The following discussion is an archived debate. Please do not modify it. To request review of this BRFA, please start a new section at WT:BRFA. The result of the discussion was Approved.
Function details: This bot was doing this task for years. The bot will look for ref tags. After that it looks for on of the following section to place the reflist template: References, Footnotes, Notes. If no section is given, it places a References section before one of the following: Further reading, External links, See also, Notes. Yobot will use AddMissingReflist function of AWB or the built-in code of WPCleaner.
Please re-review the required fields on the approval request. I believe you forgot to fill out the bottom half of it. ~ Rob13Talk 11:11, 3 February 2017 (UTC)[reply]
Done. I assumed that since they are the same with the previous BRFA the BAG can understand. -- Magioladitis (talk) 12:50, 3 February 2017 (UTC)[reply]
The top section of a BRFA is not just for BAG to understand, it's also to record the parameters of the approval. In this case, the "Function details" should describe how you will determine where to insert the references section. Experimentation during the approval process is not the time to begin determining this. I recommend you download the wikitext of recent good and featured articles, remove the "References" sections, and then do your testing offline until you have an algorithm that readds the "References" sections in the same locations as they were in the original articles. Anomie⚔ 13:30, 3 February 2017 (UTC)[reply]
"The problem is where to exactly place it." Sounds like a standard case of WP:CONTEXTBOT. You need to specify exactly where the tag and section will go and--if it's not supported by a layout policy/guideline--that there is consensus to place it that way. Sounds like tagging the articles and/or producing a list for human editors to review is more appopriate. — HELLKNOWZ ▎TALK 22:04, 3 February 2017 (UTC)[reply]
Just spitballing: If external links heading present, place above that. If external links heading not present but categories present, check for templates immediately above categories. If there are a series of templates immediately above categories, place above that (intended to handle navboxes). If there aren't a series of templates immediately above categories, place above the categories. That seems like it would handle this with a relatively high degree of accuracy if you can write regex to do it. We'd need to see a trial to determine what the error rate is, but if regex can be written to use that algorithm, I think a trial can be approved. ~ Rob13Talk 23:03, 3 February 2017 (UTC)[reply]
Hellknowz handles most fo the cases. I won;t do cases AWB does not do. BU Rob13 this is what AWB does. I will do this. -- Magioladitis (talk) 23:15, 3 February 2017 (UTC)[reply]
How will the bot deal with pages that use Harvard refs, but where someone accidentally placed a reference with <ref> tags? In these cases, the correct fix is to change the new reference to the existing style, rather than to add a reflist template to the existing page. It appears to me that manual review is needed to tell if there are existing references in another style. — Carl (CBM · talk) 03:58, 9 February 2017 (UTC)[reply]
Are you volunteering? All the best: RichFarmbrough, 23:32, 9 February 2017 (UTC).[reply]
A BRFA determines whether there is consensus for a task to be performed and whether the implementation being used is sufficient to handle the task in a proper and error-free manner. It's not an exercise in allocating volunteer time. ~ Rob13Talk 03:12, 10 February 2017 (UTC)[reply]
The task was performed the period 2010-2017 with no problems. Some bots still perform it. So the question is void. -- Magioladitis (talk) 12:38, 10 February 2017 (UTC)[reply]
Here is an example of an incorrect edit from 2016: [1]. That article uses a different citation style, which required a different fix. These are hard for a bot to avoid entirely, but in general the bot should make some sort of attempt to detect this condition. For example, the presence of any of the harvtxt family of templates, or matching the regex /\(\w+ \d\d\d\d\)/, could be signs that the page should be examined manually. — Carl (CBM · talk) 13:06, 10 February 2017 (UTC)[reply]
CBM The bot did not introduce the error. It made the error easier to find and fix by normal editors who watch the page. I could try an approach but not on my top priority since it looks a bit rare for me. Thanks for the report! -- Magioladitis (talk) 15:31, 10 February 2017 (UTC)[reply]
I am sorry that this bot request is not your top priority; perhaps you have too many of them open? I don't support the bot request if the bot task will blindly add the reflist tag without any effort to detect pages where it should not be added. That only adds a second error, rather than attempting to fix the original one. — Carl (CBM · talk) 15:37, 10 February 2017 (UTC)[reply]
This sort of wilful misinterpretation is not helpful. All the best: RichFarmbrough, 20:54, 10 February 2017 (UTC).[reply]
I need to see some example edits in semi-auto to get a feel of what this task would do. The problem originally was where to place a references section. I suggested pseudo-code to handle that, and then was told it already does that. That seems contradictory, so I'm left unsure what exactly this bot will be doing and how it will do it. ~ Rob13Talk 15:41, 10 February 2017 (UTC)[reply]
CBM do the task is top priority. Fixing a minor issue is not. -- Magioladitis (talk) 15:48, 10 February 2017 (UTC)[reply]
You are putting the cart before the horse. The point of a BRFA is to make sure that issues are addressed, not to rubber stamp the request. There is no reason overall why this needs to be done by Yobot. If a bot operator does not constructively address issues that are raised, the request should probably not be approved. — Carl (CBM · talk) 15:54, 10 February 2017 (UTC)[reply]
CBM the thing that you are here making a different request. I discuss pages that for some reason have ref tags and you ask me to do something for pages that use mixed system. At the meantime, Mediawiki auto-generates a reflist tag. As I said, I 'll look into it but do you see the problem here? -- Magioladitis (talk) 15:56, 10 February 2017 (UTC)[reply]
The page did not use a mixed system; it used Harvard refs. Then someone erroneously added a ref in a different format. That error should be fixed by changing the new ref to match the previous formatting, not by adding a reflist tag (which is a second error, because the established style did not use footnotes). The bot code should include some logic to try to detect this situation (e.g by looking for the templates in the harvtxt family), so that it does not mistakenly add a reflist tag where none should be added. This is not a different request - it is an aspect of the bot task that is being proposed. — Carl (CBM · talk) 16:02, 10 February 2017 (UTC)[reply]
This is sort of what I implied above. The bot should be able to skip pages where it finds conflicting references or any syntax that is likely to convolute things further. I'm still waiting on full function details though. I don't think we should require the bot to be able to deal with different styles, though that's a supreme task. The bot operator should acknowledge these related tasks and how their edits may or may not affect the article's ref style and what they intend to do about them and how they avoid problematic cases or further messing up the article. — HELLKNOWZ ▎TALK 16:16, 10 February 2017 (UTC)[reply]
Hellknowz The bot won't affect the existing ref style on any page. I added details and a link to the other bot that was doing the same job using python. -- Magioladitis (talk) 20:15, 10 February 2017 (UTC)[reply]
"The bot won't affect the existing ref style on any page." - that is exactly what happened in the link I gave. The existing style did not use footnotes, but one was added erroneously. Adding a {{reflist}} tag only further changes the existing style. Are you saying that the bot will now do something different? Hellknowz wrote "The bot should be able to skip pages where it finds conflicting references or any syntax that is likely to convolute things further", and I also think that would be ideal. — Carl (CBM · talk) 01:04, 11 February 2017 (UTC)[reply]
{{BAGAssistanceNeeded}} I once more underline the fact that this is a take over of a bot that was already doing it. -- Magioladitis (talk) 02:06, 23 February 2017 (UTC)[reply]
Note - Maglioladitis has been doing this process in a very automated way using their account. I will be filing with ArbCom about it. Primefac (talk) 12:31, 29 March 2017 (UTC)[reply]
Yup, and points 5 and 6 pretty clearly state that you shouldn't be doing semi-automated processes that Yobot would be doing if the Yobot task is not approved and/or halted. Primefac (talk) 12:47, 29 March 2017 (UTC)[reply]
Primefac Yobot is not doing this task. The points refer to tasks that will be approved after the ArbCom. -- Magioladitis (talk) 12:49, 29 March 2017 (UTC)[reply]
Clearly not. However, if Yobot isn't approved to do a bot task, should you really be doing the same automated task on your main account? That's one of the main reasons why you were brought before ArbCom in the first place. Primefac (talk) 12:52, 29 March 2017 (UTC)[reply]
You're an odd cat, Magioladitis. Primefac (talk) 13:12, 29 March 2017 (UTC)[reply]
Primefac True :) But this process after 2+ months is exhausting. I try to help the best way I can. Every action I try gets complaints lately. If I am not welcome somewhere I can just leave. Wikipedia is a thing I do everyday since the time I wake up till the time I go to sleep but if the rules has changed I will OK with it. This time I think I followed the rules I was given if not, so be it. Mea culpa. -- Magioladitis (talk) 13:22, 29 March 2017 (UTC)[reply]
Approved for trial (25 edits). Please provide a link to the relevant contributions and/or diffs when the trial is complete. Now that the brouhaha has died down, let's see some trials. The above objections are noted, but is nothing that should prevent a bot from running. Sspecifically, if there are Harvard citations as a dominant style, and a <references/> is missing because of one citation not following Harvard style is given, then the bot puts the error in everyone's face, which will lead to someone fixing it, and the article falling back into compliance. Editing conditions are
The reference list must actually be missing. There must be an active check that no reference list exists via <references/> or {{Reflist}} (and variants, T162492 is relevant here). If it finds anything (I suggest the "(<\s*reference|\{\{\s*reflist)" regex), the bot must skip the page, regardless if other fixes could be done.
AWB genfixes may not be done on their own as part of this task.
AWB genfixes may be enabled, but only if this is explicitly mentioned in the edit summary (T161460 is relevant here, but a 'manual' edit summary that does the same thing is acceptable)
If AWB genfixes are enabled, the skip if cosmetic-only and minor genfixes-only options must be enabled.
Edit summary must be clear about what the task is, and where to report issues.
Verify that and and other variants trigger the skip condition. This may be done in a sandbox. This specific item must be done before the trial edits are performed. Headbomb {t · c · p · b} 20:34, 17 April 2017 (UTC)[reply]
@Headbomb: If we're intending the likely errors identified above to trigger editor attention, this task should be done (a) without a bot flag, and (b) with minor edit unchecked. The error isn't "put in everyone's face" if it doesn't appear on watchlists. ~ Rob13Talk 20:38, 17 April 2017 (UTC)[reply]
The error is not very likely to occur in the first place, and would already have shown up in watchlists (likely not marked as minor either). However, I'm open to have this be considered a major edit, given it's adding a section that didn't exist. But the bot flag should be used, since this doesn't need to show up in recent changes, not should people be denied the option to hide this stuff from their watchlist. Headbomb {t · c · p · b} 20:45, 17 April 2017 (UTC)[reply]
I could get behind that as a compromise. ~ Rob13Talk 03:44, 18 April 2017 (UTC)[reply]
I ll start this after my wikibreak. -- Magioladitis (talk) 00:06, 1 June 2017 (UTC)[reply]
I'm commenting not as a BAG member but as a regular editor, per my recusal from Yobot tasks. I wanted to note that the edit summary features a redlink; I believe you intended it to link to here. Simple fix. Second, I wanted to reiterate that I think this should run without the minor edit flag both due to potential errors and because a whole section is being added, as per above. ~ Rob13Talk 00:31, 8 July 2017 (UTC)[reply]
A user has requested the attention of a member of the Bot Approvals Group. Once assistance has been rendered, please deactivate this tag by replacing it with {{t|BAG assistance needed}}. --- Magioladitis (talk) 09:29, 15 July 2017 (UTC)[reply]
@Magioladitis: Those edits are from Task 57 and Task 54. Please link the ones for the Task 52 trials. Headbomb {t · c · p · b} 18:58, 7 August 2017 (UTC)[reply]
@Magioladitis: I take it the June 20 edits were not part of the trial? Also, concerning [3], I can't reproduce the removal of the sortkey with my version of AWB. What gives? Headbomb {t · c · p · b} 00:05, 8 August 2017 (UTC)[reply]
Headbomb Yobot did 25 edits. 2 pages of those edited were probably deleted and do not show in the contributions list. This causes the previous edits to show in the list. For the removal of the sortkey: Check whether you have "Restrict DEFAULTSORT addition" in your settings list on. Since the sortkey is a substring of the DEFAULTSORT is being removed. -- Magioladitis (talk) 10:38, 8 August 2017 (UTC)[reply]
That's what it was yes. Can you give any insight on why it's turned off by default? Headbomb {t · c · p · b} 17:38, 8 August 2017 (UTC)[reply]
Headbomb For compatibility with some very old AWB settings AFAIK. I recommend that you turn it on. -- Magioladitis (talk) 18:08, 8 August 2017 (UTC)[reply]
Approved. Alright, well the bot function as intended, and I couldn't find any edit it shouldn't have done, as it always added missing reflists. Going to WP:AGF on the category sortkey thing, but if you get pushback on it, I expect you to stop until you get the thumbs up to resume that. Headbomb {t · c · p · b} 23:19, 8 August 2017 (UTC)[reply]
The above discussion is preserved as an archive of the debate. Please do not modify it. To request review of this BRFA, please start a new section at WT:BRFA.
Approved.
Approved for trial (25 edits). Please provide a link to the relevant contributions and/or diffs when the trial is complete. Now that the brouhaha has died down, let's see some trials. The above objections are noted, but is nothing that should prevent a bot from running. Sspecifically, if there are Harvard citations as a dominant style, and a <references/> is missing because of one citation not following Harvard style is given, then the bot puts the error in everyone's face, which will lead to someone fixing it, and the article falling back into compliance. Editing conditions are
Trial complete. -- Magioladitis (talk) 23:21, 7 July 2017 (UTC)[reply]
A user has requested the attention of a member of the Bot Approvals Group. Once assistance has been rendered, please deactivate this tag by replacing it with 
