Template talk:Infobox religious text

Wikidata

We should go ahead with adding wikidata params, along with options to enable editors to suppress them. This will enable improvement of both wikidata and wikipedia by synchronizing the information and reducing edit-wars on wikipedia, as controversial params can be fetched from wikidata, which do not support un-linked content for most fields. Ms Sarah Welch Frietjes -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 13:18, 27 April 2017 (UTC)

• Oppose auto-piping of image or any other field from wikidata to wikipedia article. Same reasons as above [see extended discussion below]. Would support if the specific image/field has an attached [WP:RS] reference. Ms Sarah Welch (talk) 15:33, 6 May 2017 (UTC)
• Support, good idea. Wikipedia:Requests for comment/Wikidata Phase 2 permits us to modify infoboxes to include information from Wikidata, so there is already consensus for these changes. Certainly, concerns over sources are not unfounded, but unsourced statements can be blocked like the Infobox person version does. @Capankajsmilyo: see South Pole Telescope for an excellent example. Laurdecl ^talk 01:46, 13 May 2017 (UTC)
• Strongly Oppose Until Wikidata has better enforcement of a verifibility an source policy, and until only sourced items will be used, and until a source reference is invariably included and displayed as part of the item on Wikipedia. DES ^(talk)_{DESiegel Contribs} 02:31, 25 June 2017 (UTC)
• Oppose per Wikidata's almost complete lack of policies and community-lack-of-interest in policies, per confusing the hell out of users when they click edit on a page and there's exactly ZERO indication in the wikitext where something is coming from or how to fix it, per auto-import of wikidata values bypassing page protections, per millions of unsourced&Wikipedia-sourced items I have found bypassing the "only-sourced" filter because they have circular or junk references, and assorted other intractable issues with auto-importing wikidata. Alsee (talk) 05:34, 6 November 2017 (UTC)
• Support; only 77 transclusions of the infobox, so it would be easy to handle changes. Special:RecentChangesLinked does a good job of this (in spite of the lag). I set the parameters in the link to only show Wikidata edits and it works without changing default user preferences. Module:WikidataIB can filter out unsourced statements (I've never used it myself so I won't deal with the specifics of showing references etc.). It may help to wait until after the next RfC on Wikidata to implement this. Jc86035 (talk) 06:08, 6 November 2017 (UTC)

  Jc86035, Module:WikidataIB can filter out statements that don't claim a reference, or which have a reference mentioning Wikipedia. However it can't actually filter out unsourced statements or Wikipedia sourced statements. I've found there are millions of items effectively bypassing that filter. It has been standard practice at Wikidata to run bots creating circular references en-mass within wikidata. A bot will find an unsourced statement such as "[Country] has capital [city]", and the bot logically copies that to "[City] is capital of [Country]". The fun bit is that the bot then adds a reference to the second copy of that unsourced information. The bot adds a ref saying where the information was sourced from. The reference is "Stated in: [Wikidata item for Country]". It is impossible to filter those circular refs without massively nuking wikidata imports in general. The software can't tell the difference between a circular ref and a non-circular ref. There are also massive numbers of items sourced from Wikipedia, where the reference does not contain the word "Wikipedia". They don't care about sourcing, and when they do add a ref it's often a crap ref. Alsee (talk) 07:47, 6 November 2017 (UTC)

  @Alsee: I think it's probably possible (by filtering out valid stuff from the constraint violations pages), but I guess no one's dealt with it yet if this is a problem because there's so much stuff. Is this really a big concern for just 77 pages? Wikidata infoboxes are enabled on far more, and it's not that hard to review 77 Wikidata items. Jc86035 (talk) 07:53, 6 November 2017 (UTC)

  Jc86035 I believe it would be intractable trying to set up a constraint distinguishing between presumably-citable-Q#s and presumably-uncitable-Q#s. However the problem is fundamentally unsolvable. No constraint can distinguish between Stated_in: Q(NewYorkTimes) and Stated_in: Q(NewYorkTimes), where one means published in New York Times, and the other means copied from Wikidata-Q(NewYorkTimes). The two refs are digitally-identical. Alsee (talk) 08:18, 6 November 2017 (UTC)

• Support, we're only talking about 77 articles here, that's only a small amount, and if the editors that are working on this topic want to use Wikidata then why should we stop them from doing so? I'm happy to help with any issues that arise due to the inclusion of Wikidata information. Thanks. Mike Peel (talk) 22:27, 6 November 2017 (UTC)
  • Expanding on my oppose rationale above: We're only talking about 77 articles here, that's only a small amount, and if the editors that are working on this topic don't want Wikidata auto-imported then why should we force it on them? I'm happy to help with any issues that arise due to not auto-importing Wikidata information. Thanks. Alsee (talk) 08:00, 8 November 2017 (UTC)
• Oppose. We have just had an RfC on a Wikidata-based infobox, which ended with it being reverted to a non-Wikidata version. We now have a suggested ArCom motion to start a new RFC on the use of Wikidata on enwiki. Please wait until after that RfC. Fram (talk) 10:13, 7 November 2017 (UTC)

Image (separate matter)

We can continue the discussion about other params above, this one is different. Should we include image from Wikidata, while allowing the |image=, if editors want a different one. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 19:21, 27 April 2017 (UTC)

Inviting comments from RexxS, Frietjes. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 02:40, 6 May 2017 (UTC)

Oppose auto-piping of image or any other field from wikidata to wikipedia article. Same reasons as above. Would support if the image/field has an attached reference. Ms Sarah Welch (talk) 15:33, 6 May 2017 (UTC)

Support if image field has a reference. Better than nothing. Laurdecl ^talk 01:46, 13 May 2017 (UTC)

Comment If the above discussion is not the particular focus of the RFC, can we please have the template moved so that people (like myself) who may have an idea but not understand the jargon of coding can actually respond to the question at hand rather than a general meandering discussion? Especially seeing as how both the RfC seems to be about whether to auto-pipe images from WikiData to Wikipedia, not the template coding itself, and that the question is posed in THIS sub-section, not the main "thread" as it were. Anyway, cheers, Doctor Crazy in Room 102 of The Mental Asylum 07:03, 15 May 2017 (UTC)

@Drcrazy102: I have moved the discussion below. The main topic here is whether data should be piped from Wikidata (section above). The image is slightly related but separate. Laurdecl ^talk 07:23, 15 May 2017 (UTC)

Thanks for that, it's just not the norm when responding to an RfC request to not see the actual question-statement starting the RfC. Anyway, thank you @Laurdecl: for clarifying. Cheers, Doctor Crazy in Room 102 of The Mental Asylum 07:46, 15 May 2017 (UTC)

Oppose with caution. As mentioned somewhere in the extended discussion below, WikiData does not hold itself to the same standards of verifiability as Wikipedia and it's editors do. If the image is reliable sourced AND able to be used under the various licenses, then there isn't an issue at all. However, there are far too many checks and balances at the moment for what currently sounds like an automated "data snatch-n-grab" from WikiData. Cheers, Doctor Crazy in Room 102 of The Mental Asylum 07:46, 15 May 2017 (UTC)

So you support if the image has a reference? Wikidata images are automatically used in many other infoboxes already. Laurdecl ^talk 07:50, 15 May 2017 (UTC)

Oppose per above. Alsee (talk) 05:34, 6 November 2017 (UTC)

Comment @Ms Sarah Welch, Laurdecl, Drcrazy102, and Alsee: To clarify, do you actually want image (P18) values – pictures from Commons – to have a secondary source? How does that work? Has anyone ever stuck {{citation needed}} on images on enwiki? How would we use images that users have taken themselves? Would the reference have to be a secondary source which has used the image (and wouldn't that tend to be circular referencing)? I don't think this is a very reasonable thing to require. Jc86035 (talk) 06:17, 6 November 2017 (UTC)

Jc86035 I apologize if my casual "per above" appeared to include refs. I pretty much meant every wikidata problem except refs. For example: Any user who is new or not-versed-in-wikidata will be confused-as-hell when they click to edit the article and there is ZERO indication in the wikitext where the heck a picture is coming from, or how to change it. Especially if the image is vandalism. Alsee (talk) 07:57, 6 November 2017 (UTC)

I have removed user-taken images when they have nothing to do with the subject, are amazing OR or a gross violation of MOS:IMAGES. We have zillions of completely misidentified images in commons. As such wikipedia reputation for accuracy is suspect, we shouldn't want to create auto-piping to make things worse on a widespread scale. Images are no different than text. Plus religious text / religious topics can be controversial or sensitive subjects. Let us be careful. If you really want the image pipe option, then instead of default auto true, make default false. Create a parameter "import_wikidata image=" and let the template user specify yes to enable that image to be displayed. Is that possible? Ms Sarah Welch (talk) 13:25, 6 November 2017 (UTC)

• Oppose until it is explained convincingly why we should do this. It is trivial to have an image in the infobox now. Technically, we can outsource this to Wikidata, where one image is chosen by all Wikidata editors to serve all wikis (wikipedias and other sources that use Wikidata). There is no reason why all these sites should have or prefer the same image, this is not factual information but an illustration. It doesn't help enwiki or Wikidata one bit if we take the image from there, but may cause problems (with e.g. the image being changed on Wikidata without anyone here noticing it, or with cross-wiki edit wars at worst). I see no benefit at all from this proposal. Fram (talk) 08:12, 8 November 2017 (UTC)

Extended discussion

Moved from #Wikidata

I disagree. The problem is that wikidata has yet to develop a means to establish WP:V, WP:RS and such guidelines. The data there is choke full of vandalism and much unsourced nonsense. Linking wikipedia's infoboxes to wikidata is a way to propagate misinformation. My concern with Capankajsmilyo is that even in articles he actively works on (Jainism), there has been and continues to be a disrespect for verifiability, reliable sources and other pillars of community agreed content policies. Now linking wikidata is extending the same. Not helpful. Ms Sarah Welch (talk) 13:24, 27 April 2017 (UTC)

Please note, that wikidata is optional param, enabled by not adding a particular param. It can be disabled by simply adding the particular param (with or without value). -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 13:27, 27 April 2017 (UTC)

It doesn't matter if it is optional. An optional vandalism/non-WP:V information is as bad as mandatory vandalism/non-WP:V information. The issue is content policies and guidelines. Ms Sarah Welch (talk) 13:31, 27 April 2017 (UTC)

What about the parameters which have "exactly same" values as per WP:V of English Wikipedia? -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 13:35, 27 April 2017 (UTC)

@Capankajsmilyo: I don't understand your hypothetical question, nor your evasion of the WP:V and WP:RS issue. Is there a way to require verifiability and reliable sources in wikidata? Ms Sarah Welch (talk) 13:48, 27 April 2017 (UTC)

I don't understand your dodging the question. By the way, wikidata does talk about verifiability. See this. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 13:55, 27 April 2017 (UTC)

@Capankajsmilyo: I didn't dodge your question. I don't understand what you are trying to state. To the best of my knowledge, there is no WP:V policy in wikidata. Please link the wikidata policy page on verifiability and reliable sources. The link you give merely states that "Wikidata can record not just statements, but also their sources, thus reflecting the diversity of knowledge available and supporting the notion of verifiability". That does not mean that the field data must be verifiable, leave alone they be from reliable sources. The wikidata project has yet to develop the processes we have in wikipedia for dispute resolution, mediation, 3RR, ANI, ARCA, etc. Wikipedia admins have no say over wikidata project. Automatically piping all original research and vandalism from still unstructured and unmonitored wikidata into wikipedia articles and infoboxes, is a bad idea. @RexxS: you have had some experiences with wikidata. Your experiences and guidance will be most appreciated. Ms Sarah Welch (talk) 14:10, 27 April 2017 (UTC)

@Ms Sarah Welch: Wikidata's proposed verifiability policy is at d:Wikidata:Verifiability. It is certainly comparable to WP:V, but suffers from a lack of acceptance and enforcement. The solution I've been working on over the last year is in Module:WikidataIB where I've arranged that any Wikidata content that is unsourced or only sourced to one of the Wikipedias is filtered out by default.

It's also a sub-optimal idea to rely on having to add a blank parameter if a Wikidata-enabled field is to be suppressed. Who is going to do the initial work of suppressing every unwanted field in all 62 articles using this template? How does that scale up if somebody wants to do the same for all 39,498 transclusions of {{Infobox book}}? Where do you draw the line? Also, Wiki-gnomes regularly remove blank parameters from infoboxes. How would we protect the existence of a blank parameter which is being used to suppress an unwanted field? Do we have to have a hidden comment each time and hope the gnomes notice it?

I'd strongly recommend that Capankajsmilyo studies the RfCs at Wikipedia:Requests for comment/Wikidata Phase 2 and Wikipedia:Village pump (policy)/Archive 128 #RfC: Wikidata in infoboxes, opt-in or opt-out? and tries to understand the opt-in/opt-out issues. There are solutions to each of the problems and the documentation at Module:WikidataIB/doc may help explain some possibilities, but we're a long way from gaining automatic acceptance of Wikidata-enabled infoboxes from the editing community, who rightly express concerns about sourcing Wikipedia content externally without having robust controls in place. --RexxS (talk) 16:28, 27 April 2017 (UTC)

Thanks, RexxS, for all the gems in your reply. I hope Capankajsmilyo and others active in infoboxes space will pay heed. I am also delighted to learn that d:W-V policy exists. I have been helping out with 100s of wikidata pages, but was getting tired of reverting outright vandalism, strange OR and absurd data particularly in the case of Buddhism-related pages. Now, d:W-V gives me a basis to continue helping there. I concur that there are potential and possibilities, but we are indeed a long way from the robust controls at this point. Ms Sarah Welch (talk) 16:41, 27 April 2017 (UTC)

Thanks RexxS, they were indeed good discussions. First one resolved a broad consensus on option 4. The result of second one however was not much clear. I would like to point a few things here.

1) These discussions are 1 year old, first one being much more older. Wikidata and wikipedia both has grown since then. 2) The Template you are working on is a standalone infobox. Is it possible to integrate it in the existing infobox? The documentation might need better explanations and examples on this aspect. 3) I was amused with the idea of global watch-list and would like to see it happen very much. 4) I think we should re-initiate the discussion to know where we stand now. This time I would propose the focus to be on specific info-boxes and specific parameters. Whether they can be fetched from wikidata or not. For example, "children" of "Infobox deity" or "genre" of "infobox religious text". I guess, a discussion on specific fields will help, and a hybrid solution will be much better than total "this" or "that". -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 17:24, 27 April 2017 (UTC)

I rechecked the template and tested it in preview of Rama. I used {{infobox person/Wikidata | fetchwikidata=ALL}} and it fetched only the image and dynasty. That's too less of what is available and verifiable info. Children, spouse, parents, siblings? Did I do something wrong? -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 17:39, 27 April 2017 (UTC)

@Capankajsmilyo: First one did not agree on option 4 without qualification... it reads, "There is sufficient support for option 3 however, to indicate that this modification should be done carefully and deliberately, at least at first." Option 3 stated, "Wikidata would be integrated in a very limited way, only with local editorial consensus, and can be reversed easily on that specific article." You do not have editorial consensus for all Indian religions space articles at least, and have objections to this as this discussion already shows. So, please do not pipe in wikidata, for now, either as option or default. I request you fix the Jainism articles where you have been very active in, and which heavily suffer from the lack of verifiability or from self published sources / unreliable sources (as you know, I have repeatedly raised this concern on Talk:Mahavira etc for example). By piping more unsourced information and OR from wikidata (such as on Rama etc), you are making things worse. Ms Sarah Welch (talk) 18:01, 27 April 2017 (UTC)

I haven't piped in anything. Please read again, I wrote in "Preview". Check edit history. I haven't edited Rama, since this discussion started. Please take a look before accusing. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 18:15, 27 April 2017 (UTC)

Someone did pipe in wikidata for Krishna. We discussed it yesterday and Frietjes helped fix. The issue ultimately is not who did it, the issue is that this auto-piping of wikidata is an automation of unsourced information, POV-y OR and in some cases outright vandalism, that affects 100s of wikipedia articles. See RexxS's comments above. I have looked at dozens of religions-related wikidata pages, and the field data for Christianity, Islam, Buddhism, Hinduism, Jainism, etc is unsourced/wrong and such that many fields are not worth piping into wikipedia articles. Ms Sarah Welch (talk) 18:43, 27 April 2017 (UTC)

@Capankajsmilyo: You didn't do anything wrong. Remember that {{infobox person/Wikidata}} is only in beta testing, and not all fields are necessarily implemented yet. However, you can always check what is available by previewing {{infobox person/Wikidata | fetchwikidata=ALL |onlysourced=false}} in Rama, which turns off the "unsourced" filter. You'll then see Born, Occupation, Spouse, Children and Family - all of which are unreferenced on Wikidata at Ram (Q160213). The first job would be to find reliable sources for each statement in that Wikidata entry and add them. Otherwise you'll never get agreement to fetching those values from Wikidata. Once they are sourced, you'll see them appear when you use {{infobox person/Wikidata}}. Incidentally, if you want to see all of the Wikidata items and the state of the referencing, just paste {{#invoke:Sandbox/RexxS/WdRefs|seeRefs}} into any article and preview it (please don't save!). That will give you a table showing which statements are available, but also which are unreferenced. There's still a lot of work to do on Wikidata. Cheers --RexxS (talk) 18:49, 27 April 2017 (UTC)

@Ms Sarah Welch: Please don't use words like "By piping more unsourced information and OR from wikidata (such as on Rama etc), you are making things worse", if "The issue ultimately is not who did it". Look for contribution history before such accusations. The template deity has also been fixed, and will see what I can do after you have made a sweeping revert on infobox epic character. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 18:55, 27 April 2017 (UTC)

Please stop playing this "don't blame me, check contribution history" game. It was you who added piped wikidata for consorts into wikipedia on April 24, that messed up Krishna article and wherever that template is used. Let us get our focus back on the issue here. Ms Sarah Welch (talk) 19:20, 27 April 2017 (UTC)

Thanks for the clarity RexxS, I can see that now. I have also made a test case at Template:Infobox deity/Wikidata, but the fields sibling and child doesn't seem to work. Can you please have a look? I would further suggest, that we select certain test pages and implement Template:Infobox deity/Wikidata once it is complete. Clarifying again, "only on certain test pages", which could be selected by the community of WP editors after a consensus. Coming to the reference thing, what kind of reference are we looking at here? Does imported from English Wikipedia work? Or does it need to be a book or external link. The wikipedia article has ample references on it, is there a way we can use them? -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 19:06, 27 April 2017 (UTC)

It should not be wikipedia, nor blogs, nor other questionable sources. Just be the same quality external source as what would be WP:RS in wikipedia, if the long term vision is to integrate the sister projects. It would be helpful if the wikidata fields visibly confirmed which fields include a link or source details. Ms Sarah Welch (talk) 19:15, 27 April 2017 (UTC)

They do. I just added ref in sibling->Lakshmana of Rama. See on wikidata. -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 19:41, 27 April 2017 (UTC)

Thanks. For those wikidata fields where there is a linked reference, I would not mind if the wikidata gets piped when the wikipedia infobox field is empty. This feature should stop piping if the attached reference with the wikidata field is later deleted after someone did a verifiability check. Checking such functionality on a few test articles is a good idea too (perhaps with a note posted on that article's talk page that this article's integration with wikidata is being tested). These suggestions are subject to getting consensus of the community for all this, Ms Sarah Welch (talk) 19:55, 27 April 2017 (UTC)

Module:WikidataIB and the new testcase based on that Template:Infobox deity/Wikidata is capable of doing what you are asking for. Further, I would like to know the suggested test pages, on which it can be tested. Also, is there a way, we can invite more comments? -- Pankaj Jain Capankajsmilyo (talk · contribs · count) 05:46, 29 April 2017 (UTC)

One final tip: see https://www.youtube.com/watch?v=jP-qJIkjPf0 for a visual tutorial on how to "Drag'n'drop Wikipedia references to Wikidata". You enable it by being logged on to Wikidata → go to Preferences → Gadgets → tick the box for "Drag'n'drop" → click the Save button at the bottom of the page. Enjoy! --RexxS (talk) 22:12, 27 April 2017 (UTC)

Amazing! Thanks for that tip RexxS, Ms Sarah Welch (talk) 00:44, 28 April 2017 (UTC)