On 2018-01-19 09:16 +0100 Lukas Jirkovsky via aur-general wrote:

My common sense tells me that activity that helps Arch Linux to prosper should be considered – be it packaging, triaging AUR requests etc.

From that point of view, it makes sense to not count voting as TU activity, thereby blocking the potential removal.

Just my 2 cents.

Lukas

I oppose this patch and will vote against for the reasons given in my previous reply on this list, which I repost here:

The intent of the first sectionm before the "OR", is to measure any sort of activity. Updating a package, voting or posting a comment shows that the TU is still logging in to the AUR and thus active in some sense. The point of the first section was to provide a way to remove TUs who had simply disappeared. This is as it should be. There is no mandated TU quota for package actions.

The intent of the second section, after the "OR", is to ensure that TUs who repeatedly disregard votes and possibly prevent quorum from being established can be removed.

The special removal procedure was only added to handle special cases where the TU has clearly abandoned Arch altogether (no detectable activity in the last two months) or the TU has simply ignored votes and thus jeopardized quorum, again over a period of two months or more. The normal procedure should be used to remove a TU who has not impeded other TUs in their mission and who has been active within the last two months, which gives them time to offer an explanation or a resignation. Regards, Xyne

On 2018-01-21 10:04 +0100 Lukas Fleischer via aur-general wrote:

So you suggest to remove the first part of the condition (before the "OR") altogether?

I made no such suggestion. With the current bylaws, any 2 TUs can start a regular removal process for any reason. This suffices to remove TUs who aren't "doing enough" according to the majority of other TUs. The only difference from the special removal process is a few more days and a higher quorum. The current bylaws are explicit and fine as they are. There is no need to change them. A TU who still shows signs of recent online activity should be given the full discussion period of a regular removal process to offer an explanation or resignation. After all, all TUs have contributed to this community and were recruited due to qualities that we recognized in them at the time of their application. A few extra days to hear their them out before kicking them out is a costless courtesy. Unless they are preventing quorum from being established, there is absolutely no harm nor pressing need to speed up the process. Just to be clear, I support the proposed removals given the cited inactivity, and I agree that doing nothing other than vote for over a year is not the intended mission of a TU. The silence from both during the discussion period is also strange given that they still log in to vote (vote timers?). And again, the intention of the current bylaws was not to disregard voting as an activity. The first part was to speed up the process for TUs who show no visible online activity (i.e. TUs who have abandoned Arch completely, for whatever reason), and the second part was to remove those who prevent the establishment of quorum. A TU who votes is clearly still logging in to his or her account and not preventing quorum. So yeah, a TU who does nothing but vote for over a year should be removed, but by a regular removal process, which the bylaws already handle. Btw, we discussed this together over 4 years ago and we were in agreement then about both the intention and the formulation. Here's a message in which you clearly agreed with this intention: https://lists.archlinux.org/pipermail/aur-general/2013-August/024783.html You later approved my patch. Regards, Xyne

On 01/21/2018 04:19 PM, Lukas Fleischer via aur-general wrote:

On Sun, 21 Jan 2018 at 21:40:43, Xyne wrote:

...

On 2018-01-21 10:04 +0100 Lukas Fleischer via aur-general wrote:

...

So you suggest to remove the first part of the condition (before the "OR") altogether?

I made no such suggestion.

By your logic, there is no situation where the first part of the condition applies but the second part (after the "OR") does not. So we could as well remove the part before the "OR" -- and, if that were the case, we would have removed the statement before the amendment of the bylaws.

If what I am saying is not true, there might be a misunderstanding: in this case, please describe a situation where only the part before the "OR" applies but the part after the "OR" does not.

Well, strictly speaking in the event that there have been no prospective votes in more than two months, but the TU in question did vote for the votes from, say, three and four months ago but then dropped off the face of the community for the last two months, they would qualify for special removal under the part before the OR, but not the part after the OR. I think that having to contrive that circumstance indicates that it is, in fact, contrived and going against the intent.

In fact, even if you really want to be pedantic and take everything in the bylaws literally, the first condition before the "OR" applies if a Trusted User does nothing but vote: in today's terminology, the "AUR" is the "Arch User Repository" (the collection of packages uploaded and maintained by users) and the web interface is called "aurweb", see [1]. So, while voting is a TU action performed in aurweb, it is clearly not "any action that required TU privileges on the AUR" because the AUR (package collection) is not touched.

Indeed! I tried to make this evident in my initial proposal to this thread. -- Eli Schwartz Bug Wrangler and Trusted User

On 01/23/2018 02:16 PM, Balló György via aur-general wrote:

...

Does anyone have any last-minute proposals to modify the wording for grammar etc. in the event that this is accepted?

Sounds good for me. But how can we check if a TU modify a user account or do anything other than resolving package requests (which is tracked on aur-requests mailing list)? Currently these actions are untracked on aurweb, so we don't really know the last action. I think we need a 'last privileged action' field or 'privileged action log' or something similar to be implemented in aurweb, so any TU could easily check if the condition met or not.

We probably do need some way to track non-requests package deletions, use of AUR_PRIVILEGED in the git backend, etc. But I will probably defer to Lukas/Johannes for that, it sounds like it would involve php. It's not like we were keeping track of this as we should up to now, so that wouldn't really change. -- Eli Schwartz Bug Wrangler and Trusted User

On 01/24/2018 11:18 AM, Eli Schwartz wrote:

On 01/23/2018 12:54 PM, Eli Schwartz wrote:

...

On 01/18/2018 06:18 PM, Eli Schwartz wrote:

...

Not everything that is available only to an aurweb account of the Trusted User type, qualifies as a TU "privilege"

Signed-off-by: Eli Schwartz ---

Handy link to context and surrounding discussion:

https://lists.archlinux.org/pipermail/aur-general/2018-January/033789.html

The current wording of the bylaws indicates that there are two ways for a TU to qualify for special removal due to inactivity:

1) Do not participate in voting, thereby potentially blockading a quorum.

2) Do not participate in general TU'ish activities like maintaining [community], administrating the AUR and the packagers and users therein, being representative of TUs in general on this mailing list by being awesome and stuff, i.e. posting (hopefully useful information that helps AUR users), and... um... voting?

Point #2 calls out "performed any action that required TU privileges on the AUR", but does the tu voting interface on aurweb count as that or not? Moreover, do we *want* it to count? It seems to be somewhat defeating the purpose of the process, i.e. as long as a TU doesn't actually block quorum during a vote, they can remain while not actually performing any of the inherent functions of a TU.

Now, I would argue that under a common sense interpretation the original intent of the bylaws was almost certainly that voting does not count as a "TU privilege", since a TU is someone who has the "privilege" to administrate AUR packages and users in order to keep good order, and select good packages to upload to [community].

But bylaws exist in order to prevent people from having different interpretations of common sense. So this should be clarified no matter what.

Thus far, we've (I think) only seen people argue that:

1) this is what the bylaws really mean, let us clarify it for the sake of less confusion some other day,

2) The bylaws do not mean this and should not do this.

Can I assume that means there is no one who feels this *should* be true, but currently *isn't*?

...

Does anyone have any last-minute proposals to modify the wording for grammar etc. in the event that this is accepted?

The discussion period is over, time to vote!

https://aur.archlinux.org/tu/?id=102

The results are in! Yes No Abstain Total Voted Participation 26 9 4 39 Yes 81.25% Seems like overall people thought this was a reasonable interpretation. Rather than holding *another* round of votes to decide whether the previous voting results (for #100 and #101) should be upheld two weeks later, I think it is safe to declare that the previous proposals were indeed valid. -- Eli Schwartz Bug Wrangler and Trusted User