Wikipedia:Bots/Requests for approval/BattyBot 47

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.

BattyBot 47

Operator: GoingBatty (talk · contribs · SUL · edit count · logs · page moves · block log · rights log · ANI search)

Time filed: 23:18, Wednesday, September 9, 2015 (UTC)

Automatic, Supervised, or Manual: Automatic

Programming language(s): AutoWikiBrowser

Source code available: AWB + ~~User:BattyBot/Persondata~~ User:BattyBot/Blank Persondata as custom module

Function overview: Remove deprecated {{Persondata}} template ~~ONLY when its values are found elsewhere in the article~~ when the only value is in the |NAME= parameter

Links to relevant discussions (where appropriate): Wikipedia:Village pump (proposals)#Removing Persondata

Edit period(s): One time run, and then periodically as the custom module is improved

Estimated number of pages affected: Thousands

Exclusion compliant (Yes/No): Yes

Already has a bot flag (Yes/No): Yes

Function details: Although {{Persondata}} has been deprecated, some editors are concerned about deleting those templates that may contain valuable data. There are other editors concerned that we have not yet started an automated process to delete the templates.

This bot request is intended as a compromise to ONLY delete those templates where ALL of the {{Persondata}} values are found elsewhere in the article, such as an infobox, the lead, and/or categories. If the module can not completely remove all values from {{Persondata}}, it will not make any edits. If the module can completely remove {{Persondata}}, AWB will also perform the usual general fixes.

Eliminating the simple cases should make it easier for editors to find those articles where {{Persondata}} values may be valuable and improve them before deleting {{Persondata}} (e.g. add categories, update the infobox, resolve discrepancies between different birth/death dates/places). Wikidata editors will still be able to populate Wikidata fields based on the infobox, the lead, and/or categories. GoingBatty (talk) 23:18, 9 September 2015 (UTC)[reply]

This bot request has been changed to be a small first step to ONLY delete those templates where the ONLY value in {{Persondata}} is in the |NAME= parameter. This bot would run in mainspace and the Draft namespace. If there are values in any other parameter, it will not make any edits. If the module does remove {{Persondata}}, AWB will also perform the usual general fixes. GoingBatty (talk) 01:47, 10 September 2015 (UTC)[reply]

Discussion

@GoingBatty: I appreciate your effort to move the ball forward, but the issue is not whether data found in Persondata is found in the infobox, lead or elsewhere within the existing Wikipedia articles. This rather misses the point. The issue is whether all usable Persondata information has been transferred to Wikidata. Before pushing forward with this effort, I would urge you to coordinate with BU Rob13, who is drafting a bot proposal that will remove Persondata templates that do not contain potentially useful information the idea being to focus efforts to transfer remaining accurate information to Wikidata. Please consider. Dirtlawyer1 (talk) 23:37, 9 September 2015 (UTC)[reply]

@Dirtlawyer1: I'm not actually working on any bot proposal. I was just brainstorming ideas about how a potential bot would work. This is a bit of a firestorm at the moment. At the very least, it is uncontroversial to remove all Persondata templates that are completely blank except for the name. I sampled 1,000 articles to check the frequency of such a use of the template, and that suggested that it would remove between 2,500 and 7,500 transclusions using that criteria alone, which is a non-trivial amount. I highly recommend starting there, where the controversy is lowest, and moving outward. At the very least, a proposal like yours should not look for information in the lead. Transferring data from complete sentences that are organized in no particular way into the fields of WikiData is significantly harder than from the fields of Persondata to the fields of WikiData. Infoboxes are more plausible, but you'd probably need to query those with experience moving data from enwiki to WikiData to see whether that's as easy as using Persondata fields. ~ Rob^Talk 23:53, 9 September 2015 (UTC)[reply]

Rob, removing empty Persondata templates is a sensible place to start. As I understand it, most of the empty ones were added by bot in the first instance, or by newbie editors using the article creation wizard. Dirtlawyer1 (talk) 00:00, 10 September 2015 (UTC)[reply]

@Dirtlawyer1 and BU Rob13: If someone gets approval to run a bot to transfer usable Persondata information to Wikidata, I will happily withdraw my proposal. My proposal should not prevent the continuing important work being done by you and others to manually populate Wikidata based on Wikipedia data. I would be happy to start by removing all Persondata templates that are completely blank except for the name. The only data my custom module is currently checking in the lead is the birth and death dates, which is normally organized in a particular way. GoingBatty (talk) 00:10, 10 September 2015 (UTC)[reply]

@GoingBatty, BU Rob13, and C678: "I would be happy to start by removing all Persondata templates that are completely blank except for the name." To be crystal clear: that is exclusive of alternate names, which are probably the single largest pool of usable Persondata information that should be transferred to Wikidata. GB, that is completely unobjectionable, and I see no reason why that should not be implemented immediately by bot action, in light of previous discussions. No one is arguing to keep Persondata indefinitely, or to unwind the existing consensus regarding deprecation. We are arguing about what remain is useful and usable, and how it should be transferred. At current count, there are about 1.2 million transclusions of Template:Persondata; if we could eliminate several hundred thousand of them without controversy, I think that would be advancing the ball, and would permit the larger community's rank and file to better focus on valid data that remains to be transferred to Wikidata. Dirtlawyer1 (talk) 01:10, 10 September 2015 (UTC)[reply]

Comment. Please note there is already an open BRFA regarding this task and is on hold pending the outcome of the discussion at the Village Pump.—^cyberpower_Chat:Online 00:05, 10 September 2015 (UTC)[reply]

@C678: Wikipedia:Bots/Requests for approval/Yobot 24 appears to be a request to delete ALL {{Persondata}} templates. This request is a intended to be a more surgical approach to ensure that there is no data loss, and includes a link to the source code for transparency. GoingBatty (talk) 00:17, 10 September 2015 (UTC)[reply]

I also think that initiating a BRFA before a decision is met is premature.—^cyberpower_Chat:Online 00:19, 10 September 2015 (UTC)[reply]

@C678: I am optimistic that submitting this BRFA with the source code will help the community reach a decision that has been eluding us for months. GoingBatty (talk) 00:30, 10 September 2015 (UTC)[reply]

The sounds to me like you're trying to force the solution through by saying "here's the code, now deal with it." We shouldn't launch BRFAs to push for a solution, but rather as a result of a decided solution.—^cyberpower_Chat:Online 01:05, 10 September 2015 (UTC)[reply]

@C678: See my comment above: removing all templates that are empty except for the primary name field is completely unobjectionable and consistent with prior discussions regarding deprecation. Dirtlawyer1 (talk) 01:13, 10 September 2015 (UTC)[reply]

Fair enough. Is this what this bot will do? I get the impression it will delete non-blank templates if the information can be found elsewhere. I might be thinking too much into this though. I'm trying to stay impartial to this so I can provide objective comments and opinions.—^cyberpower_Chat:Online 01:16, 10 September 2015 (UTC)[reply]

@C678 and Dirtlawyer1: I was thinking more like "here's the code that I suggest, will this help to make everyone happy?" I have reduced the bot scope (and provided a new custom module) per your suggestions. If you would like to leave this request open for a few days to give others a chance to weigh in on the scope, that's fine with me. If this small scope is approved, I would be happy to file a separate BRFA in the future with a larger scope in the hopes we can "eliminate several hundred thousand of them without controversy" upon request. Thanks! GoingBatty (talk) 01:47, 10 September 2015 (UTC)[reply]

@GoingBatty and C678: I have no objection to proceeding immediately, and I think that can be easily justified based on the May RfC and the subsequent BRFA discussions. No one is objecting to the deletion of Persondata templates in principle, only throwing out the baby with the bathwater, i.e., deleting remaining Persondata such as alternate names which has not been transferred by previous bot action. Frankly, I am as curious to see the numbers produced by your modified proposal above. CyberPower, what say you? Dirtlawyer1 (talk) 01:57, 10 September 2015 (UTC)[reply]

Sure. I have no issues with it. It seems uncontroversial enough. Unfortunately, I'm no BAGer, so I can't approve for trial.—^cyberpower_{Chat:Limited Access} 02:02, 10 September 2015 (UTC)[reply]

Well, how do we make you one? Alakzi (talk) 02:04, 10 September 2015 (UTC)[reply]

I would either go and self-nom or have someone nom me, at WT:BAG by following the instructions in the first section. Since I already applied for it twice and failed, I prefer to be nominated at this point.—^cyberpower_{Chat:Limited Access} 02:56, 10 September 2015 (UTC)[reply]

@Dirtlawyer1: When I see your post about throwing the baby out with the bathwater, it makes me think I did not do a good job at making my initial proposal clear enough. To continue the analogy, if someone discovered I emptied one bathtub (e.g. Persondata), they would be happy to see the baby (e.g. alternate names) is still safe and sound splashing around in the other bathtub (e.g. the infobox's |birth_name= or |other_names=), ready to be transferred manually or future bot action. GoingBatty (talk) 02:23, 10 September 2015 (UTC)[reply]

Approved for trial (100 edits). Please provide a link to the relevant contributions and/or diffs when the trial is complete. — Earwig ^talk 02:10, 10 September 2015 (UTC)[reply]

@The Earwig:

Trial complete. - please see these 100 edits. Thanks for your support! GoingBatty (talk) 02:55, 10 September 2015 (UTC)[reply]

Just went through all 100 edits. It looks to be performing exactly as GoingBatty promised. I say turn it loose. Dirtlawyer1 (talk) 03:31, 10 September 2015 (UTC)[reply]

No. We must perform 3 more extended trials each seperated by a week. We can't have BRFAs running through the process so fast. :p Seriously though, no issues here.—^cyberpower_{Chat:Limited Access} 03:39, 10 September 2015 (UTC)[reply]

I reviewed all one hundred and most of it looks good. I'm quite happy about this method of reducing the persondata backlog. One question: why is the bot adding this DEFAULTSORT? That doesn't look right. Here is another example. I know it's part of genfixes, but I thought genfixes handled that case correctly. And regarding future more aggressive removals, what should the procedure be? — Earwig ^talk 03:46, 10 September 2015 (UTC)[reply]

First, thanks GoingBatty for being receptive to reducing the scope. A lot of people on both sides of this issue have dug in their heels, and it's refreshing to see a little more willingness to compromise. I would recommend more aggressive removals involve a BRFA to allow the BAG to ensure consensus is there (which has been somewhat tricky in this area). If consensus has been reached and there are no serious objections, a speedy approval makes sense. It's a tiny bit tedious to file BRFAs, but I'm not fully comfortable with only a single person judging consensus in this area no matter how good their judgement is because many people (myself included, if I'm being honest) do not fully understand the technical details of this conversion and WikiData as a whole. ~ Rob^Talk 04:05, 10 September 2015 (UTC)[reply]

Tracked in Phabricator
Task T112227

@The Earwig and BU Rob13: I have opened an AWB bug report for the {{DEFAULTSORT}} issue, and will install any update that the AWB developers release. I'm open to someone else proposing other unobjectionable increase to the scope for this request. However, if this request is approved, I agree that any more aggressive removals (such as the original scope above) should come through a new BRFA. Thanks everyone! GoingBatty (talk) 02:40, 11 September 2015 (UTC)[reply]

@GoingBatty: So, what's the status here? Are we waiting on the AWB fix for defaultsort, or are you ready to move forward? Dirtlawyer1 (talk) 02:51, 11 September 2015 (UTC)[reply]

@Dirtlawyer1: I'm not opposed to waiting for the AWB developers to respond. They generally respond to bug reports very quickly. GoingBatty (talk) 03:06, 11 September 2015 (UTC)[reply]

Your call, sir. I'd like to see this initial phase of non-controversial bot removal go forward ASAP for a variety of reasons, not least because I want to see the statistics it produces. One commonality I noticed about the 100-article test run was that virtually all of the Persondata templates had either been added by bot action or by an editor using AWB, and none of them had ever been edited or updated. I am rather curious to see what percentage of the current 1.2 million instances of the template are virtually empty. Dirtlawyer1 (talk) 03:13, 11 September 2015 (UTC)[reply]

This is technically sound (aside from the defaultsort issue) and has consensus. I agree that it would be nice to get this done soon so we can judge how much persondata will be trivial to remove and how much will require more care. Once we get the defaultsort out of the way, I have no problem approving this, leaving more aggressive removal to future BRFAs. — Earwig ^talk 03:24, 11 September 2015 (UTC)[reply]

@The Earwig, GoingBatty, BU Rob13, C678, and Alakzi: If the 100-article test run is representative (100 articles in 20 minutes), we can extrapolate to 300 articles in an hour, or 7,200 per day. If that is a reasonably accurate estimate of the bot's progress, it would take two weeks to remove Persondata from 100,000 articles. That's fairly time-consuming, and it affords plenty of time for contemporaneous transfer activities. If we're really serious about transferring still-usable information from Persondata to Wikidata, I think we need to contemplate using a second bot to notify all active editors about what is being done, and to provide an introduction to Wikidata and instructions for the manual transfer of usable information from Persondata to Wikidata. I strongly urged that such a plan should be formulated in early June, and the idea was largely ignored by the most ardent proponents of immediate deletion. Dirtlawyer1 (talk) 04:01, 11 September 2015 (UTC)[reply]

Bots are never as fast as we'd like them to be. :p—^cyberpower_Chat:Offline 06:17, 11 September 2015 (UTC)[reply]

Your proposal was not ignored; it was dismissed as unworkable; and as contrary to the consensus at the RfC which preceded it. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:44, 11 September 2015 (UTC)[reply]

T112227 has been closed as resolved. Given that's the only issue:

Approved. As discussed, approval is for this version of the task only and future modifications will require a new BRFA. Thanks! — Earwig ^talk 18:12, 11 September 2015 (UTC)[reply]

Supportthe proposed removals, without prejudice against a further bot operation to remove all instances of the deprecated Persondata template. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 08:37, 10 September 2015 (UTC)[reply]

@Pigsonthewing: Thanks for your support, but since there are those who disagree with the wholesale instant removal, I'm trying to find a happy medium. Guess I overshot this time, but we'll keep working on it. GoingBatty (talk) 02:40, 11 September 2015 (UTC)[reply]

They may disagree - and do so loudly - but they keep ignoring both the fact that an RfC found consensus to remove Persondata; and the points made in the discussion that led to that consensus both here and on Wikidata (witness the bizarre claim that the voices of those commenting in that discussion have been "absent"). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:40, 11 September 2015 (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.