Private filters should not be discussed in detail here; please email an edit filter manager if you have specific concerns or questions about the content of hidden filters.
The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
For those of you that do not know him, 1AmNobody24 has been quite an active patroller of EFFPR spanning a little more than 700 edits in the past few months, and he would be a great asset to the edit filter team in order to review false positives that involve private filters, and to assist with improving and creating private filters. Some of his suggestions include Wikipedia:Edit filter noticeboard/Archive 12#Filter 1112, and Special:Permalink/1211462999#Improving Filter 1045.
Outside of edit filters, he does a great job of reverting obvious vandalism and spam, has decent UAA, AFC, CSD and SPI logs, fixes references (including but not limited to bare URLs, CS1 errors), adds wikilinks, and has signed the confidentiality agreement for nonpublic information per this diff on Meta.
OpposeWeak support: I'm slightly concerned by this, as non-EFH/EFM/sysop should not generally be actioning reports involving private filters, regardless of how obvious the result may be. The key problem is that the person responding doesn't have access to all relevant logs, nor access to the necessary filters to check the report in full. A similar idea applies to this one. While I think they could be responsible with EFH, I'm not a fan of granting it to someone who recently (within 2 months and even 1 month) has shown to be actioning reports as described. EggRoll97(talk) 21:50, 25 March 2024 (UTC)[reply]
@EggRoll97 I agree that most private filter hits should not be actioned by non-EFH/EFM, but those two reports I can easily explain why I responded. The first one one triggered the Rapid disruption private filter and filters 61 and 636 in the same attempt. When looking at the public filters hits, one can see the obvious reason why that attempt is disruptive. The second report is also for an attempt thar hit both a private and a public filter. By looking at the public hit, one can easily see what part got hit for looking like a email. These were both obvious cases of disruptive attempts and even if I don't see the private filters it's obvious that the hits weren't false positives. Nobody (talk) 05:27, 26 March 2024 (UTC)[reply]
Comment:WP:EFH only talks about requesting the right for yourself, nothing about nominating others. This feels like it misses the candidate's own statement on why they want the right (even if it's obvious). That said, this reads like it was made using a template, so is this just undocumented? (Also the confidentiality agreement diff link is broken, as I've mentioned, please fix that) – 2804:F1...01:18F4 (talk) 21:02, 28 March 2024 (UTC)[reply]
There is a tradition of nominating others, even if it isn't written down on the guidelines. About the statement on why they want the right, I don't really know if it is needed in this case but it could be. – PharyngealImplosive7(talk)21:26, 28 March 2024 (UTC)[reply]
I'll also concur that it doesn't matter too much if they nominate themselves, so long as they're available to answer questions from others. Ultimately the test that is applied is whether the candidate can be trusted, and while self-nominations are fully acceptable, some also like the reassurance that comes from a nominator. Also, confidentiality noticeboard diff updated. (I hope you don't mind my fixing that diff, @Codename Noreste:.) EggRoll97(talk) 22:34, 28 March 2024 (UTC)[reply]
In addition to that, I have some of my own comments to say of what I've learned despite my two failed nominations:
Regardless if they're obvious or not, I also agree that reports that only involve private filters should be left to the ones that can view such log entries. I believe I have learned that the hard way.
Despite so many responses, account age probably matters (almost two years or more is recommended).
I'm not going to use scare quotes anymore as somebody mentioned, including if it's in the edit or summary.
Understood, and I will give myself a year at most to address these issues. In addition, I have written and proposed a filter by emailing its conditions to an EFM (1292 to be exact). Codename Noreste 🤔 talk02:25, 29 March 2024 (UTC)[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Hello all. I am presenting myself here to the community today to request that I be granted edit filter manager rights as a non-administrator. I've thought about this for a bit, and it's 0xDeadbeef's response to my request for a bit of advice and his encouragement of boldness here that has pushed me to bite the metaphorical "bullet", so to speak, and write this up. (As a side note, I've hovered over the publish changes button now for about an hour, uncertain if everything is perfect yet.)
Edit filter managers need demonstrated competency with the edit filter to be considered, as well as being trusted by the community to safely utilize the edit filter. As for trust, it's largely a factor that differs by person, though I of course will present that I have been an edit filter helper handling private filters for just over four months now without spilling the beans, and have signed the confidentiality agreement for non-public information (see m:Special:Diff/20180422). For technical competency, I have attached a few links below for both public and private filter changes I have requested. I've attempted to summarize the private filter changes as best I can without compromising private filter integrity.
Further, I have also passed by more than a few false positives reports that had small changes proposed to the filters that just needed an EFM to make them. This is something I would plan to work on a lot if granted the userright. The EFM right would also allow me to use filters filter 1 (public testing) and filter 2 (private testing) which can be more efficient than Special:AbuseFilter/test as it only tests the last 100 edits (though User:Suffusion of Yellow/FilterDebugger works wonders). I plan to extensively test any edit filter changes I implement, and with new edit filters as applicable, enable on log-only until fine-tuning has kept the false positive count to a low and reasonable degree. I am aware of the confidentiality expectations applicable to the private filters, and am aware of the extensive damage that edit filters can cause if recklessly implemented. I thank you for your consideration, and am fully open to and will respond to any questions and queries as applicable. EggRoll97(talk) 00:11, 29 March 2024 (UTC)[reply]
Question: what type of public and/or private filters do you intend to create using your knowledge of regular expressions and edit filter syntax other than filters 1 and 2? Codename Noreste 🤔 talk00:27, 29 March 2024 (UTC)[reply]
@Codename Noreste: I'd likely take inspiration from the requests at Wikipedia:Edit_filter/Requested, though I believe the bulk of my contribution would come through fine-tuning existing filters based on false positives and filter history. We do, after all, already have a massive number of filters, so I'd be more likely to test changes to an existing filter and merge them in when properly vetted, then to create a new filter, if possible. For reference, at the time of writing, we have over 300 enabled filters. (314, to be exact.) To answer your question directly though, I'd probably start by enabling Wikipedia:Edit_filter/Requested#No_rcats? in log-only and monitoring. EggRoll97(talk) 00:35, 29 March 2024 (UTC)[reply]
May sound somewhat off-topic, but it would be useful if you are a moderator of that edit filter mailing list who can accept or deny (and respond to) requests. Codename Noreste 🤔 talk00:52, 29 March 2024 (UTC)[reply]
Support per 0xDeadbeef. I’ve seen them edit and test filters on a test wiki, and I am confident that they will not cause intense disruption to thousands of editors. Codename Noreste 🤔 talk02:31, 29 March 2024 (UTC)[reply]
@PharyngealImplosive7: It's not really "unavailable". It's available, and used on a few projects if I remember correctly, but it's considered a restricted action because if someone does something that merits that, they probably should just be blocked by an admin instead. The filters also don't auto-block because the idea in this project is that all blocks should be made by an admin, similarly to how userrights should be managed by admins in the community's view. It was discussed somewhere if I recall correctly. EggRoll97(talk) 01:19, 30 March 2024 (UTC)[reply]
Correct. There isn't an automated edit filter action that blocks the user from editing. This template refers to when the "Prevent the user from performing the action in question" and "Revoke the user's autoconfirmed status" actions are used within an edit filter. ~Oshwah~(talk)(contribs)03:01, 2 April 2024 (UTC)[reply]
Extending time for EFH discussions
Historically, EFH has been considered a relatively high trust role. I appreciate opinions on this can vary, and so the "need" to grant has been decided by precedence at this board, and evaluated on a case-by-case basis. Since we do not evaluate EFH discussions against a set criteria (like we do TE in Special:Permalink/1215492787, for example), participation is quite important.
Since many editors in the edit-filter community aren't around every day, to maximise participation, I'd like to suggest we extend the time for EFH discussions to the standard 7 days. ProcrastinatingReader (talk) 12:18, 30 March 2024 (UTC)[reply]
If we'd want participation, we need to formulate explicit requirements for how many !votes are needed as minimum participation. A low participation pass for EFH was actually me from a year ago with only 4 support. 0xDeadbeef→∞ (talk to me) 12:23, 30 March 2024 (UTC)[reply]
I think that determination may be up to the closer. I personally wouldn't say it's low participation simply because your request had support from SoY, zzuuzz, Compassionate727 and Red-tailed hawk. It was obviously passing even if left open for more comments. That's in part because the number of active EFMs was quite small, and those two have been persistently dedicated to filters. But also in part because SoY and zzuuzz have criteria which I think roughly matches this board's criteria holistically. As a combination of both factors, I personally would tend to trust their judgement on a request, and I imagine others here do too. ProcrastinatingReader (talk) 12:32, 30 March 2024 (UTC)[reply]
I don't think low participation is really a thing on this noticeboard. It's a very small community of people that have the desire and technical knowledge to modify the filters. There are millions of accounts on this site. Of those, 864 at the time of writing are administrators. A lot of the edit filter managers are admins, and didn't go through a consensus !vote on EFN, but rather self-granted as admins. Depending on their interests, they may or may not be involved in edit filters to a deep degree, they may have just given themselves the bit for one edit and forgotten to take it off. Out of the 139 EFMs, I counted at one point and a little over 10 aren't admins. Next, there are 23 edit filter helpers. That means overall, there are maybe 30 non-admins plus a few unflagged who are involved in edit filters. Add in the few admins who are also involved here, and we maybe have 40-50 people in the edit filter "community". I'm sure my count is probably a bit off, but those who toil and tinker with the edit filters aren't a large community, so unless there's serious concerns about someone as a candidate, I don't see that much of a problem with low participation. EggRoll97(talk) 15:56, 30 March 2024 (UTC)[reply]
It'd be real nice to stop these third party nominations for both EFH and EFM, at least those that are absent the nominee providing a few sentences of why they need the role or what benefit they can bring. The request for 1AmNobody24 passed without any writings from them besides a four word acceptance and a response to a specific point brought up in an oppose !vote. I'm sure some rationale was included in Codename Noreste's nomination after the two of them had a discussion who-knows-where, but we should really be hearing from the nominees themselves in these discussions.
Totally agreed. This is a process where it makes sense that self-noms are a norm.
Extending EFH to seven days also makes sense as we have not failed to process the backlog. The need for EFH will be decreased for each additional EFH we take. 0xDeadbeef→∞ (talk to me) 14:07, 30 March 2024 (UTC)[reply]
If there is consensus to disallow. The reason for not allowing third-party noms has been said above: we should really be hearing from the nominees themselves in these discussions. 0xDeadbeef→∞ (talk to me) 14:55, 30 March 2024 (UTC)[reply]
Oh, I see. I also don't see any objection to not allow third party nominations, but we are going to need consensus for that and to extend the EFH nomination for one week, just like running for EFM. Codename Noreste 🤔 talk15:21, 30 March 2024 (UTC)[reply]
I'd also like to introduce another possibility where non self-noms are allowed but the nominee must also add a few extra sentences at least with some sort of rationale so we can hear from them too. – PharyngealImplosive7(talk)00:04, 31 March 2024 (UTC)[reply]
For example, this might include: it's 0xDeadbeef's response to my request for a bit of advice and his encouragement of boldness here that has pushed me to bite the metaphorical "bullet". Codename Noreste 🤔 talk00:41, 31 March 2024 (UTC)[reply]
Support amending earliest close time to 1 week for efh. In the future may also want to consider requiring at least one bolded !vote from an efm. No efm participation, or low efm/efh participation in the discussion, suggests to me that the discussion should be open longer so those folks can chime in. –Novem Linguae (talk) 15:41, 30 March 2024 (UTC)[reply]
Support on extending to one week, and on requiring self-noms. I don't think third-party nominations are really very helpful here. EggRoll97(talk) 15:56, 30 March 2024 (UTC)[reply]
Support both extending to 1 week and only self-nominations(for EFH and EFM), more opportunity for input from people who aren't here so often can only do good, and self-noms are easier than having to coordinate third-party nominations with the candidate so they can provide rationale. – user in the /32 - currently 2804:F1...54:171E (talk) 22:45, 30 March 2024 (UTC)[reply]
Support one week. I also think we need a statement from any applicant, and I'll just say that I don't like 3rd party nominations for this role. Just ask for a support statement instead. (I can vaguely imagine scenarios where some highly experienced user might want to introduce an application, but it can probably be done with just a supporting statement). I'd also support some type of quorum, though I think admins closing these applications should probably know what they're doing here, and not rush to close something just because it's past the due date. Being a minor venue, I think admins should be using a wide discretion when it comes to process here. In a sense, the closing admin also needs to be a supporter and not just a closing-bot, which should add to the quorum. -- zzuuzz(talk)00:23, 31 March 2024 (UTC)[reply]
Their are a few promotional accounts whose names have 'corporations' in them instead of simply 'corporation' which is currently filtered out. I suggest that we should change the syntax to also log these accounts. We could change the related part of the regex to CORP(?:S?.?$|ORATE|ORATIONS?\b). – PharyngealImplosive7(talk)00:29, 31 March 2024 (UTC)[reply]