Wikipedia:Bots/Requests for approval/TheSandBot
- 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.
Operator: TheSandDoctor (talk · contribs · SUL · edit count · logs · page moves · block log · rights log · ANI search)
Time filed: 00:47, Wednesday, October 17, 2018 (UTC)
Automatic, Supervised, or Manual: automatic
Programming language(s): Python
Source code available: https://github.com/TheSandDoctor/election_converter
Function overview: Looks through the linked csv file, converting from the old title format to the new one.
Links to relevant discussions (where appropriate): [1], RfC on election/referendum article naming format
Edit period(s): Run until done
Estimated number of pages affected: Approximately 35, 227
Exclusion compliant (Yes/No): No
Already has a bot flag (Yes/No): No
Function details: The bot goes through the compiled csv file (in GitHub repo, more easily read form is the Excel document, which is also included). The bot then pulls up the individual page objects, double checks that they are not themselves redirects (ie that they haven't been moved) and exist. If both conditions are satisfied, the bot moves the page (leaving behind a redirect) to the corresponding title in column B (.xlsx doc). This corresponds with the latest RfC on election/referendum article naming format and was created per request of Number 57.
The code itself is relatively straight forward, with most of the heavy lifting being handled by the mwclient Python library's move function, which is a part of the page object.
Due to the large number of page moves required, I would also request that the bot flag be assigned should this request be approved. The bot is not exclusion compliant as that is non-applicable given the context.
Discussion
[edit]- Would it be possible for you to generate a randomized list of 100-500 articles to be moved, and the titles that they would be moved to in your userspace please? SQLQuery me! 01:11, 17 October 2018 (UTC)[reply]
- @SQL: Certainly! It can be viewed here:
User:TheSandDoctor/sandbox2User:TheSandBot/sandbox2. The first 150 (of 151) were generated at complete random (though I did ensure no duplicates were chosen) by a Python script. Since there are over 35 thousand and only 150 selected at (near as computers can get) random (so ~0.43% of the articles, assuming my math is correct), I included a variant not shown by the randomly generated list (#151). --TheSandDoctor Talk 01:49, 17 October 2018 (UTC); sandbox moved. updated to reflect this 05:20, 29 November 2018 (UTC)[reply]- Here is the random generation code, if you are interested. Incase it wasn't clear before, this is the script itself. I should really clean up the repo... --TheSandDoctor Talk 01:54, 17 October 2018 (UTC)[reply]
- This is a very large request. I'd love some more opinions from other recently active BAG members. Xaosflux, MusikAnimal, Anomie, The Earwig, Headbomb, Cyberpower678. SQLQuery me! 02:12, 17 October 2018 (UTC)[reply]
- If it is determined that throttling would be required, the addition of a wait timer would be relatively trivial. All I would need is the time you want it to wait between moves and that could be added rather painlessly. --TheSandDoctor Talk 02:04, 17 October 2018 (UTC)[reply]
- @SQL: my concern is determining if there was sufficient input to that RfC, it closed with ~69% support over only 16 editors despite a claim that it would be advertised to "as many relevant WPs as I can find". At the very least notice of this at WP:VPR would be a good "last chance" notification to editors at large. — xaosflux Talk 03:17, 17 October 2018 (UTC)[reply]
- I would support that. The absolute number of pages involved isn't very large for a bot, but it's the kind of change that can be very contentious, especially if people feel they weren't notified. I generally support the proposal, for what it's worth, but consensus isn't as clear as I would've liked. — Earwig talk 03:22, 17 October 2018 (UTC)[reply]
- It seems like a WP:VPR discussion and people reviewing this list would both be helpful here. Anomie⚔ 12:34, 17 October 2018 (UTC)[reply]
- Number 57 would you like to do the honours, or would you like me to? I agree that bringing attention to this BRFA at WP:VPR might be a good idea. --TheSandDoctor Talk 20:34, 17 October 2018 (UTC)[reply]
- I'll give it a go. Please amend what I've done if it's not right (never started a discussion there before). Cheers, Number 57 21:06, 17 October 2018 (UTC)[reply]
- I'll keep an eye out for it Number 57 --TheSandDoctor Talk 21:14, 17 October 2018 (UTC)[reply]
- I'll give it a go. Please amend what I've done if it's not right (never started a discussion there before). Cheers, Number 57 21:06, 17 October 2018 (UTC)[reply]
- Number 57 would you like to do the honours, or would you like me to? I agree that bringing attention to this BRFA at WP:VPR might be a good idea. --TheSandDoctor Talk 20:34, 17 October 2018 (UTC)[reply]
- This is a very large request. I'd love some more opinions from other recently active BAG members. Xaosflux, MusikAnimal, Anomie, The Earwig, Headbomb, Cyberpower678. SQLQuery me! 02:12, 17 October 2018 (UTC)[reply]
- Here is the random generation code, if you are interested. Incase it wasn't clear before, this is the script itself. I should really clean up the repo... --TheSandDoctor Talk 01:54, 17 October 2018 (UTC)[reply]
- @SQL: Certainly! It can be viewed here:
- No objections here at running it at full speed. Getting it done as fast as possible would present minimal disruption to people's watchlists. We just need to make sure the renames work correctly. We wouldn't want a bunch of moves going to "Test move" or "Oopsie".—CYBERPOWER (Trick or Treat) 02:37, 17 October 2018 (UTC)[reply]
- "Test move" and "oopsie" are not in the list, I can assure you that . --TheSandDoctor Talk 02:41, 17 October 2018 (UTC)[reply]
- Conventional wisdom is that you want to go slowly enough to allow for manual review and to minimize disruption if something goes wrong, especially for a task that is not time-sensitive like this one. Unless the bot flag doesn't work for moves (?), watchlist disruption shouldn't be an issue. I would recommend at least a few seconds between edits. — Earwig talk 03:16, 17 October 2018 (UTC)[reply]
- @Earwig: Easily done. The script is currently configured with a 4 second delay, but that could be changed in less time than it took to write this sentence. --TheSandDoctor Talk 04:02, 17 October 2018 (UTC)[reply]
- I'm not convinced all of the proposed wordings are correct. For example, Oregon Ballot Measure 58 (2008) should probably not be moved. I see a lot of similar issues in the .csv with other propositions/ballot measures; e.g. California Proposition 10 (1998) going to 1998 California Proposition 10 seems wrong. There are some other strange wordings: should Polish presidential election, 1922 (special) go to 1922 Polish presidential election (special) as suggested or 1922 Polish presidential special election or similar? Perhaps we can come up with a tighter definition of the grammar for acceptable renames, like leaving titles with parentheses for manual review. — Earwig talk 02:54, 17 October 2018 (UTC)[reply]
- @The Earwig: If a title contains parenthesis anywhere, it could certainly be compiled into its own list, recorded, and skipped over in the move. Would only add a couple of lines. The thing is, for that sort of thing, I need well defined and clear rules in order to write the regex to test for. --TheSandDoctor Talk 02:57, 17 October 2018 (UTC)[reply]
- Right, and I'm not sure what that would look like yet. I noticed another strange phrasing, which would currently move Ohio's 13th congressional district election, 2006, to 2006 Ohio's 13th congressional district election. So at the very least, we can probably have extra eyes on titles with parentheses or apostrophes, and titles without "election" or "referendum"? — Earwig talk 03:16, 17 October 2018 (UTC)[reply]
- @The Earwig: If a title contains parenthesis anywhere, it could certainly be compiled into its own list, recorded, and skipped over in the move. Would only add a couple of lines. The thing is, for that sort of thing, I need well defined and clear rules in order to write the regex to test for. --TheSandDoctor Talk 02:57, 17 October 2018 (UTC)[reply]
- The change file includes moving non-year values to the start of the title, the MOS doesn't appear to address this, nor did the RfC. e.g. French constitutional referendum, October 1946 (Guinea) --> October 1946 French constitutional referendum (Guinea). Do you mean to sort these with "O"? — xaosflux Talk 03:21, 17 October 2018 (UTC)[reply]
- @Xaosflux: The list was compiled by Number 57, so they would probably be the best to ask. That said, it does appear to be the case and does make logical sense, given the RfC and its approaches with the postfix years. The whole purpose of the RfC appears to be moving like this. Moving otherwise would not make sense. This is not to comment on the above point or "(Guinea)" not being moved, just "October 1946". --TheSandDoctor Talk 04:01, 17 October 2018 (UTC)[reply]
- @TheSandDoctor: so in this example, the SORTKEY is currently under "French con..", now it will be under "October" (any very specifically NOT under "1946") - unless additional sortkey adjustments are made. What is the category sorting goal? — xaosflux Talk 04:04, 17 October 2018 (UTC)[reply]
- @Xaosflux: Number 57 is going to have to answer that one. I just saw the WP:BOTREQ and made the bot. I will happily answer or give my assessment on the technical side of things (related to the script), but the excel document and the RfC was Number's brainchild. --TheSandDoctor Talk 04:09, 17 October 2018 (UTC); expanded for clarity 04:22, 17 October 2018 (UTC)[reply]
- @TheSandDoctor: so in this example, the SORTKEY is currently under "French con..", now it will be under "October" (any very specifically NOT under "1946") - unless additional sortkey adjustments are made. What is the category sorting goal? — xaosflux Talk 04:04, 17 October 2018 (UTC)[reply]
- @Xaosflux: The list was compiled by Number 57, so they would probably be the best to ask. That said, it does appear to be the case and does make logical sense, given the RfC and its approaches with the postfix years. The whole purpose of the RfC appears to be moving like this. Moving otherwise would not make sense. This is not to comment on the above point or "(Guinea)" not being moved, just "October 1946". --TheSandDoctor Talk 04:01, 17 October 2018 (UTC)[reply]
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 new pages with "–" in the titles, now at the start for ranges, are you going to make redirects from the common redirect "-"? — xaosflux Talk 03:25, 17 October 2018 (UTC)[reply]
- Unless I'm missing something, or don't understand - I don't see any titles matching " - " in the docx. SQLQuery me! 03:52, 17 October 2018 (UTC)[reply]
- @SQL: For example new page will be July–August 1990 Bulgarian presidential election, may need a redirect from July-August 1990 Bulgarian presidential election created. — xaosflux Talk 04:02, 17 October 2018 (UTC)[reply]
- @Xaosflux: The word "may" could be problematic here as scripts can't do human thought. That would probably necessitate output to be posted in the user space for human editors to look over and make the call if we go down that route (definitely possible, would just need to decide between the bot posting it or it generating a file and myself periodically updating the page). Aside from that though (ignoring it momentarily, if you will), a regex could be crafted to scan a title looking for '–'s and then launch another method to create a redirect without (assuming '–'s are present). --TheSandDoctor Talk 04:14, 17 October 2018 (UTC)[reply]
- I'm only calling it out for commentary here, I'm not that current on MOS about dashes and don't want to fall asleep on my keyboard reading the MOS right now! — xaosflux Talk 04:20, 17 October 2018 (UTC)[reply]
- @Xaosflux: Not a problem and I hear you . I am happy to work with the community on this and share the technical knowledge I have. I am hoping that we can iron out the details regarding this. I still believe that some sort of a bot is needed for this, should the RfC stand, since 35k+ articles is a tad too much to do by hand very easily. --TheSandDoctor Talk 04:27, 17 October 2018 (UTC)[reply]
- I'm only calling it out for commentary here, I'm not that current on MOS about dashes and don't want to fall asleep on my keyboard reading the MOS right now! — xaosflux Talk 04:20, 17 October 2018 (UTC)[reply]
- @Xaosflux: The word "may" could be problematic here as scripts can't do human thought. That would probably necessitate output to be posted in the user space for human editors to look over and make the call if we go down that route (definitely possible, would just need to decide between the bot posting it or it generating a file and myself periodically updating the page). Aside from that though (ignoring it momentarily, if you will), a regex could be crafted to scan a title looking for '–'s and then launch another method to create a redirect without (assuming '–'s are present). --TheSandDoctor Talk 04:14, 17 October 2018 (UTC)[reply]
- @SQL: For example new page will be July–August 1990 Bulgarian presidential election, may need a redirect from July-August 1990 Bulgarian presidential election created. — xaosflux Talk 04:02, 17 October 2018 (UTC)[reply]
- Isn't this already covered by Anomie Bot? ~ Amory (u • t • c) 15:22, 17 October 2018 (UTC)[reply]
- I think you're right; as an example, when 2018–19 Southern Football League was created, Anomie Bot created 2018-19 Southern Football League a few hours afterwards. Number 57 16:11, 17 October 2018 (UTC)[reply]
- I looked through the bot's user page and did not see anything covering this BRFA. That said, after reading Number's response (which occurred while I was looking), I realize that that appears to have not been what you meant. In that case, then there probably wouldn't be any issues whatsoever on this particular point/thread. --TheSandDoctor Talk 16:22, 17 October 2018 (UTC)[reply]
- Yes, sorry — this was threaded to be in reply to Xaosflux. It's AnomieBOt 74, and works like a charm (too well, really; I see plenty of these at G8 patrolling). ~ Amory (u • t • c) 19:24, 17 October 2018 (UTC)[reply]
- I looked through the bot's user page and did not see anything covering this BRFA. That said, after reading Number's response (which occurred while I was looking), I realize that that appears to have not been what you meant. In that case, then there probably wouldn't be any issues whatsoever on this particular point/thread. --TheSandDoctor Talk 16:22, 17 October 2018 (UTC)[reply]
- I think you're right; as an example, when 2018–19 Southern Football League was created, Anomie Bot created 2018-19 Southern Football League a few hours afterwards. Number 57 16:11, 17 October 2018 (UTC)[reply]
- Unless I'm missing something, or don't understand - I don't see any titles matching " - " in the docx. SQLQuery me! 03:52, 17 October 2018 (UTC)[reply]
- To respond to some of the various points above:
- The RfC was advertised to the following WikiProjects: Elections and Referendums, Politics, Politics of the United Kingdom, U.S. Congress, Pakistani politics, New Zealand/Politics, Indian politics, Chinese politics and Australian politics; these are all the politics-related WikiProjects that I could find. On a vote-counting basis it is 69% in favour, but a couple of the oppose !votes are on dubious grounds (one being because an editor didn't believe that redirects appear in search, and another one who claimed they had never seen a year at the start of an article title), so I think the consensus (in terms of what a closing admin would determine) from the discussion is pretty undeniable.
- The proposed moves to 1998 California Proposition 10 etc are in line with the naming guideline (see the last bullet at WP:NC-GAL#Elections and referendums). The Oregon ones are currently incorrectly titled, so the move is to bring them in line with the guideline. We could have a conversation at a later date about whether the year is required at all for these types of articles (I'm not convinced it works), but currently they are in the guideline as such. If it's really a problem, perhaps we could drop them from the run?
- Polish presidential election, 1922 (special) is at the wrong title (one a few days before is at Polish presidential election, 9 December 1922, so the other one should be at Polish presidential election, 20 December 1922). This should therefore probably be moved to 20 December 1922 Polish presidential election;
- 2006 Ohio's 13th congressional district election is again a correct move in terms of the guideline (sixth bullet of WP:NC-GAL#Elections and referendums). The issue here is more around the awkward naming of the districts (e.g. Ohio's 13th congressional district) and is perhaps something that should be raised separately;
- The RfC did include discussion about titles that would start with a month (see Impru20's comments). The sortkey for articles like this would still be the year, followed by a numeral representing the month (e.g. "1946 1", "1946 2" for elections held in two separate months in 1946)
- Cheers, Number 57 09:02, 17 October 2018 (UTC)[reply]
- Looks like some bot coding is going to be needed to add/alter sort keys following the moves for new titles not starting with years. — xaosflux Talk 10:56, 17 October 2018 (UTC)[reply]
- I'm not sure this is really needed; the current format means that articles don't automatically sort by year, so in many cases a sortkey has already been added. For instance, French constitutional referendum, October 1946 (Guinea) mentioned above is sorted in Category:Referendums in Guinea by the key 1946. This might be something better to do manually for the 380 articles with months in the year if not already in place. Number 57 13:00, 17 October 2018 (UTC)[reply]
- For the other 4 categories it is in it is sorted only by page title such as Category:October 1946 events. I'm not sure what the 'best' answer for this is, but if doing it manually is the way that could be done prior to the page title moves to prevent issues in category view. — xaosflux Talk 13:08, 17 October 2018 (UTC)[reply]
- I think the question there is what is the best category sortkey for the article in Category:October 1946 events. Number 57 13:10, 17 October 2018 (UTC)[reply]
- Pages can be sorted differently in each category with a directive, but they only get one "default sort" for all undirected categories, so if in general sorting these be "month name" is undesirable a default sort should be defined for what the best general sorting should be. — xaosflux Talk 20:43, 17 October 2018 (UTC)[reply]
- There doesn't seem to be any firm conventions around this – some articles starting with a year are sorted by the year in the events categories, and others by the first word after the year. However, in election categories sorting by year would definitely be desirable, so I guess the year followed by the month would be the best sorting (e.g. 1946 01, 1946 02 to 1946 12). Happy to add a DEFAULTSORT manually to these articles if it will resolve this concern for you? Cheers, Number 57 21:06, 17 October 2018 (UTC)[reply]
- Pages can be sorted differently in each category with a directive, but they only get one "default sort" for all undirected categories, so if in general sorting these be "month name" is undesirable a default sort should be defined for what the best general sorting should be. — xaosflux Talk 20:43, 17 October 2018 (UTC)[reply]
- I think the question there is what is the best category sortkey for the article in Category:October 1946 events. Number 57 13:10, 17 October 2018 (UTC)[reply]
- For the other 4 categories it is in it is sorted only by page title such as Category:October 1946 events. I'm not sure what the 'best' answer for this is, but if doing it manually is the way that could be done prior to the page title moves to prevent issues in category view. — xaosflux Talk 13:08, 17 October 2018 (UTC)[reply]
- I'm not sure this is really needed; the current format means that articles don't automatically sort by year, so in many cases a sortkey has already been added. For instance, French constitutional referendum, October 1946 (Guinea) mentioned above is sorted in Category:Referendums in Guinea by the key 1946. This might be something better to do manually for the 380 articles with months in the year if not already in place. Number 57 13:00, 17 October 2018 (UTC)[reply]
- Looks like some bot coding is going to be needed to add/alter sort keys following the moves for new titles not starting with years. — xaosflux Talk 10:56, 17 October 2018 (UTC)[reply]
- The WP:VPR post can be viewed here. Xaosflux, MusikAnimal, Anomie, The Earwig, Headbomb, Cyberpower678 --TheSandDoctor Talk 19:27, 18 October 2018 (UTC)[reply]
- Oppose this job. Per my comments at Wikipedia talk:Naming conventions (government and legislation)#RFC_inadequate,_bot_not_justified I think that
- the RFC was inadequate for the scale of changes proposed
- I see no consensus anywhere to bypass WP:RM's consensus-forming process for the 35,226 articles involved.
- WP:NCGAL is a guideline, which like other guidelines says prominently at the top
It is a generally accepted standard that editors should attempt to follow, though it is best treated with common sense, and occasional exceptions may apply
. The use of a bot on such a huge scale precludes the consideration of exceptions. --BrownHairedGirl (talk) • (contribs) 03:03, 19 October 2018 (UTC)[reply]
- PS The discussion above about some exceptions such as the naming of some Polish elections should not be conducted on this page. I am sure that @Number 57 is making recommendations on a well-reasoned basis, but decisions on the titles of individual articles should not buried in a technical page such as this; they belong at WP:RM. --BrownHairedGirl (talk) • (contribs) 03:08, 19 October 2018 (UTC)[reply]
- This isn't RFA. Bolded "votes" aren't needed, or helpful. I do believe that this idea needs more discussion, and I'm surprised (in a good way!) at how much attention, and the overall quality of comments that this request is getting. I'm not sure about listing ~35,000 pages at WP:RM. I could see some parties seeing that as disruptive. SQLQuery me! 03:28, 19 October 2018 (UTC)[reply]
- @SQL: no mater how many good comments there are here, this remains a technical page whose remit is to decide whether and how to use bots to implement a consensus. It is not a suitable place to form a consensus on whether to bypass WP:RM on this scale; that is right outside the remit of WP:BAG. --BrownHairedGirl (talk) • (contribs) 03:42, 19 October 2018 (UTC)[reply]
- Right. I have an idea of how this works. I've been around a BRFA or two. You can see comments above questioning the RFC by other BAG members. I'm not sure where you get the idea that this is just going to pass without addressing that. SQLQuery me! 04:15, 19 October 2018 (UTC)[reply]
- My point is simply that BRFA is not the place to address those issues. BRFA's role starts when they have been resolved. --BrownHairedGirl (talk) • (contribs) 04:50, 19 October 2018 (UTC)[reply]
- Right. I have an idea of how this works. I've been around a BRFA or two. You can see comments above questioning the RFC by other BAG members. I'm not sure where you get the idea that this is just going to pass without addressing that. SQLQuery me! 04:15, 19 October 2018 (UTC)[reply]
- @SQL: no mater how many good comments there are here, this remains a technical page whose remit is to decide whether and how to use bots to implement a consensus. It is not a suitable place to form a consensus on whether to bypass WP:RM on this scale; that is right outside the remit of WP:BAG. --BrownHairedGirl (talk) • (contribs) 03:42, 19 October 2018 (UTC)[reply]
- This isn't RFA. Bolded "votes" aren't needed, or helpful. I do believe that this idea needs more discussion, and I'm surprised (in a good way!) at how much attention, and the overall quality of comments that this request is getting. I'm not sure about listing ~35,000 pages at WP:RM. I could see some parties seeing that as disruptive. SQLQuery me! 03:28, 19 October 2018 (UTC)[reply]
- PS The discussion above about some exceptions such as the naming of some Polish elections should not be conducted on this page. I am sure that @Number 57 is making recommendations on a well-reasoned basis, but decisions on the titles of individual articles should not buried in a technical page such as this; they belong at WP:RM. --BrownHairedGirl (talk) • (contribs) 03:08, 19 October 2018 (UTC)[reply]
- I have raised this at WP:Village pump (policy)#Mass_renaming_of_election_articles,_bypassing_WP:Requested_moves. --BrownHairedGirl (talk) • (contribs) 03:36, 19 October 2018 (UTC)[reply]
- Using some regex magic, I have split the original csv (which is still in the repository) into two camps. format1.xlsx (and .csv) contain the "odd ball" formats which could conceivably be the more contentious of the two groups, given the above discussion. format2.xlsx (and .csv) contain the "election(s), year" (where "year" is the end of the title), which appear to be less contentious per the above. While that doesn't necessarily lessen the problems above, they are now in two distinct datasets easier analyzed. It appears that 21972 are in the latter of the two, with 13253 being in format1. It should, however, be noted that the format1 dataset could be trimmed down by multiple thousands further if the words "referendum", "measures", and each of the state names were considered in the format (instead of just "election(s)"). --TheSandDoctor Talk 05:06, 19 October 2018 (UTC)[reply]
- If anyone wants it, I will split based on those words in 2 other files as well. A bot could conceivably run through format2 and leave the first for human intervention as it appears few (if any) concerns were raised about that case directly. That said, I made this bot based on the original bot request above, which was simple to implement, and submitted it accordingly. If another RfC is desired to further solidify the consensus and address concerns raised, I am all for it and do not wish to rush anything. Pinging all active participants: Number 57, Xaosflux, MusikAnimal, Anomie, The Earwig, Headbomb, Cyberpower678, SmokeyJoe, BrownHairedGirl, SQL. (hope that's everyone, if I missed anyone I apologize). --TheSandDoctor Talk 05:40, 19 October 2018 (UTC)[reply]
- Thanks, @TheSandDoctor. I am sure that you acted in good faith in respinse to the bot request. However, I do think that this request should be placed on hold pending a fresh RFC, and that the bot should not be run unless and until it is clear that there is a v broad consensus on a) the guidelines, and b) the use of a bot to bypass RM for >35k articles. I really don't see any basis for asserting that there is a consensus to rename e.g. the 860 relevant articles under WikiProject Ireland's scope with zero notification to WP:IRE or on any one of the 860 article pages. --BrownHairedGirl (talk) • (contribs) 05:55, 19 October 2018 (UTC)[reply]
- User:TheSandDoctor, I suggest putting a random selection, maybe ten, through the standard RM process. This will draw in critical comments from RM regulars. --SmokeyJoe (talk) 06:48, 19 October 2018 (UTC)[reply]
- @SmokeyJoe: I think that's a good idea, but wildly insufficient. This needs to go way beyond the RM regulars, who are few in number. And Joe, you are rightly critical of how CFD tends to be dominated by regulars. Same goes here.
- This needs to draw in editors who who have sufficient experience of each sub-topic (e.g. Spanish local elections, or Kenyan parliamentary elections) to assess how the broad principle works in their field and hopefully to look for any exception. --BrownHairedGirl (talk) • (contribs) 07:40, 19 October 2018 (UTC)[reply]
- Put ten through RM this week, then go back to RFC next week. The previous RFC was pretty sad in drawing attention, despite the attempt at publicising. —SmokeyJoe (talk) 08:14, 19 October 2018 (UTC)[reply]
- If anyone wants it, I will split based on those words in 2 other files as well. A bot could conceivably run through format2 and leave the first for human intervention as it appears few (if any) concerns were raised about that case directly. That said, I made this bot based on the original bot request above, which was simple to implement, and submitted it accordingly. If another RfC is desired to further solidify the consensus and address concerns raised, I am all for it and do not wish to rush anything. Pinging all active participants: Number 57, Xaosflux, MusikAnimal, Anomie, The Earwig, Headbomb, Cyberpower678, SmokeyJoe, BrownHairedGirl, SQL. (hope that's everyone, if I missed anyone I apologize). --TheSandDoctor Talk 05:40, 19 October 2018 (UTC)[reply]
- The RfC supposedly supporting these mass moves looks very very dubious on a quick glance. The closing statement is terribly inadequate. This looks a tad overenthusiastic. 35,226 page moves with nothing mentioned at WP:RM? --SmokeyJoe (talk) 05:14, 19 October 2018 (UTC)[reply]
- I will echo what Joe already said. That discussion is clearly insufficient for a change of this magnitude coupled with the bland closure. To avert unnecessary crisis I will suggest a new RFC on Village pump with detailed rationale and be well advertised . –Ammarpad (talk) 08:31, 19 October 2018 (UTC)[reply]
- I knew nothing about this proposed task until today when it was mentioned at WT:RM. Given the number of pages involved it would be best to advertise this in all the relevant places where editors of such articles would be watching and get further input. Thanks. — Frayæ (Talk/Spjall) 08:39, 19 October 2018 (UTC)[reply]
- Number 57 and others, for the record I have now changed my vote from oppose to support. However, I don't agree with my oppose vote being characterised in that way. I would generally advise against editors attempting to explain the reasons for other people's votes. Onetwothreeip (talk) 11:23, 19 October 2018 (UTC)[reply]
- Strong oppose. This proposal does not have an adequate consensus given the large number of pages concerned, long-standing titles, and the high profile nature of the pages. The proposal should be tested in a few RMs first to see if it really has consensus. — Amakuru (talk) 22:52, 19 October 2018 (UTC)[reply]
- Number 57, I'm sorry, but I feel a bit misled that you cite WP:NCGAL as justification for moving California Proposition 10 (1998) to 1998 California Proposition 10, when it was you who changed the guideline in response to the RFC four days ago. I don't see any discussion in the RFC that supports this unnatural wording. Perhaps I am being pedantic, but I think this is an important distinction because "California Proposition 10" is a proper noun that external sources use directly, while "1998 California Proposition 10" really isn't. If there was only one Proposition 10 in California, I see a strong argument for excluding the year (c.f. California Proposition 46, though there aren't many examples), further supporting that the year acts as disambiguation and not as part of the proper name. I'm open to discussing this point further, but I don't feel it's clear-cut enough for the bot. — Earwig talk 02:36, 20 October 2018 (UTC)[reply]
- Hi Earwig. I'm not really sure what the issue is here. The guideline previously stated that propositions should be of the title format "California Proposition 10, 1998" (the article itself is not named correctly according to the guideline by using parentheses). The RfC proposal was to move the year from the end to the start of the title, so California Proposition 10, 1998 would therefore become 1998 California Proposition 10. Number 57 10:40, 20 October 2018 (UTC)[reply]
- Number 57, understood, but when I look through the category series starting at Category:California ballot propositions, 1994, I see almost no titles using the previously guideline-supported format. This leads me to believe the guideline was never really correct or widely applied in this area (it's a guideline, after all), so I have trouble using the guideline now as the sole justification for moving these pages automatically. — Earwig talk 17:06, 20 October 2018 (UTC)[reply]
- Actually, let me move this discussion to the RfC; it's a better place. — Earwig talk 17:14, 20 October 2018 (UTC)[reply]
- Number 57, understood, but when I look through the category series starting at Category:California ballot propositions, 1994, I see almost no titles using the previously guideline-supported format. This leads me to believe the guideline was never really correct or widely applied in this area (it's a guideline, after all), so I have trouble using the guideline now as the sole justification for moving these pages automatically. — Earwig talk 17:06, 20 October 2018 (UTC)[reply]
- Hi Earwig. I'm not really sure what the issue is here. The guideline previously stated that propositions should be of the title format "California Proposition 10, 1998" (the article itself is not named correctly according to the guideline by using parentheses). The RfC proposal was to move the year from the end to the start of the title, so California Proposition 10, 1998 would therefore become 1998 California Proposition 10. Number 57 10:40, 20 October 2018 (UTC)[reply]
- @TheSandDoctor, SQL, Xaosflux, Cyberpower678, Amorymeltzer, The Earwig, Ammarpad, and Frayae: The original RfC has been reopened for further input: Wikipedia talk:Naming conventions (government and legislation)#Proposed change to election/referendum naming format. Number 57 15:08, 20 October 2018 (UTC)[reply]
- @Number 57: Thanks for the note, the BRFA process does include many components, and this page is best for going over the technical issues of the execution, but not the best for determining community consensus as the the overall scope of edits. Article titles haven proven to be touchy subjects in the past. I brought up some points about possible unintended consequences above to category SORTKEYS, feel free to include that in the RfC if you want a definitive answer. — xaosflux Talk 15:12, 20 October 2018 (UTC)[reply]
Post-RFC Closure
[edit]- @Number 57, SQL, Xaosflux, Cyberpower678, Amorymeltzer, The Earwig, and Ammarpad: The RfC was closed in support by GorillaWarfare. Would it be possible to get this BRFA back on track since the question of its need has now been addressed by a wider audience
and to {{archive top}}/bottom the concern sections with some sort of a note linking the RfC addressing the concerns raised? --TheSandDoctor Talk 03:40, 20 November 2018 (UTC); struck part of my comment 14:30, 20 November 2018 (UTC)[reply]- Well, that's now upto BAG members to approve. The concern then was inadequate consensus, now that's over. –Ammarpad (talk) 09:00, 20 November 2018 (UTC)[reply]
- Now that the RFC is closed, are there any outstanding technical concerns that would prevent us from moving forward with a trial? SQLQuery me! 04:56, 20 November 2018 (UTC)[reply]
- As part of the extended discussion, I produced a full list of article changes at the links below. Perhaps these might be of use for a bot to move the articles in the same way that Cydebot processes the category moves listed at Wikipedia:Categories for discussion/Working? One of the shorter lists could potentially be used for the trial
- Cheers, Number 57 10:47, 20 November 2018 (UTC)[reply]
- These lists were generated off of the original dataset in the excel document, right Number 57? --TheSandDoctor Talk 14:27, 20 November 2018 (UTC)[reply]
- Yes, although I've just removed a couple that I'm aware have been merged since (the Polish presidential ones from 1922). Number 57 15:02, 20 November 2018 (UTC)[reply]
- @Number 57: I was going to ask that it be reformatted, but in the process of going it myself automatically, I realized that that was a pointless extra step. Anyhow, new source code is written to work with the pages you linked above. Bot is ready to go so long as the format of User:Number 57/Elections/Propositions is maintained wherever it runs. (The format being "*wikilink to wikilink "). --TheSandDoctor Talk 21:16, 20 November 2018 (UTC)[reply]
- Excellent, thanks! Is it just a case of people approving it now? I don't really know how it works. In the meantime I might start moving articles that were created since the lists were created. Cheers, Number 57 21:18, 20 November 2018 (UTC)[reply]
- @Number 57: I think we wait a couple of days and if no one comments about the technical aspect of things or no technical issues are discovered we can move on to a trial. Once the trial is complete and if there weren't any issues, then it will probably be approved. If there are issues, those addressed and then a new trial. Do I have the gist of that right SQL? It's been a while since I've been at BRFA. --TheSandDoctor Talk 21:45, 20 November 2018 (UTC)[reply]
- Have we decided what the trial will be? Onetwothreeip (talk) 23:27, 20 November 2018 (UTC)[reply]
- @Onetwothreeip: Not yet. That is up to the BAGs. --TheSandDoctor Talk 23:42, 20 November 2018 (UTC)[reply]
- TheSandDoctor, Yep, you've got it exactly right. SQLQuery me! 23:50, 20 November 2018 (UTC)[reply]
- Have we decided what the trial will be? Onetwothreeip (talk) 23:27, 20 November 2018 (UTC)[reply]
- @Number 57: I think we wait a couple of days and if no one comments about the technical aspect of things or no technical issues are discovered we can move on to a trial. Once the trial is complete and if there weren't any issues, then it will probably be approved. If there are issues, those addressed and then a new trial. Do I have the gist of that right SQL? It's been a while since I've been at BRFA. --TheSandDoctor Talk 21:45, 20 November 2018 (UTC)[reply]
- Excellent, thanks! Is it just a case of people approving it now? I don't really know how it works. In the meantime I might start moving articles that were created since the lists were created. Cheers, Number 57 21:18, 20 November 2018 (UTC)[reply]
- @Number 57: I was going to ask that it be reformatted, but in the process of going it myself automatically, I realized that that was a pointless extra step. Anyhow, new source code is written to work with the pages you linked above. Bot is ready to go so long as the format of User:Number 57/Elections/Propositions is maintained wherever it runs. (The format being "*wikilink to wikilink "). --TheSandDoctor Talk 21:16, 20 November 2018 (UTC)[reply]
- @TheSandDoctor: Just a thought: Will the bot produce a report of all the articles it has failed to move (I guess there may be a small number of targets that are locked for whatever reason)? Cheers, Number 57 15:23, 21 November 2018 (UTC)[reply]
- Yes Number 57, they will all be logged in the output file, with failed moves clearly indicated, being prefixed with "CONVERTION FAILED ", followed by what the page name is and what it was attempted to be moved to. This can be seen on the second last line of election_converter2.py. --TheSandDoctor Talk 18:54, 21 November 2018 (UTC)[reply]
- @SQL and Number 57: Actually, instead of reading out of that list, when I do do a run I will probably put it as a fully protected subpage of my userspace or fully protect the page it is currently generating the list from if that is okay (and would undo the protection after it is done). That is just as an added precaution as we are talking about a lot of articles and anyone editing it maliciously while it is reading could be a concern/risk I don't really want to take. I would be open to template protection as well, I am just not comfortable with that risk of it being unprotected, however small. Of course, the bot status subpage will be the way to shut it down if need be and will remain EC protected. --TheSandDoctor Talk 22:27, 21 November 2018 (UTC)[reply]
- Yes Number 57, they will all be logged in the output file, with failed moves clearly indicated, being prefixed with "CONVERTION FAILED ", followed by what the page name is and what it was attempted to be moved to. This can be seen on the second last line of election_converter2.py. --TheSandDoctor Talk 18:54, 21 November 2018 (UTC)[reply]
- @TheSandDoctor: I can protect the lists too if you like (I'm also an admin), but that does sound sensible. Cheers, Number 57 22:33, 21 November 2018 (UTC)[reply]
- Haha, you're right. I knew you were an admin, my script told me you were an admin, yet when I sent that for some reason that didn't come to mind. I was just concerned with making sure you could edit it. That said, it still stands that I believe fully or template protecting is a smart move. The bot only reads the page in question containing the list of changes for a moment before storing it in memory and working from there, but if it was changed just before, then there could be potential issues. The probability is probably extremely small, but it is a risk as automated scripts don't think. So long as a malicious change kept the same format the bot would happily execute it unless it didn't have the rights to. --TheSandDoctor Talk 23:26, 21 November 2018 (UTC)[reply]
- @TheSandDoctor: I can protect the lists too if you like (I'm also an admin), but that does sound sensible. Cheers, Number 57 22:33, 21 November 2018 (UTC)[reply]
Trial run
[edit]- I don't see any other comments on the technical end, so Approved for trial. Please provide a link to the relevant contributions and/or diffs when the trial is complete. - 150 moves. Please make sure to clearly link to this BRFA when moving. SQLQuery me! 04:40, 22 November 2018 (UTC)[reply]
- @SQL: I granted the bot (temp) confirmed as it was triggering an edit filter (MediaWiki:Abusefilter-warning-badmove) resulting in the script failing. Trial underway. --TheSandDoctor Talk 05:22, 22 November 2018 (UTC)[reply]
- Trial complete. SQL. Issues were initially encountered, but they had to do strictly with the aforementioned edit filter and the API rate limits since the account is not currently flagged as a bot. In the event that this task is approved, I would have to request the bot flag so that page moves wouldn't fail every 5 or so, killing the script along with it. --TheSandDoctor Talk 05:46, 22 November 2018 (UTC)[reply]
- Looks good to me – I didn't spot any errors in the moves. I'm guessing there were no failures? Cheers, Number 57 09:55, 22 November 2018 (UTC)[reply]
- @Number 57: No failures other than the aforementioned external forces the script was fighting for not being confirmed (I could fix) and not having bot status. It is worth noting that it skipped California Proposition 64 (2016) since it is a redirect elsewhere at the moment, but otherwise did not fail. --TheSandDoctor Talk 14:02, 22 November 2018 (UTC)[reply]
- Looks good to me – I didn't spot any errors in the moves. I'm guessing there were no failures? Cheers, Number 57 09:55, 22 November 2018 (UTC)[reply]
- @SQL: Any chance this could be approved today (not sure if only one editor can approve)? It would be great if the run could happen over Friday night (UK time) as I can then spend the weekend sorting out the loose ends. Cheers, Number 57 12:41, 23 November 2018 (UTC)[reply]
- {{BAGAssistanceNeeded}} @Xaosflux, Cyberpower678, Amorymeltzer, The Earwig, and Ammarpad: Editors have started mass moving high profile articles (like 2016 United States presidential election), so it would be good if we could get the bot going as soon as possible to ensure the entire set of moves are done in the shortest window possible. As the trial went ok, would anyone be willing to approve the bot? Cheers, Number 57 23:08, 25 November 2018 (UTC)[reply]
- Given the delay in getting this approved, I'm wondering whether editors are worried about some kind of backlash if they're the person that approves the bot? Several high profile articles have been moved (all US Senate elections, several recent US presidential elections and all the UK general elections) and there has only been one editor react negatively, whilst a couple of editors who weren't involved in the discussion have noted that they support the change.[2][3] Now that pages have started to be moved manually, the longer the bot run doesn't happen for, the more likely it is that problems will crop up because of inconsistency in the format. Cheers, Number 57 14:51, 27 November 2018 (UTC)[reply]
- Number 57, I was mainly waiting for feedback one way or another. A lot of the delay can also be attributed to there being very few actually active approvals group members at this time. Typically, I prefer not to approve bots that I've trialed, but this has seen significant input from many other members of the approvals group. It appears that the bot is working properly, and that there is consensus for this task. Approved. SQLQuery me! 20:46, 27 November 2018 (UTC)[reply]
- Given the delay in getting this approved, I'm wondering whether editors are worried about some kind of backlash if they're the person that approves the bot? Several high profile articles have been moved (all US Senate elections, several recent US presidential elections and all the UK general elections) and there has only been one editor react negatively, whilst a couple of editors who weren't involved in the discussion have noted that they support the change.[2][3] Now that pages have started to be moved manually, the longer the bot run doesn't happen for, the more likely it is that problems will crop up because of inconsistency in the format. Cheers, Number 57 14:51, 27 November 2018 (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.