On 06/12/14 at 13:26 -0800, Steve Langasek wrote: > On Mon, Dec 01, 2014 at 02:37:30PM +0100, Lucas Nussbaum wrote: > > Rationale > > --------- > > First, I think that there is wide agreement that a more regular > > turn-over among TC members would be a good thing. And both Stefano's > > and this proposal aim at addressing this, by ensuring that at least 2 > > members of the TC are replaced every year. > > > However, too much turn-over, with more than 2 replacements at one point > > of time, might have negative effects too. The TC might be temporarily > > weakened by having more young members; replacing more than two members > > at one point will cause less replacements later; it increases the > > difficulty of finding new members. > > > The recent situation, with three TC members resigning, should not be > > treated as exceptional in the context of this resolution. If it were to > > happen again, I don't think that we should add one or two automatic > > expirations to the three resignations. > > > This proposal differs from the original proposal by counting all > > resignations and removals as part of the desirable "2 per year" > > replacement rate, so that the total number of replacements does not > > exceed two if only one or two younger members decide to resign. > > > This version of the proposal could even result in an internal TC > > discussion: "OK, the Project wants two members to be replaced. Are there > > members that feel like resigning now? Or should we just fallback to the > > default of expiring the two most senior members?". I think that such a > > discussion would be a healthy way for each TC member to evaluate its > > status. The orignal proposal could have the detrimental effect of > > pushing inactive/demotivated members to stay on the TC until their > > expiration, to avoid causing additional churn. > > The pathological corner case here appears to be that the longest-serving > member of the TC could evade the term limit indefinitely. A scenario that > assumes good faith on the part of all TC members is: > > - The longest-serving member of the TC spends a minimum amount of time > engaging with TC issues. They vote on all resolutions, but don't spend > much time cross-examining the petitioners, nor do they participate in > resolution drafting. From their perspective, they are doing their duty > on the TC, but other members of the TC have a faster response time to > issues and therefore wind up doing the bulk of the work. > - The other members of the TC all are very passionate about their work on > the committee. (They've all been serving less than 3 years, so they have > a lot of passion for it.) They engage with every issue, spend several > hours each week on trying to make the TC serve the needs of the project > as best they know how. And once or twice each year, there is a big issue > that lands on the TC's desk, with social and technical issues intertwined > and that require a lot of energy to pick apart. Once a year, one of > these issues further devolves into a public flamewar where the ethics of > the TC members themselves are called into question. And as a result, two > members of the TC per year resign. > - With the minimum turnover requirement met, the longest-serving member > continues to serve as long as they are comfortable doing so. > > Did you consider this corner case in your analysis? If you think this > corner case is less important than the risk of high turnover in the TC, > could you elaborate why you think this? Your scenario describes a case where a member of the TC fullfills their voting duties, but does not otherwise really participate in TC work. This can happen, but I don't really see a correlation between this happening, and the seniority of that specific TC member. One could imagine a scenario where a recently-appointed TC member goes semi-MIA very early, and still stay on the TC for 4 years. After all, in Debian teams, people go MIA for various reasons, and this is not correlated with their seniority in those particular teams. I think that the root goal of this GR is to force more turn-over in the TC, which is a very desirable thing. Doing that by removing the most senior members every year is a reasonable default choice. However, the goal of this GR is NOT to provide a mechanism to automatically 'expire' poorly-performing TC members. I am not sure that this is necessary: we already have a mechanism to remove members of the TC (§6.2.5), which has already been used in situations where members of the TC had become inactive. I doubt that we need more than that. Also, if the version of the GR I proposed gets chosen, I hope that the fact that resignations or removals can 'save' other members from expiration will result in yearly discussions where the status and activity level of each member gets reviewed, which could actually help address the general problem of semi-MIA TC members. Lucas
Attachment:
signature.asc
Description: Digital signature