven: reverse shadow.

gunnar: reality check.

didier: chairs delegation time-based?

faw: (not about chairs, in general) having a checklist, year after year, "before you burn out, teach someone".

sven: delegate, please.

moray: in principle I like the idea, but some people are only interested in a specific role.

tincho: this happens with companies.

sven: eventually people will get bored and annoyed by a job. stay in contact, pass knowledge.

madduck: keep the teams small and delegate.

List of suggested subteams

Talks, bursaries, fundraising, registration, visa, video, frontdesk, sysadmin, accomodation, food, network, summit/penta, daytrip (social events), commitee (bid selection), accounting, sponsorship, volunteers coordination.

Meeting #4: August 30th 2014

Slides: File:DebConf workshop 4.pdf

• Review of consensus goals from past meetings
  • Clearer decision-making processes
  • Better-defined subteam membership
  • Better-defined subteam responsibilities
  • Subteam membership that continues from year to year
  • Clearer timeline for the DebConf year
  • Clearer documentation on agreed processes

• Proposed new roles: leaders, shadows, wizards, local liasons (agreed)
• Proposal: related tasks grouped into fewer subteams, avoid overlapping roles (agreed)
  • 22 subteams were listed, we can not have more subteams than people (we aim for +/-5 subteams)
  • Suggestion: there could be a coordination subteam, not the same as (but can include) volunteer coordination
  • There could be a "local facilities team" which would group food, venue and accomodation
  • Mentors for bids? Is it a task for the committee?
• Proposal: all global team members should belong to a subteam (agreed)
• Proposal: recruit team leaders by canvassing past team members (agreed)
  • Possible questions:
    • "Which DebConf teams have you worked in before and what would you like to do in the future in DebConf?"
    • "Who would make good leaders for those teams?"
    • "If you see yourself as a team leader, who would be good alongside you?"
  • Agreed: we need to have 1st and 2nd person in each team always (not 2 leaders)
• Proposal: encourage rotation of team leaders before burn-out (agreed)

• Vorlon has volunteered to draft subteams structure

vorlon:

• can we call it "team lead" instead of "leader"?
• good leaders have good management skills, they should be able to manage people
• useful concept from Ubuntu code of conduct "step down gracefully"
• it is understood that you have to hand off when you are no longer doing the job

moray: even more, it should happen even if you are still doing a good job

• What about having cycles?
• What about chairs: yes, there should also be a rotation, the time scale is not clear

Agreed timeline

• 1st week of October: agreement on list of subteams and responsibilities
• 1st week of December (might go to January 2015, since December is Holidays anyway): agreement on...
  • Team members with appointed leads, shadows, local team liaisons and possibly wizards
  • Default timeline of DebConf year
  • Push documentation of subteam processes

Comments about the role of Chairs

maduck: yes, the chairs are necessary. At this point, we don't really need to re-discuss the existance of chairs

vorlon: the delegates should not be powers of the DebConf team should be in parallel with the orga team. Can we re-discuss the exclusion of local team members being chairs?

patty: the local team also sees points and process that need to be improved

moray: local team needs a rest after the conference

lucas: in the new structure there should be no that division local/global

moray: chairs and local leaders are certainly part of the coordination group

madduck: relistically, there will always be a local team (that meet locally)

lucas: answer to the second question: you are doing great

Comments about the bid process

Composition of the committee: It should include not only debconf people, since DebConf has an impact in many other Debian areas/sub-projects

madduck: a smaller group that are always there

lucas: some appointed members by the DPL

vorlon: delegates in addition to subteam leaders

The bid process

patty: deadline needs to be fixed

madduck: lowering the requirements, earlier involvement of the committee, having mentors

moray: if we do that, we need to have a "flexible", open decision

patty: at the end of march was perfectly fine

vorlon, madduck: march was late

moray: two stage-approval?

pietro: 17th and 16th at the same moment

madduck: we don't do 2 at the same time very well

moray: too much in advance, team cannont commit: kids, moving countries, etc

maduck: against 2 stages: we are going to lower our expectations

faw: beggining of the year

madduck: whatever the deadline is, it should not change

patty: bid deadline should be fixed.

Agreed timeline

• December 1st 2014: bid proposals submission
• January 31st 2015: bid decision