Re: [ripe-list] PDP Appeals Process
On 12 Apr 2021, at 22:14, Peter Koch wrote: … … This could be continued, And it should in the context of the PDP review! but I'll stop here to focus on Daniel's question. That question remains: “Is it worth the effort to take a step back and describe how we deal with conflicts and their resolution within the RIPE community or is tweaking the PDP appeals procedure enough for now?” Any answers? Suggestions? Daniel
Re: [ripe-list] PDP Appeals Process
On Sun, Apr 11, 2021 at 10:31:17AM +0200, Daniel Karrenberg wrote: > Again: Should we go beyond tweaking and fundamentally review the PDP appeals > procedure? the problem that lead us here seems to be much deeper than the appeals part at the far end of the PDP. One source of confusion is that there are natural limits to the jurisdiction of the PDP or RIPE policies in general: that's the numbers management of the RIR and, to some extent, the community services performed by the NCC for the RIPE community. That's also how far enforcement can reach. The 'controversial' proposals all tried one way or another to overstep this capacity to "enable" the NCC to take away resources for "violation" of rules that aren't in the natural remit of an RIR. One feature of the PDP that complicates things is that the author/proposer keeps the pen (other than, at least in theory, within the IETF). This may be going back to the times where policy would have originated from 'common sense' within the community and the author(s) were the poor stuckees doing the legwork. It doesn't fit well in cases mentioned above. This could be continued, but I'll stop here to focus on Daniel's question. -Peter
Re: [ripe-list] PDP Appeals Process
On 4/10/21 11:13 AM, Daniel Karrenberg wrote: I have some implementation ideas already, similar but not identical to the RIPE NCC arbitration procedure. However before I get to those I would like to have some feedback on the general idea. Wearing my arbiter hat, I agree that some of the challenges you describe here are not dissimilar from those encountered during the RIPE NCC Dispute Arbitration process. While there's by no means a 100% solution to them all there either (e.g. the risk of bad-faith use of the process to clog things up), it's indeed possible some elements of the arbitration process might be helpful to address points you outline. If that is, reform of the PDP appeals process is the path chosen, which I don't have enough familiarity with to have a view on. Keith - Appeals require a large amount of community resources. - The process involves too many people. - The process involves people who have not consciously signed up for it, e.g. all WG chairs. - The process involves significant number of people who feel they have to recuse themselves. - Documentation and Openness of the process leave to be desired. Trying to apply incremental improvements to the existing procedure will not solve these significant shortcomings. Therefore I suggest to make more fundamental changes that do address these shortcomings. Here are three generic suggestions: 1) There should be a higher threshold to make an appeal because appeals are costly to the community. 2) Appeals should be handled by a small number of people who commit to handling it properly within a defined time line because someone has to take responsibility. 3) Appeals should be fully and transparently documented from the first submission until the conclusion, because this is the RIPE standard.
Re: [ripe-list] PDP Appeals Process
Daniel, My own feeling is that the appeals process worked basically well enough the first and only time it has been used, and that we would not be able to get any significant gain by a redesign at this point. Since we've only had 1 appeal since the process was actually documented 6+ years ago, maybe we should wait to see if the appeals process is actually broken before scrapping it? More thoughts inline and at the end... On 11/04/2021 10.31, Daniel Karrenberg wrote: Randy, colleagues, we are already past the station where we wonder whether to review the appeals process. Our chairs collective is already actively thinking about significant tweaks. This is what prompted my reaction. I wouldn't call the proposed changes significant: 1. Clarifying who should recuse themselves. (This was already implicit, but to quote the Zen of Python: "explicit is better than implicit".) 2. Setting deadlines. 3. Clarifying the role of informal discussion before policy proposals. In fact, I don't see any much work at all there. To make my point clearer let’s look at this from the perspective of cost to the community: The chairs are proposing costly tweaks to the existing procedure, such as writing and agreeing on a playbook and giving courses to WG chairs that may never use the PDP. I ask whether we should fundamentally review the procedure instead. That has a cost too. I expect this cost to be less or equal to the cost of the proposed tweaks. I also expect that we can come up with a good procedure that costs significantly less to *run* each time than a tweaked procedure. My message gives the general idea on how I propose to achieve that. This is the question we have to answer first. The engineering comes after that. And, as always, the engineering will include trade-offs: the less costly the execution of the procedure is, the lower the threshold to invoke it can be and vice-versa. Again: Should we go beyond tweaking and fundamentally review the PDP appeals procedure? We now have N=1, since we've had exactly one appeal. I was the one who facilitated the process and wrote the draft of the response after a special meeting of the working group chairs collective (WGCC) to discuss the appeal. We had a relatively small number of chairs who actively participated in the process. I suspect that any handling of an appeal would look pretty much the same, although possibly with different people. I don't think we would save work by picking different people to handle the appeals process; we would just shift it around. As for including people who did not sign up for this task, the role of the WGCC in appeals was set in the PDP in 2014 (ripe-614), so this is hardly new (although I see that when we updated the working group chair job description in 2017 we missed that bit; technically redundant but really it should be there). In the end we basically have the same unpleasant options we usually have for such things: either re-use a body that was not designed for this purpose or create a new one. Both options kind of suck. In principle the appeals procedure could be used as a denial-of-service attack on the RIPE community by bad faith actors. But I think that the RIPE community is highly resistant to such problems and I trust that our RIPE Chair will be able to handle any such situation in a common sense manner. Cheers, -- Shane OpenPGP_0x3732979CF967B306.asc Description: application/pgp-keys OpenPGP_signature Description: OpenPGP digital signature
Re: [ripe-list] PDP Appeals Process
Randy, colleagues, we are already past the station where we wonder whether to review the appeals process. Our chairs collective is already actively thinking about significant tweaks. This is what prompted my reaction. To make my point clearer let’s look at this from the perspective of cost to the community: The chairs are proposing costly tweaks to the existing procedure, such as writing and agreeing on a playbook and giving courses to WG chairs that may never use the PDP. I ask whether we should fundamentally review the procedure instead. That has a cost too. I expect this cost to be less or equal to the cost of the proposed tweaks. I also expect that we can come up with a good procedure that costs significantly less to *run* each time than a tweaked procedure. My message gives the general idea on how I propose to achieve that. This is the question we have to answer first. The engineering comes after that. And, as always, the engineering will include trade-offs: the less costly the execution of the procedure is, the lower the threshold to invoke it can be and vice-versa. Again: Should we go beyond tweaking and fundamentally review the PDP appeals procedure? Daniel On 10 Apr 2021, at 19:31, Randy Bush wrote: Therefore I suggest to make more fundamental changes that do address these shortcomings. Here are three generic suggestions: 1) There should be a higher threshold to make an appeal because appeals are costly to the community. 2) Appeals should be handled by a small number of people who commit to handling it properly within a defined time line because someone has to take responsibility. 3) Appeals should be fully and transparently documented from the first submission until the conclusion, because this is the RIPE standard. how may appeals has ripe had? how many appeals were upheld? how much sturm, drang, and omplaloskepsis are we willing suffer to tune it? imiho, your point one is the toughie. you want to require N signatures? I have some implementation ideas already, similar but not identical to the RIPE NCC arbitration procedure. However before I get to those I would like to have some feedback on the general idea. i fear we have to go through this. if so, i respect and value your start. randy --- ra...@psg.com `gpg --locate-external-keys --auto-key-locate wkd ra...@psg.com` signatures are back, thanks to dmarc header butchery
Re: [ripe-list] PDP Appeals Process
> Therefore I suggest to make more fundamental changes that do address > these shortcomings. Here are three generic suggestions: > > 1) There should be a higher threshold to make an appeal because >appeals are costly to the community. > > 2) Appeals should be handled by a small number of people who commit to >handling it properly within a defined time line because someone has >to take responsibility. > > 3) Appeals should be fully and transparently documented from the first >submission until the conclusion, because this is the RIPE standard. how may appeals has ripe had? how many appeals were upheld? how much sturm, drang, and omplaloskepsis are we willing suffer to tune it? imiho, your point one is the toughie. you want to require N signatures? > I have some implementation ideas already, similar but not identical to > the RIPE NCC arbitration procedure. However before I get to those I > would like to have some feedback on the general idea. i fear we have to go through this. if so, i respect and value your start. randy --- ra...@psg.com `gpg --locate-external-keys --auto-key-locate wkd ra...@psg.com` signatures are back, thanks to dmarc header butchery
[ripe-list] PDP Appeals Process
Mirjam, Niall, WG Chairs, Thank you for sharing your minutes about the Appeals Review and PDP Evolution. As far as PDP appeals are concerned I have the impression that the discussion is at the wrong level. We seem to be trying to tweak the procedure without fully recognising some significant shortcomings: - Appeals require a large amount of community resources. - The process involves too many people. - The process involves people who have not consciously signed up for it, e.g. all WG chairs. - The process involves significant number of people who feel they have to recuse themselves. - Documentation and Openness of the process leave to be desired. Trying to apply incremental improvements to the existing procedure will not solve these significant shortcomings. Therefore I suggest to make more fundamental changes that do address these shortcomings. Here are three generic suggestions: 1) There should be a higher threshold to make an appeal because appeals are costly to the community. 2) Appeals should be handled by a small number of people who commit to handling it properly within a defined time line because someone has to take responsibility. 3) Appeals should be fully and transparently documented from the first submission until the conclusion, because this is the RIPE standard. I have some implementation ideas already, similar but not identical to the RIPE NCC arbitration procedure. However before I get to those I would like to have some feedback on the general idea. Best Daniel Full disclosure: RIPE Participant since RIPE 0 and NCC staff from the beginning.
