Template talk:Infobox video game

From Wikipedia, the free encyclopedia
  (Redirected from Template talk:Infobox VG)
Jump to: navigation, search
WikiProject Video games (Rated Template-class)
WikiProject icon This template is within the scope of WikiProject Video games, a collaborative effort to improve the coverage of video games on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
 Template  This template does not require a rating on the project's quality scale.
 

Wikidata[edit]

Does anyone know the status of Wikidata integration with infoboxes? It would be nice to leave most of the infobox data there, or to set an image once across all (participating) Wikipedias when a free use version is available... czar 13:59, 12 April 2016 (UTC)

Official status: Approved for use. Unofficial status: Inclusion of Wikidata in infoboxes is being only-slowly adopted. Template:Infobox telescope is the only one on en.WP using it of which I know. --Izno (talk) 14:03, 12 April 2016 (UTC)
There are a few more listed at Category:Templates using data from Wikidata. Is the idea that each infobox group starts experimenting on its own? Is anyone opposed? czar 14:44, 12 April 2016 (UTC)
Basically, I think that's what is happening. I wouldn't be opposed to experimentation here--VGs are pretty easy to handle. --Izno (talk) 15:25, 12 April 2016 (UTC)
  • @Ferret, ready for your next project? czar 02:30, 26 April 2016 (UTC)
  • Sure, I can tackle this when we're done with the review wikidata ideas. However this will be a little different... The reviews benefit from custom Lua code. Infobox convention is to rely on Module:Wikidata, which takes away some freedom. For example, in the review score module, an early issue brought up was how the systems were sorted, so I implemented a simple alphabetical sort. VG Infobox tends to have a lot of carefully ordered fields such as release date and what order to place writers or composers in, etc.... By default, these would be returned in the order they were input into Wikidata. I'm not sure Module:Wikidata provides a way to address that, though I know it currently doesn't look at rankings. Most of the data is there, but presentation may be nitpicked. -- ferret (talk) 11:38, 26 April 2016 (UTC)
  • The released field in particular will be difficult, getting it to follow local project guidelines like "Only English, unless developer region" and grouping systems, etc. -- ferret (talk) 14:34, 27 April 2016 (UTC)
For the last example, couldn't we check whether the QID's developer is Japanese and then only display the date if so? I don't think we'll hit all the edge cases at once, but we should start somewhere. This might be the type of application that shows the Wikidata devs what kind of contingencies we'll need. Is it worth thinking about rewriting the infobox in Lua? czar 20:34, 27 April 2016 (UTC)
I'm wary of rewriting the infobox in Lua, as I suspect that would break typical convention of using Template:Infobox as a base. A better solution I think would be to have a "Template:Infobox video game/released" sub-template that wraps and hides the custom logic required. This lets the primary template keep normal format. -- ferret (talk) 21:29, 27 April 2016 (UTC)
That makes sense czar 22:56, 27 April 2016 (UTC)

@Czar and Izno: I've implemented a few easy fields on the sandbox: Developer, publisher, director, composer, series, genres, modes, engine and platforms. I'll add some more as I have time to dig up the right properties. The release date field still gives me the most concern but for most fields, the simple list format generated by Module:Wikidata should suffice. Plus it can always be overridden. To test it, visit any article, like Dark Souls III (I populated some of the properties for testing), replace the entire infobox with {{Infobox video game/sandbox}}, and press Preview (Don't save!). -- ferret (talk) 18:29, 16 May 2016 (UTC)

The first value popping out of genre and mode should probably be upper case first letter; not sure how to implement that in Lua, though there is a parserfunction at {{ucfirst:VALUE}} which probably has an mw API representation (or in the parserfunction API). I'll do some more poking later. --Izno (talk) 18:48, 16 May 2016 (UTC)
The genre call might want to check that the genres popping out are instances (P31) or subclasses (P279) of video game genre (Q659563) and display only those genres, since I'm aware of some genres on Wikidata video games items reference literary genres. I personally consider that a good thing in Wikidata, and it might be neat to start displaying that on our templates, but it's not what 'genre' exactly means in our context. --Izno (talk) 19:08, 16 May 2016 (UTC)
When more than one output is produced for a particular key, include "s" e.g. "genres", else "genre", per #Getting rid of these ugly (s)s below. --Izno (talk) 19:12, 16 May 2016 (UTC)
Upper casing first letter should be fine, ucfirst should work without issue. Checking for instance/subclasses may require a sub-template with lua module, as I do not believe Module:Wikidata will go to a second level to check values like that. I do not believe I can handle programmatically deciding if a field is singular or plural (Is "Microsoft Windows" plural? The template doesn't know the context, just that there's a space or comma or ends in s...etc.) -- ferret (talk) 19:15, 16 May 2016 (UTC)

Singular: Yes, I'm looking at the sandbox implementation and it's using the functions from Module:Wikidata rather than a module (e.g. Module:Infobox video game using Module:Wikidata and Module:Infobox rather than Template:Infobox video game using Template:Infobox and Module:Wikidata). @RexxS: Maybe it would be good for Module:Wikidata to provide a function to "count" the set of values e.g. getValueCount.

Regarding p31/p279: just a thought; I'm not particularly concerned about it. That's something for deeper discussion at Module talk:Wikidata, since it's non-trivial probably. --Izno (talk) 19:24, 16 May 2016 (UTC)

Counting the number of items is trivial from within Module:Wikidata: if you look at function getValues, it constructs a table of values called "out", so #out will be the size of the table, hence the number of items. Just duplicate that bit of code and strip out all the stuff about sitelinks and labels, then return #out -- or even just increment a counter as it iterates through for k, v in pairs(claims) do.
By the way, function getValues does look at ranks and returns only the best in cases where the value is not a linkable article. I have a sandbox function Module:Sandbox/RexxS/ImageLegend that demonstrates handling ranks for the image (P18) image property, but haven't included it in the main module because I get moaned at now when I add code to it.
By the way II, if you use the parser functions like this: {{ucfirst:{{lc:multi Case WORDS}}}} you get Multi case words i.e. sentence case, which we usually want. --RexxS (talk) 19:52, 16 May 2016 (UTC)
Sentence case implemented for genre and mode. I'd like to keep this template in typical (?) infobox format, using Infobox as a template base, but if we need to Lua-it-up I can go that route. That does provide for some possible performance enhancements. -- ferret (talk) 21:09, 16 May 2016 (UTC)

Hoping to come back around to this as most of the work on Template:Video game reviews is completed. -- ferret (talk) 00:51, 1 July 2016 (UTC)

I'm going to try to test the sandbox at Fallout 4: Far Harbor. @RexxS: A few thoughts you might be able to help with... {{ucfirst:}} doesn't seem to be working for genre and modes. I'm not sure why. They are coming out in all lower case currently. Second... The series should be italicized. Is there a way to have getValue do this? -- ferret (talk) 13:09, 1 July 2016 (UTC)
@Ferret: I've fixed the sandbox template so that it italicises the series if it exists - that's generally how we deal with returned values that need post-processing in Wikidata-enabled templates. I can see that {{ucfirst:{{lc:action role-playing game}}}} works (-> Action role-playing game), but {{ucfirst:{{lc:[[action role-playing game]]}}}} doesn't (-> action role-playing game), because the 'magic words' are too dumb to realise that [ isn't actually the first character of the name. I'll go and have a look for a string function in a Lua module that will do the job in a smarter fashion, otherwise I'll have to write one - we can see that there's a need for it. --RexxS (talk) 14:07, 1 July 2016 (UTC)
Ugh, didn't even occur to me that '[' is technically the first character. -- ferret (talk) 14:11, 1 July 2016 (UTC)
It turns out that I'd written Module:Sandbox/RexxS/String2 in 2013 - just chalk it down to senility (see update below). I've now updated it to find the first real letter and capitalise that.
  • {{#invoke:String2|sentence|action GAME}}
  • Action game
  • {{#invoke:String2|sentence|[[action Game]]}}
  • Action game
See if that works for you. If not, give me another ping. --RexxS (talk) 15:11, 1 July 2016 (UTC)
Update: I'd forgotten I'd actually created Module:String2, so that's available instead of my sandbox. --RexxS (talk) 15:25, 1 July 2016 (UTC)
Another update: @Ferret: I needed to also handle piped wiki-links. Now done.
Hope that should be enough to solve your concerns above. --RexxS (talk) 22:58, 2 July 2016 (UTC)
Yes check.svg Done @RexxS: Between real life and working on Module:Video game wikidata I had not swung back around to this. String2 has been added to the sandbox and the first letter of genre and mode is now capitalized. -- ferret (talk) 13:00, 15 July 2016 (UTC)

Progress[edit]

I made some progress today. Here's some issues that need addressed, ignoring for the moment the released field, which has all kinds of arcane rules...

  1. "Producer" has a property available, P162, however, Wikidata seems to have defined it narrowly as only for film and TV.
  2. "Artist" has no suitable property, with the closest matches being creator P170 (Too generic) or cover artist P736 (Too specific).
  3. "Writer" is implemented as P50, which is "author" in Wikidata. This should be fine, but in case anyone has concerns I've listed it.
  4. Except for CPU, NONE of the arcade fields have suitable wikidata properties. In fact, the scope of the arcade fields seems so narrow that I almost think they should be removed. If we cannot wikidata-source all of them, I believe we should leave CPU out as well.

Thoughts in general? -- ferret (talk) 19:38, 15 July 2016 (UTC)

  1. Nothing in the constraints restricts P162 to film and TV, so I don't see an issue there. However, what about d:Property:P272?
  2. Yes, we'll need a property here, I think. Not sure how to make it sound sane to the Wikidata peoples.
  3. No issue here.
  4. Those are leftovers from Template:Infobox arcade's merge here. They seem fine to include specifically for arcade games. I'll look into what Wikidata properties are needed, though it sounds like "all" is the answer".
--Izno (talk) 22:09, 15 July 2016 (UTC)
P272 seems oriented towards companies/businesses rather than people. What bothered me about d:Property:P162 was the instance Of claims. Are those not soft constraints for usage? There appears to be no "Wikidata property for items about video games" item, though perhaps we should consider creating one. -- ferret (talk) 22:16, 15 July 2016 (UTC)
That's just people on Wikidata getting classification-happy (it's driving me nuts too ;D). The constraint on the talk page saying that all claims must be a subclass of "work" is more telling about the use of the property. --Izno (talk) 22:21, 15 July 2016 (UTC)
I've implemented d:Property:P162 for Producer. I've submitted a request for Artist. I believe its a sound enough rationale. (Though I just now found d:Property:P110, but it seem narrowly defined and doesn't really mean the same thing.) -- ferret (talk) 22:36, 15 July 2016 (UTC)
Just noticed cs:Šablona:Infobox_-_videohra, another Infobox video game that's already data enabled. They also had to leave Artist blank. At least they picked the same properties I did. :) -- ferret (talk) 22:53, 15 July 2016 (UTC)
  • I agree—I'm not sold on the necessity of the Arcade parameters. They tend to be dubiously sourced and of little overall value. czar 19:09, 16 July 2016 (UTC)

Waiting on the artist property request to fail, then I will have to convert the entire infobox to Lua. Module:Wikidata doesn't have a method currently for handling sub-querying of qualifiers, so need a little bit more logic. I don't anticipate it being overly difficult, and it has some performance boons. -- ferret (talk) 12:53, 21 July 2016 (UTC)

  • I commented there (finally). Still have to get to your talk page. ;D --Izno (talk) 13:23, 21 July 2016 (UTC)

Release dates[edit]

This is going to be impossible to get exactly right for the expectations of the project, probably, but I believe we can get an approximate that works in most cases. Likely, release dates will be the most commonly overridden field in the end though. Please read the proposed logic and let me know your thoughts:

The documentation, for quick reference: "Add release dates according to the platforms field, for English-language regions and the developer's region. Use only general public release dates, not festival, preview, or early access dates. If possible, use the game's exact release date ("July 21, 2016"). Use the {{Video game release}} template: {{Video game release}}. If there are many release dates, enclose them all with the {{Collapsible list}} template[2] and add the field titlestyle=font-weight:normal;background:transparent;text-align:left; followed by title= set as the earliest release date. Platforms can be abbreviated to fit in one line and should be listed as bolded section titles without colons, separated with commas (e.g. PS2, NGC, Xbox)."

  1. Pull all Publication date.
  2. Pull the developer, and check for claim country.
  3. On each, check for qualifier place of publication.
  4. For each place of publication, check a table of languages. If not found, tried to populate by pulling official language. If it is English, or the country matches the developer's country, then store in a table. (The table above is necessary because good ol' USA has no official language to check)
  5. At this point we have all of the publication dates that are from English regions or the developers country. (To best of our ability to determine).
  6. For each publication date, pull the qualifier platform if available.
  7. Group dates/regions by platform.
  8. For each platform, sort publication date.
  9. For each platform, output: Platform {{video game release}}.
  10. If more than X total platforms, wrap the entire thing in {{Collapsible list}}.

Caveats: If one platform, exclude platform.

This is a very rough draft of the logic. It should produce something akin to typical displays, but there's going to be no way to make it perfect for the more nitpicky editors. -- ferret (talk) 13:53, 21 July 2016 (UTC)

Would this be easier by adding the region on the video game item? I think we'd need a new property for that, but it sounds like it would mostly fix the mess above. --Izno (talk) 13:56, 21 July 2016 (UTC)
Need more details on what you're thinking. Would region be a qualifier on the publication date? Note the diff of this edit, I adjusted the logic some because I got off in the weeds. We don't need the language of the developer, only it's country. Language is only needed to filter place of publications that do not match developer. -- ferret (talk) 14:11, 21 July 2016 (UTC)
Yes, a qualifier on the publication date. --Izno (talk) 14:21, 21 July 2016 (UTC)
Which would naturally (duh) publication location, which I believe exists presently. --Izno (talk) 20:29, 21 July 2016 (UTC)
Yes, place of publication. -- ferret (talk) 21:23, 21 July 2016 (UTC)
  • I think a lot of this struggle could be for naught. More often than not, none of these dates are sourced in articles. We end up having a giant list of dates that are only off by a day or two that should really be explained better in prose. For the minimum viable infobox, I would think that you only need to have the first release date up, similar to how the film etc. infoboxes do it. You can try to code something that approximates the mess that we currently use but personally I don't see the use. czar 19:09, 21 July 2016 (UTC)
    • I agree, I don't really see the point in making this needlessly more complex. ~ Dissident93 (talk) 01:02, 22 July 2016 (UTC)
      • @RexxS: Is there a Module:Wikidata call that will pull the earliest date for a property? GetValue looks like it grabs everything. If not I'll whip something up quickly. -- ferret (talk) 16:34, 3 August 2016 (UTC)
        • @Ferret: There's no call written to do that filter; getValue is intended to return the list of values for a given property. You'll need to iterate through the property values and look for start date qualifiers. The values of those are time stamps which are strings, but they will still sort/compare cleanly, except that you'll have to watch out that some might be a year (["precision"] = 9) stored as something like ["time"] = "+2016-00-00T00:00:00Z" which sorts/compares "earlier" than 1 January 2016. You may also need to allow for the theoretical case when two values of the property have the same start date. Cheers --RexxS (talk) 18:24, 3 August 2016 (UTC)

Go forth?[edit]

While we are still pending final options for Artist and Release Dates, there's nothing to prevent moving forward with the rest. Is there any opposition to moving forward with this sandbox version? It is inline with other infobox implementations that use Module:Wikidata. We can expand and tweak from there. I would like to get this installed before looking into other requested changes such as the "(s)" labels. -- ferret (talk) 16:40, 3 August 2016 (UTC)

I will be implementing the version above today. There has been plenty of talk page traffic and no opposition voiced. -- ferret (talk) 11:15, 8 August 2016 (UTC)

Preceded by[edit]

Can we have a field like this for games in a series like Resident Evil 2 so it is easier to click on what comes before and after without scrolling to bottom for a chart? 64.231.169.3 (talk) 01:33, 28 June 2016 (UTC)

It's been discussed several tmes before, and it's not a good idea, as the question with several games if you put it by order of release date or by order of story chronology, and do you include DLC, and the like. It's too tricky a field to use. --MASEM (t) 01:49, 28 June 2016 (UTC)
Also the predecessor/sequel tend to be linked in the lede too czar 01:51, 28 June 2016 (UTC)

Platform field order[edit]

We have discussed this before but haven't seen any resolution and the last convo seems to be 2013. Lately, I have been seeing changes to platform ordering in the infobox to be alphabetical regardless of any other circumstances, though at one point I though it was generally considered to be PC versions then consoles then handhelds then mobiles. I'm not against going alphabetically but we should make this a standard if we want to go this way, including if we alphabetize on "Windows"/"Microsoft Windows", "Nintendo 3DS"/"3DS", etc. Note that this should also apply to the Series infobox too. --MASEM (t) 14:13, 2 July 2016 (UTC)

  • It's not something I'm particularly bothered by, but I would say list in chronological order, and then alphabetical for simultaneous releases (to match the order used in the release field). For old games, I think it would more logical for the original platform to appear first and rather than platforms the game was ported to years later. --The1337gamer (talk) 14:23, 2 July 2016 (UTC)
    • I thought about that, but that's what the release date field should be doing, making it clear on such ordering, so seems unnecessary to duplicate it.
    • There's also the case that I'm not sure about for a game that was came out exclusive for one platform and then later got ported across the board (eg Limbo for 360 for example). But again, if the release field is formatted correctly that should be apparent from a glance. --MASEM (t) 14:25, 2 July 2016 (UTC)
      • I definitely think that the best way to list them is to list the PC versions first, followed by consoles, followed by handheld, and followed by mobile. It just really seems odd to have Android listed as the first platform for Mortal Kombat X when the game is primarily known as a console game instead of a mobile game. Alternatively we can list the platforms they are primarily associated with, then the ports. If the PC game is going to be ported to consoles (like Elite: Dangerous), then PC should be listed first. If a console game was announced exclusively for a console then was announced for PC, then list the console first (like Quantum Break). AdrianGamer (talk) 16:05, 2 July 2016 (UTC)
  • My understanding is that precedence has been to order first by release and secondarily by alphabet. If there is little difference in the release window, I would ignore the order by release and go for full alphabet—the difference of a week or even a month is usually insignificant but the idea is to avoid putting "Android" first in Doom's infobox, for instance, as that would be confusing for readers. But putting PlayStation first on an article narrowly first released on Xbox isn't a big deal.
I'll add that I think this is related to infobox bloat—the platforms field does not need to duplicate what we already see in the release dates field. WP has an idiosyncratic obsession with specific release dates, even in eras where the release dates were not specific nor actually enforced with strict street dates (see moral panic over the official release date of Super Mario Brothers). The benefits of seeing all release dates for Doom or Limbo in the infobox at a glance are minimal (especially as few infoboxes are actually fully sourced like Limbo's) and would be best covered in a separate Development section template, in my opinion. If we did indeed only list the original release date, like the film infobox, the platforms field could go back to simple alphabetization, which would be easier for readers to read, and they would be encouraged (as they should) to read the corresponding section in prose for more details, hopefully with real sourcing. czar 17:13, 2 July 2016 (UTC)
    • I do think that we need to encourage the use of the collapsed list for release if there's anything more than 2 dates. And I'm not against alphabitization as long as we are consistent (and stick to our guns about avoiding emulated game released like PS2 classics on PS4, while allowing remasters to be accounted for, ala the BioShock collection), but we do need to agree to be consistent here. --MASEM (t) 17:25, 2 July 2016 (UTC)
  • It's supposed to be listed chronologically first, per WP:VG/STYLE, and only alphabetically if it was released the same day. ~ Dissident93 (talk) 22:59, 2 July 2016 (UTC)

"Microsoft is just a disambiguator"[edit]

As changed by Czar ealier this morning, according to him, the "Microsoft" in "Microsoft Windows" is just there to disambiguate from the plural of "windows". It has so far only been him who is enforcing this way of using Microsoft Windows, as he is forcing this into the Microsoft Windows article as well. As far as I know there was no consensus on this. Windows truly is the common term, but if we consider it the term, we should get it moved to "Windows" instead of always having a redirect or a pipe. But first we would need to know, does everyone agree with this? All Windows logos up until Windows XP's included the "Microsoft", so it at least was the official, full title, and I do believe it still is, just cut off since Windows Vista's logo for whatever reason. So what do you think, leave or change? Lordtobi () 15:02, 2 July 2016 (UTC)

All of our sources in the overwhelming majority of cases describe releases for Windows and/or PC, but never "Microsoft Windows". The latter terminology has been used and reused most likely because "Microsoft" is used as a natural disambiguator for the operating system's Wikipedia page (i.e., "Windows" was taken and "Microsoft Windows" is preferable to "Windows (operating system)"). Article titles on WP only occasionally go by their "official" title—more importantly, our use of terms should follow the precedent of reliable, secondary sources.
As for the other points: "Microsoft" as natural disambiguation means that it used and piped the same way that "(operating system)" disambiguates—it doesn't get special treatment or have special need to become the primary topic for the phrase "Windows". (I would think that physical windows are the primary topic here, but note that as of this comment "Windows" does redirect to "Microsoft Windows".) And as for official titles including the developer, I invite you took at Nintendo GameCube and Sega Dreamcast naming discussions. czar 16:59, 2 July 2016 (UTC)
I think this type of discussion is one that needs to be normalized with other software projects across WP. There is similar inconsistency within {{Infobox software}} whether to spell it out. My concern is that to anyone in computing "Windows" is crystal clear in context, but would not make sense to a layperson who has no idea about software. --MASEM (t) 17:33, 2 July 2016 (UTC)
Even if it's context-dependent, whether it's listed as a platform in the infobox or among other platforms in prose, I find it hard to consider contexts where "Windows" without "Microsoft" is any more ambiguous than any other operating system name. And this aside, context is a prose issue. I wouldn't introduce "Windows" the same way that I wouldn't introduce "Xbox" without context that it's a gaming platform. czar 06:35, 30 July 2016 (UTC)
I've considered making this change many times myself. It also seems clear to me that the "Microsoft" in Microsoft Windows' article title is there purely for disambiguating the common name, and can be piped-away like any other, especially given its context in a list of software platforms. – Quoth (talk) 11:15, 13 August 2016 (UTC)

Why is there no "Written in" field?[edit]

I was just wondering why this template has no "Written in" field that the software infobox has, as it seems like an appropriate field seeing how video games, by definition, are pieces of computer software and hence are written in at least one programming language. Brenton (contribs · email · talk · uploads) 02:23, 22 July 2016 (UTC)

  • While games are software, some fields present in the software infobox are omitted in the video game infobox for not being as relevant to video games specifically. We don't include programming languages. But if a notable game engine was used, we do have a field for that. Reach Out to the Truth 02:40, 22 July 2016 (UTC)
  • Because the game's engine is more important to know, and whatever language the engine was coded in is listed on the article for it, making this redundant. ~ Dissident93 (talk) 08:28, 22 July 2016 (UTC)
  • Also, this information is very rarely sourcable reliably for video games and would require OR for almost every game. It's just not common practice in sources to focus on (or even mention) languages when games are concerned. —  HELLKNOWZ  ▎TALK 09:23, 22 July 2016 (UTC)
  • Most importantly, it's out of scope. The infobox is supposed to be for quick access to key facts (sigh) and the game's technical details, including the language in which it was programmed, are not important to 99% of readers of a generalist encyclopedia. In a more technical wiki, sure czar 19:38, 22 July 2016 (UTC)

Multiple release dates in the infobox[edit]

Per the discussion at Template talk:Video game release new#Font size is not accessible, I think we should reconsider our standard of shoving every release date into the infobox. Here are a few reasons:

  1. The content is almost always unencyclopedic. For all the parading of the difference between NA/EU/AUS region releases, the difference is at most a week, on average. In all of my editing, I cannot recall this date difference being of importance to any of our sources, nevertheless a fact that readers need to see. (Is there any encyclopedic purpose for knowing that the official release date was first in NA instead of Europe? Not according to our sources.) I doubt whether these specific dates are in our purview.
  2. They are rarely or atrociously sourced. Like most of the infobox parameters, they are more likely to be unsourced than if they were included in the text. (Really they're supposed to be included in the text but more on that below.) We don't have any canonical source for release dates. We have yet to verify the origins of IGN's dates (yes, I've tried contacting them), but they were associated with the site's user-contributed wiki for some time. We dropped GameFAQs/MobyGames as a source for dates a long time ago. In the end we're really trying to pull what historic press releases and routine reporting says, but I can even attest with Videoball's recent release and all kinds of reporting on breaking street date that not even modern dev release dates necessarily match what was reported...
  3. Which brings us to citogenesis. Gaming culture, and our editors likewise, have a fruitless obsession for exact release dates when often ranges or simply "MMMM YYYY" would be a better fit. We host reams of unsourced, specific dates, that often make their way back into user-submitted databases and make us just as guilty for perpetuating inaccuracies. Our dates are supposed to be sourced—full stop—and if we can't do that well, we need to look at whether we are a good source for compiling that information. I have my own doubts about the accuracy of information we now pull from the newer Nintendo sites (Nintendo Life, even NWR) as they often have release dates in our format that match our exact dates, yet I haven't been able to find any secondary source verification even in historic newspapers for such dates. I assume that they pulled their data from us or another untrustworthy source.
  4. They belong in prose. We all agree that the release dates, if important, should come from secondary sources and be cited in prose. But we also agree that they make for really, really boring prose. Prose that I don't think should even withstand a FAC review. I've grown into the habit of relegating the cited release date prose into an endnote so it's there for those who want it but not in the text to bore anyone else. It would be better visualized...
  5. That a game was released first in North America or Europe by a few days is unadulterated trivia. This is one of those relics of the age in which Wikipedia articles had more trivia and game guide sections and less focus on Reception and Development. In my years of editing, I have never seen a source besides us ever substantively care that the title released first in a region, apart from its listing as a matter of fact. It would be more accurate to say that the game released in the beginning of a specific month, or to similarly generalize the release windows. If this information is too clunky for prose, I think a visualization by way of a Development/Release section template would be in order, similar to how we've visualized review scores or chronologies within a tricky series when warranted by the sources. For instance, we help very few with the mess that is our Doom smorgasbord of infobox dates, but perhaps we can visualize the releases better within their appropriate section and focus less on the regional releases of the minor releases, which are truly inconsequential.
  6. It is the most Frankensteined infobox parameter that I can recall out of any infobox I've seen, with multiple formatting styles (bold subheaders, superscript regions, regular dates requiring multiple templates). The infobox is meant for reporting the most important information simply, and the key detail here is when the game was released. All other specifics beyond that original release are supposed to be in prose. We would be following the standard of every other major media infobox template by only listing the date of first release.
  7. In advance of any handwringing over making a major change, we had the same situation with every other major template change (dropping media format params from the infobox, dropping GameRankings as a vg reviews necessity) but there has never been an issue as long as the consensus is documented and the reasoning is sound.

My suggestion is to remove all but the first major release date from the infobox, à la the other media infobox templates, and to brainstorm a better template for illustrating the differences in release dates when they are consequential (significant differences across multiple platforms or regions). Perhaps even a timeline? I would also suggest that we adopt language in our WikiProject guidelines to encourage release dates to be put in engaging prose with relative dates or else relegate them to endnotes. All in all, we should follow the lead of our sources by putting less emphasis/importance on specific dates, especially for older releases, and more emphasis on relative dates (MMMM YYYY or with early/mid/late prefixes or date ranges instead of X in NA, Y in Europe, Z in Australia). czar 06:36, 30 July 2016 (UTC)

  • I agree with most of what you said, exact release dates don't really matter in the long run. The lead sections of (well written) articles already have generalized release info, which is simply better prose if read out loud. ~ Dissident93 (talk) 09:44, 30 July 2016 (UTC)
  • I would say that for games that take a long time to get released in other regions than the one it originated from, such as Mystery Dungeon: Shiren the Wanderer (1995 in Japan, 2008 in the West), this time difference is important enough to be in the infobox. I think I would prefer to include the first release date for each major region in the infobox, while the rest should be moved elsewhere (but not deleted entirely). --IDVtalk 10:53, 30 July 2016 (UTC)
    • Yeah, info should be generalized, not simply removed all together, as then you could make the case if release dates matter at all. ~ Dissident93 (talk) 11:12, 30 July 2016 (UTC)
  • I agree that having regional release dates in the infobox is unadulterated trivia and that they must be removed ASAP. In fact, I previously proposed ditching them from the infobox, but no consensus was reached. However, I disagree with the fact that only one release date should be included in the infobox. Unlike films, albums, or novels, video games are tied to a platfrom, and knowing the original platform for which the game was developed is an important piece of information that should be included. I propose the following: combine the release date and platform fields together and simply call the resulting field "Release date(s)". This would include a list with two things: the first release date and its corresponding abbreviated platform. Here's a Resident Evil 3 example:
    • Sept 22, 1999 (PS)
    • June 16, 2000 (PC)
    • Nov 16, 2000 (DC)
    • Jan 23, 2003 (GC)
  • I agree with this in a general sense and originally stated it at the other talk page, including making something akin to {{Video game timeline}} to hold more detailed information (Not necessarily that format, but in that style of template format). Niwi3 just to ensure we have a clear example, in your setup how would you present multiple platforms from the same date? Would it be: Sept 22, 1999 (PS, PC)? Note that the outcome of this discussion could lead to a much more consistent format I can use in Wikidata pulls. -- ferret (talk) 14:12, 30 July 2016 (UTC)
  • There are some exceptions I would make, but would argue that release dates that are all within the same week, perhaps month, should be treated "equivalently" for purposes of the infobox. This would make cases like how most Japanese games come to the West stand out, how some ports take months to come to some systems, but normalize games that come out in different regions and on different systems within the same week as having the same effective release date. I will note that sometimes a matter of days is not as trivial (for example, No Man's Sky, there was actual sourced discussion about a two-date delay in the UK version which was resolved to get it released at the same time as the EU), but that's all good for prose, not infobox. --MASEM (t) 14:51, 30 July 2016 (UTC)
    • No no no no no no no no. We had this exact discussion just six months ago, proposed by Czar as well, and concluded that it was not a good idea. Calling the release date field of the infobox "unencyclopedic" is fundamentally misunderstanding its purpose. The true purpose is to tell me "at a glance" whether a game has been released in my (English-speaking) region. I shouldn't have to sift through prose to figure that out. If the release date was in fact simultaneous worldwide, collapsing it into one date without calling it out specifically makes me uncertain if it was actually released in my region, especially if I live in a "secondary" market like Australia (this is a US WP:BIAS problem).
      • Hmmm. So if you look in the box and you don't see your region, does that mean it's not released, or they just didn't add it? You can't know, so that renders it useless. Further, if this is presented it's useful for what, a week? And is this information not trivially available on many other websites? Maury Markowitz (talk) 17:02, 30 July 2016 (UTC)
    • The solution to #2 is to source them. "But it is le hard" is not an excuse to just remove that information altogether. Proper sourcing fixes #3 as well. For #4, I actually disagree that it belongs in prose. It only belongs in prose if, as you say, there's something notable to talk about differences in release dates. If not, then there's no need to repeat it in prose. It's a complete straw-man to call release dates "trivia" in the same way that a list of guns in Call of Duty is trivia. In the case of Doom, it's true that a better visualization would benefit the reader more, as long as the data remains whether that's in the infobox or somewhere else.
    • As I said last time, this is a solution in search of a problem. We already have a perfectly functional answer to the release date parameter getting too long in Template:Collapsible list. Wikipedia is not paper and there's no need to excise perfectly good sourced information in the name of brevity. Axem Titanium (talk) 16:51, 30 July 2016 (UTC)
      • Pinging all the participants of the previous discussion. @Rhain, Maplestrip, Favre1fan93, Maury Markowitz, AdrianGamer, Hellknowz, Famous Hobo, JDC808, Hahnchen, and Sergecross73: Axem Titanium (talk) 16:51, 30 July 2016 (UTC)
      • The issue is not to wipe out the regions, but the granularity of them. Say a title is released on a tuesday in the US, on the immediate thursday in the EU, and the next week in AUS/NZ. From the standpoint of PAST the game's release, those all are the same time, so indicating a single WW release would be appropriate. Yes, for seven-odd says, people from Australia may be confused, but we're not writing for immediacy but enduring information, and that's the goal here. --MASEM (t) 16:58, 30 July 2016 (UTC)
That wasn't the conclusion from the last discussion... If the "solution" of sourcing the dates were viable, it would have been done already. The vast majority of our release dates are not sourced anywhere in the article and many need to be referenced directly from print. It's not a solution to leave up junk dates until someone decides to take on that gargantuan sourcing effort—until then, we have an obligation to WP:V. We have removed such trivia sections in the past when poorly sourced. (Yes, if we don't have a reliable, secondary source for regional release data, it is trivia (primary source, low-impact factoids) and not at-a-glance facts.) Your Call of Duty is a perfect example: Call of Duty 4: Modern Warfare is a FA and look at its release date sourcing—either unreliable or nonexistent. And for what? The main release dates are mere days apart. I find it impossible to trust our release date content as is. czar 19:48, 4 August 2016 (UTC)
  • I'm already irritated/tired of all the constant revisions/arguments that pop up on my watchlist about not putting every platform for a game in the infobox. This sort of thing is going to double that sort of issue. It's just not intuitive - outside of the core 10-20 of us at WP:VG, everyone is going to think it's missing information. I don't like these guidelines that just increase the need for constant maintenance by our core editors. I guess it's up to you guys if you want to spend your time meddling away with this sort of thing. I'll leave that to you all. Sergecross73 msg me 17:00, 30 July 2016 (UTC)
  • This is a very good point, I think.--IDVtalk 17:17, 30 July 2016 (UTC)
  • Agreed, and to add my opinion to this discussion, I still hold my stance from the first discussion. --JDC808 04:30, 31 July 2016 (UTC)
I think I addressed this above, but the issue of drive-by editors changing infoboxes has more to do with compulsion—they see a hole in the list and go find whatever date they find online and fill it in. That's a problem with its design. We've been over fear of change before and have not had any sustained problems with enforcing, for instance, removing GameRankings from the standard vg reviews box or removing ports from the infobox. If anything maintenance should be much easier once we've reduced the reason to fiddle with the infobox. Like always, users will adjust—the question is whether the change is right czar 19:48, 4 August 2016 (UTC)
Agreed. The simpler the release date field is, the easier it is to maintain, and the less likely drive-by editors will pay attention to it. I think the question is right to a certain degree, mainly because regional release dates are not just difficult to source, but are trivia that shouldn't belong to the infobox. They should belong to the body if and only if they can be backed up with reliable sources. Release dates for multiple platforms, on the other hand, are much easier to source. --Niwi3 (talk) 09:29, 6 August 2016 (UTC)
Given we're only talking about the infobox (that is, date information can be in prose where it can be readily and properly sourced), what if we took the approach that the Film project does, which is the first release date and a second release date if that first release is not necessary in the country of origin of the game (if that even really happens?) It gives the critical bit of data - the approximate time the game was released that properly serves the general reader using the infobox, and allows proper expansion in prose or in tables if the situation gets any more complex. (And regardless of any change we make, grandfathering in existing articles and only applying this new criteria in a mandatory fashion at GA/FA review levels needs to be done) --MASEM (t) 17:25, 30 July 2016 (UTC)
  • Comment : List only first release dates by regions in infobox (|released={{vgrelease new|JP|18 Dec 1987 (FC)|NA|12 Jul 1990 (NES)|EU|14 Mar 2003 (PS1)}}), and change |platforms= to |platforms=Famicom (1987), MSX2 (1989), WonderSwan Color (2000) ... , then move other things into a new infobox of development section? --43.242.155.141 (talk) 03:17, 31 July 2016 (UTC)
Personally, I like the second part, but I really doubt we could accurately source the first part (or should) czar 19:48, 4 August 2016 (UTC)
Release years
Platform JP NA EU
FC 1987 1990 N/A
MSX2 1989 N/A N/A
WSC 2000 N/A N/A
  • I hate to see structured content disappear into thin air. The "Release years" infobox to the right of my post seems okay, though. It would be a good substitute. SharkD  Talk  10:30, 31 July 2016 (UTC)
  • As long as the regional release dates are still kept somewhere in the article then I am ok with it. The release years template seems to look quite decent, and would be a great alternative. However, I am concerned that editors are just going to throw all the release dates out of the windows instead of turning them into prose/putting them into a new template if a consensus is reached. In addition, I also thought that this template would be useless in articles like Mad Max and Star Wars Battlefront, in which they would probably only have one row. And if we are really going to use the new template, port releases should not be mentioned in the infobox as well. AdrianGamer (talk) 11:54, 31 July 2016 (UTC)
I like this template as a start. I'd recommend adding a 1px black border between the column/row headings and the data czar 19:48, 4 August 2016 (UTC)
  • Don't put release dates in prose. Lists of dates makes for godawful reading in the same way that strings of review scores do. I complain at FAC when video game articles have lists of release dates clogging up the development section. In prose, I will usually only state the month and year of the release, and have the details in the infobox. - hahnchen 17:49, 6 August 2016 (UTC)
  • So...just to confirm, I'm not seeing any consensus for change really, right? Thought it should be outwardly stated, as sometimes editors misinterpret "having the last word in" with discussion dying off as some sort of consensus in their favor. Sergecross73 msg me 17:32, 8 August 2016 (UTC)
    • I don't see consensus but I really think we need to work this out better, maybe starting with straw polls (eg "Should we have detailed release dates by region/platform in our articles to start?"). Release date stuff is kudzu that clogs up articles, whether infobox or in prose (as Hanchan identifies above), and I do feel pushing the bulk of it to tables or footnotes or something else will make our articles read much more smoothly. --MASEM (t) 17:39, 8 August 2016 (UTC)
  • I created the Draft:Template:Video_game_release_table, and tested it in Final Fantasy V#Ports and remakes and Final Fantasy VI#Re-releases. --117.136.26.138 (talk) 12:46, 13 August 2016 (UTC)

Wikidata Go-Live[edit]

Please be aware that Template:Infobox video game will become "Wikidata aware" today. A basic guide to Wikidata editing is available at WP:VG/WD. If you encounter any major issues, please open a section at Template talk:Infobox video game. The implementation we are using is typical and relies on Module:Wikidata, which helps standardized retrieval and formatting of Wikidata properties.

Note that a blank template parameter will prevent Wikidata pulls. If you begin populating Wikidata and expect to see that data in the article, make sure you remove placeholders. This is by design based on the last RFC that allowed for Infoboxes to pull Wikidata. -- ferret (talk) 11:31, 8 August 2016 (UTC)

Live The template update is live. I've reviewed across some 40+ articles now and have not found any outstanding issues at this time. -- ferret (talk) 13:56, 8 August 2016 (UTC)

Could the release dates be put in a "Wikidata aware" template as well? SharkD  Talk  03:13, 9 August 2016 (UTC)
Because of all the arcane rules about the release field, as well as the on going discussion just above this one, I have not tackled the release date field yet. It will require a bit of custom coding to accomplish. It will however be included in the future. -- ferret (talk) 11:24, 9 August 2016 (UTC)
I'm not real familiar with Wikidata. Will each game have its own data "item" (not sure what to call pages at Wikidata), or is the data spread across multiple "items"? Can multiple templates pull data from the same item? We could generate a separate table for release dates as long as the data itself is stored in the same Wikidatabase. SharkD  Talk  01:46, 10 August 2016 (UTC)
"Item" is usually what it's called. Each item corresponds to an entity (fictional or otherwise), wherein we make statements about the entity (and not Wiki*'s description of such). Most of the Wiki*-notable games have an item, and all should. A separate template could pull that information (I'm against (re)moving that information from the infobox; the table is thin at-best and the infobox really is the place for it). --Izno (talk) 11:24, 10 August 2016 (UTC)
@Ferret: The "[Edit on Wikidata]" on the bottom of the infoboxes looks out of place - I'd suggest removing it. Also note that other infoboxes don't have such either. Maybe the pencil could stay (it should have a tooltip saying "edit") if it's moved out of the extra row to some appropriate place in the box. --Fixuture (talk) 21:13, 9 August 2016 (UTC)
(edit conflict) @Fixuture: It's sensible to indicate to editors that they may need to make changes by editing the associated article on Wikidata, and we should be encouraging infobox designers to include such links. The edit link as text was adapted from the Spanish Wikipedia - where it is common - and has been in use in {{Infobox person/Wikidata}} since June 2015. The updated {{Infobox telescope}} has been in use for several months and normally pulls all its data from Wikidata. You can see the [edit on Wikidata] at the bottom of the infobox of any telescope article. If you browse through Category:Templates using data from Wikidata, you'll doubtless find other examples. The pencil does have a tooltip that says "Edit this on Wikidata". If you're not seeing that when you hover over the pencil, please let us know your OS and browser and we'll look for a fix for you. --RexxS (talk) 21:50, 9 August 2016 (UTC)
Some first comments: I went to implement the below in Dota 2, and I got some interesting display:
{{Infobox video game
| title = Dota 2
| image = DotA2.jpg
| producer = Erik Johnson
| writer = {{Unbulleted list|[[Marc Laidlaw]]|[[Ted Kosmatka]]|Kris Katz}}
| released = '''Microsoft Windows'''{{Video game release|WW=July 9, 2013}}'''OS X''', '''Linux'''{{Video game release|WW=July 18, 2013}}
}}
  • Where I expected to see only Valve in the publishing organizations (in an attempt to get to the same infobox as present), I see the other 2 institutions on the Wikidata item. Both Nexon and Perfect World are/were correctly publishers outside the main regions (China, Korea, and Japan; Nexon no longer publishes the game--to be qualified on Wikidata shortly--but Perfect World still does). How should we square that in the infobox?
  • Single and multiplayer aren't getting/using the short names as previously discussed. I'm not sure if we implemented that, but that's a point at which to poke.
  • Source 2 has a redirect to Source. I don't think there's any way (besides using a Wikidata-side hack of a thing) to get that to display locally instead of the link to Wikidata. The question of engines (and their various redirects) deserve some discussion, I think.
--Izno (talk) 21:47, 9 August 2016 (UTC)
The same infobox guidelines should allow apply to this, right? So Nexon and Perfect World shouldn't be listed there too. I also see other errors (Like Chet being a credited writer), but everytime I've tried to remove the info, it wouldn't let me. ~ Dissident93 (talk) 21:59, 9 August 2016 (UTC)
That's essentially my question. Nexon and Perfect World however are correctly distributors in those countries, so removing them from Wikidata is out of the question (also out of the question would be to use d:Help:Ranks to resolve the issue, since while we might prefer Valve, the Chinese WP might not). Let me know if you need help with it--you might have been edit-conflicting me. --Izno (talk) 22:03, 9 August 2016 (UTC)
Well, it needs a way to not show on the English Wikipedia then. Personally, I do not like the transition to Wikidata for various reasons. Is this a site-wide guideline shift, or something I've missed on WP:VG? Oh, and BTW, Nexon isn't the publisher in Korea/Japan anymore, Valve directly took over last fall. ~ Dissident93 (talk) 22:05, 9 August 2016 (UTC)
I know that it needs a way--that's why I asked to see if there's a way (besides RexxS's comment below). You should review Wikipedia:Requests for comment/Wikidata Phase 2 regarding your question. Lastly, regarding Nexon, please review d:Help:Qualifiers--they basically allow for that nuance. --Izno (talk) 22:11, 9 August 2016 (UTC)
From a three year old discussion? I wonder what WP:VG members would think now (assuming I haven't missed anything else), because to me, all this is doing is making editing and maintaining infoboxes more confusing for no reason. ~ Dissident93 (talk) 22:20, 9 August 2016 (UTC)
It's only 3 years old because no-one has implemented it yet :D. And as regards what VG members thing, that's a non-starter (though their feedback is welcome) per WP:CONLOCAL. You don't have to use a Wikidata infobox if you don't want, but I think most people will, for a number of different reasons. As it happens, I expect there will be a general tendency toward "let's improve the Wikidata version of this when we do our quality push for a particular article, of whatever level" (this is what inspired me to work on Dota 2 right now), but for now, there's obviously no requirement to do so. --Izno (talk) 22:24, 9 August 2016 (UTC)
(edit conflict × 3) @Izno: infoboxes aren't the place to expound the nuances of which publishers no longer publish. If the Wikidata doesn't match the consensus summary (just Valve displaying), then supplying |publisher=Valve is the most sensible fix. We could at a later stage make use of the qualifiers in Wikidata for publisher (P123), but surely we should be looking to implement a working wikidata-aware infobox before trying to fix all the edge cases? Incidentally, I don't think that removing claims from Dota 2 (Q771541) to force it to fit the en-wp idea of who publishes the game is a sensible way forward. --RexxS (talk) 22:07, 9 August 2016 (UTC)
I agree regarding your incidental. Will think on the rest. --Izno (talk) 22:11, 9 August 2016 (UTC)

@RexxS: Regarding your "shouldn't we shoot for working": I think it is working (besides video game artist, which will probably have its property approved shortly, and dates, per ferret's comment above). So I think now is the time for "let's talk about edge cases" and sometimes not-so-edge cases, as is the question with video game engines (which we generally redirect on en.WP).

Aside: Part of the problem, I think, is that we've gotten problematic on what isn't allowed in this infobox. I see little issue with displaying everything, or near everything, regarding publishers. Maybe dates and platforms will be a different question, but there are rarely so many publishers/distributors that we can't just list them all. --Izno (talk) 22:20, 9 August 2016 (UTC)

I'm not sure I'm content about the wikidata-aware infobox working, when the data that it's drawing on is almost entirely unreferenced. I'm working my way through Dota 2 (Q771541) adding references to get an idea of the size of the problem (and make an useful example to show off the new infobox). I'd be happy to look at implementing calls that filtered out unwanted information, e.g. publishers that no longer publish the game, but there's a broader set of issues about how to determine what filters to apply under what circumstances. I sympathise with Dissident93 and I'm not sure how we would implement filtering the Wikidata in infoboxes in a way that editors will not find it more effort than what they do now. If we can always 'hard-code' a filter into the infobox design, then it's worth doing; but as soon as we find we need one filter for one article and a different filter for another, we're never going to sell that to the editors who will have to implement it on an article-by-article basis. --RexxS (talk) 22:45, 9 August 2016 (UTC)

Regarding the specific case of referencing, it's a pain right now for any source which is not one we can re-reference over and over (as is the case with most video games). Citoid support for Wikidata will help fix that, but that's a ways away.

Generally agreed regarding filters. --Izno (talk) 22:53, 9 August 2016 (UTC)

Well, adding a few refs to Dota 2 (Q771541) wasn't too bad, Magnus' drag-and-drop-references gadget makes life easier (it's in your Wikidata preferences - see https://www.youtube.com/watch?v=jP-qJIkjPf0 if anybody watching this page hasn't seen it). I only had to manually add the archived refs that the gadget doesn't do yet. I had the article open on one monitor to search for words and references and the Wikidata entry open on the other. It's a productive way of spending a relaxing hour, and might even take over from editing Wikisource as a way of getting away from the battles on Wikipedia. --RexxS (talk) 23:14, 9 August 2016 (UTC)

Good evening. Lot of replies in a short while here... Izno and RexxS seem to have things well in hand as far as replies go. Just my quick few points:

  1. We will never get this to pull from Wikidata and fully obey the various arcane rules this template has wrapped around some fields. The release date is the worst offender, of course. If the Wikidata format doesn't work for a particular field at a particular articles, we'll just need to use a local value there. See my subsection above about the release date where I tried to map logic that would obey the current rules.
  2. WP:VG was asked a couple times for input on this effort, and it was developed over several months with discussion here. I believe enough notification was provided, and as long as there are no immediate issues caused by the template update, it should not harm any existing article infoboxes, which are fully populated with local values. Even empty local values prevent Wikidata pulls.
  3. Wikidata being unreferenced is something we can easily fix. Presumably, our articles contain references already that cover the infobox. This can be added into Wikidata fairly easily. I've done it by hand without much trouble and as RexxS noted, there's a few gadgets to aid.
  4. Regarding the publisher/distributor topic: From an encyclopedic point of view, why should Nexon be removed? It should be noted with a date range to show they are no longer publishing, but from a historic view point, they should be captured somewhere in the article. Thoughts?
  5. We'll definitely keep working on this and refining it.

Just quick thoughts, been very busy IRL but am still watching over this for issues. -- ferret (talk) 00:35, 10 August 2016 (UTC)

"why should Nexon be removed?" - It's a lot of work and results in a lot of clutter. Not sure if that's a encyclopedic answer, but there you have it. SharkD  Talk  01:42, 10 August 2016 (UTC)
The article itself does say "In November 2015, Nexon announced they would no longer be operating servers for Dota 2, allowing Valve to take over direct distribution and marketing of the game in those regions." if that's what you mean. ~ Dissident93 (talk) 02:21, 10 August 2016 (UTC)
I think the issue is less that Wikidata contains Nexon, and therefore shows it, so much as the current way of pulling the data doesn't integrate the qualifiers and add necessary context. For example, Nexon should say "Nexon (JP, 2004-2007)" or something similar. This can be accomplished with a sub-template/module. I can create {{Infobox video game/publishers}} and have it format the Wikidata pull with further information like that. THis is likely what will be required for the Released field anyways, i.e. {{Infobox video game/released}}. -- ferret (talk) 10:31, 10 August 2016 (UTC)

Wikidata - On going[edit]

I'm creating a new section to discuss specific future enhancements and efforts. The above section was intended more to capture issues directly stemming from implementation. I am starting this section with three sub-sections that further work or discussion have already brought up. Please create a new subsection if you have concerns/ideas/thoughts on a specific field -- ferret (talk) 11:02, 10 August 2016 (UTC)

Publisher[edit]

In the above section for Go-Live there's a lot of discussion centering on the publisher field, such as displaying publishers for non-English regions or publishers who no longer publish. From a historic point of view this data is accurate and therefore should not be removed from Wikidata, which does not obey or honor the display rules of this template. For now, this means if Wikidata has more than you want, you'll have to use a local value like in the past. I suspect there's very few articles that lacked a local publisher value, so I don't think we're causing any trouble on current articles, but we need a solution to better integrate qualifiers here.

General suggestion: A sub-template like {{Infobox video game/publisher}} will pull publisher information along with their qualifiers, so that more concise data can be generated, such as "Nexon (JP, 2004-2007)". -- ferret (talk) 11:02, 10 August 2016 (UTC)

Any further thoughts on how to qualify/filter/display publishers from Wikidata? Otherwise I will look into trying to implement a "Publisher (region, startyear-endyear) format. -- ferret (talk) 14:07, 19 August 2016 (UTC)

Released[edit]

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.

This is going to be impossible to get exactly right for the expectations of the project, probably, but I believe we can get an approximate that works in most cases. Likely, release dates will be the most commonly overridden field in the end though. Please read the proposed logic and let me know your thoughts:

The documentation, for quick reference: "Add release dates according to the platforms field, for English-language regions and the developer's region. Use only general public release dates, not festival, preview, or early access dates. If possible, use the game's exact release date ("July 21, 2016"). Use the {{Video game release}} template: {{Video game release}}. If there are many release dates, enclose them all with the {{Collapsible list}} template[2] and add the field titlestyle=font-weight:normal;background:transparent;text-align:left; followed by title= set as the earliest release date. Platforms can be abbreviated to fit in one line and should be listed as bolded section titles without colons, separated with commas (e.g. PS2, NGC, Xbox)."

  1. Pull all Publication date.
  2. Pull the developer, and check for claim country.
  3. On each, check for qualifier place of publication.
  4. For each place of publication, check a table of languages. If not found, tried to populate by pulling official language. If it is English, or the country matches the developer's country, then store in a table. (The table above is necessary because good ol' USA has no official language to check)
  5. At this point we have all of the publication dates that are from English regions or the developers country. (To best of our ability to determine).
  6. For each publication date, pull the qualifier platform if available.
  7. Group dates/regions by platform.
  8. For each platform, sort publication date.
  9. For each platform, output: Platform {{video game release}}.
  10. If more than X total platforms, wrap the entire thing in {{Collapsible list}}.

Caveats: If one platform, exclude platform.

This is a very rough draft of the logic. It should produce something akin to typical displays, but there's going to be no way to make it perfect for the more nitpicky editors. -- ferret (talk) 13:53, 21 July 2016 (UTC)

Would this be easier by adding the region on the video game item? I think we'd need a new property for that, but it sounds like it would mostly fix the mess above. --Izno (talk) 13:56, 21 July 2016 (UTC)
Need more details on what you're thinking. Would region be a qualifier on the publication date? Note the diff of this edit, I adjusted the logic some because I got off in the weeds. We don't need the language of the developer, only it's country. Language is only needed to filter place of publications that do not match developer. -- ferret (talk) 14:11, 21 July 2016 (UTC)
Yes, a qualifier on the publication date. --Izno (talk) 14:21, 21 July 2016 (UTC)
Which would naturally (duh) publication location, which I believe exists presently. --Izno (talk) 20:29, 21 July 2016 (UTC)
Yes, place of publication. -- ferret (talk) 21:23, 21 July 2016 (UTC)
  • I think a lot of this struggle could be for naught. More often than not, none of these dates are sourced in articles. We end up having a giant list of dates that are only off by a day or two that should really be explained better in prose. For the minimum viable infobox, I would think that you only need to have the first release date up, similar to how the film etc. infoboxes do it. You can try to code something that approximates the mess that we currently use but personally I don't see the use. czar 19:09, 21 July 2016 (UTC)
    • I agree, I don't really see the point in making this needlessly more complex. ~ Dissident93 (talk) 01:02, 22 July 2016 (UTC)
      • @RexxS: Is there a Module:Wikidata call that will pull the earliest date for a property? GetValue looks like it grabs everything. If not I'll whip something up quickly. -- ferret (talk) 16:34, 3 August 2016 (UTC)
        • @Ferret: There's no call written to do that filter; getValue is intended to return the list of values for a given property. You'll need to iterate through the property values and look for start date qualifiers. The values of those are time stamps which are strings, but they will still sort/compare cleanly, except that you'll have to watch out that some might be a year (["precision"] = 9) stored as something like ["time"] = "+2016-00-00T00:00:00Z" which sorts/compares "earlier" than 1 January 2016. You may also need to allow for the theoretical case when two values of the property have the same start date. Cheers --RexxS (talk) 18:24, 3 August 2016 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

The above is the original subsection for the Released parameter. I want to let the old section archive off so copied here. I think we have to wait on the general discussion about the Released field before we can do something "final here". General suggestion if we want something quick for now: A sub-template named {{Infobox video game/released}} that will pull and display only the earliest publication date. No region will be listed. This is simply to add some functionality, and would not be expected to be the permanent solution.

@RexxS: Regarding the dates, Module:Date is "Wikidata" aware now. Johnuniq updated it to understand their date structure recently in support of the review score module. -- ferret (talk) 11:02, 10 August 2016 (UTC)

Current discussion appears to support having Wikidata only pull "the earliest" publication date, at least to start. More advanced formatting would be done as a local value. Anyone got other ideas? -- ferret (talk) 14:08, 19 August 2016 (UTC)

I think we should bite the bullet on this one and leave it Wikidata-disabled for now, since a) there's not a lot of pull at present to implement it and b) the rules are pathological. We can revisit this in the future. --Izno (talk) 14:24, 23 August 2016 (UTC)

Distributor[edit]

Can you hide Steam as the distributor? We don't include digital-only ones in the normal infobox, but it was showing. ~ Dissident93 (talk) 03:50, 11 August 2016 (UTC)

Another arcane rule set that will be somewhat difficult to implement right. How do I determine if a particular distributor is "digital-only"? Will require a special sub-template {{Infobox video game/distributor}} to filter either a set list of values out, or handle the data pulls to determine if a given distributor is "digital-only". -- ferret (talk) 11:22, 11 August 2016 (UTC)
I think "pathological" is more correct, since that rule is new from within the past year or two. :D --Izno (talk) 11:40, 11 August 2016 (UTC)
It's still no excuse to ignore what passed consensus though. And again, all this has done so far is create new issues/more work for editors for no real gain. ~ Dissident93 (talk) 20:32, 11 August 2016 (UTC)
I have to disagree. It's about enriching Wikipedia as a whole, not just enwiki. This helps populate Wikidata and allow every Wiki to use the same data. Where it doesn't work here on enwiki, a local value can be supplied, exactly as has always been done. There is no extra work, nor any mandate to remove existing local values. Just do what you've always done if a particular field needs it. If Wikidata just doesn't suit you at all, completely ignore it and do as you've always done in the past. It is simply a new option in our toolbox, and we'll continue to improve it. -- ferret (talk) 21:02, 11 August 2016 (UTC)

At this time I do not have a suitable way to filter distributor, or any suggestions to implement. If distributors from Wikidata do not fit a particular enwiki article, a local value should be used. -- ferret (talk) 14:10, 19 August 2016 (UTC)

I just used the hidden markup to override it, but this still isn't perfectly ideal. ~ Dissident93 (talk) 20:12, 19 August 2016 (UTC)
That is the correct/accepted way to do it per the last major RFC for Wikidata inclusion in Infoboxes (Which yes, I know, was 3 years ago). The hidden comment isn't necessary but is nice and informative. If someone can suggest a reliable way for us to mark distributors as "online only", I can build a method for excluding them from display here on enwiki. -- ferret (talk) 20:59, 19 August 2016 (UTC)
  • Is there a way we can detect if the publisher is the same as the distributor? In cases like Fallout 4, Bethesda is listed twice, as both publisher and distributor. (I have personally removed it from the wikidata) I thought this is our consensus, or at least our common practice? AdrianGamer (talk) 12:58, 25 August 2016 (UTC)
    • @AdrianGamer: It should only be removed from Wikidata if it's wrong. To suppress it from displaying on enwiki, specify an empty distributor with |distributor=. Please add back to Wikidata. See WP:VG/WD for more detail. -- ferret (talk) 13:01, 25 August 2016 (UTC)
    • "distributor" was misspelled at Fallout 4. Fixed. -- ferret (talk) 13:12, 25 August 2016 (UTC)
      @Ferret: I think it would be pretty easy to do a check in the template using just the getValues and some simple parser functions (or maybe a lua function to do comparison) to do the suppression as requested. --Izno (talk) 13:23, 25 August 2016 (UTC)
      Tested in this diff. This works except in one case. If the local |publisher=[[Bethesda Softworks]], and distributor pulls the same value from Wikidata, distributor will still display. This is because Module:Wikidata is using a piped link (Wikidata EN label, piped to Enwiki article). The outputted display is the same, but the values during the comparison differ so it doesn't catch it. If |publisher=[[Bethesda Softworks|Bethesda Softworks]] is used, distributor is suppressed. -- ferret (talk) 13:35, 25 August 2016 (UTC)

Hide "edit on Wikidata" when infobox pulls no data from Wikidata[edit]

Animal Farm
Author George Orwell Edit this on Wikidata
Genre roman à clef, fable, satiric novel [*], novella, dystopian literature [*], political novel [*] Edit this on Wikidata
Publication date
17 August 1945 Edit this on Wikidata
Pages 112 Edit this on Wikidata
823/.912 20 Edit this on Wikidata
LC Class PR6029.R8 A63 2003b Edit this on Wikidata

This one seems like sugar, but it doesn't make sense to stuff the link in someone's face if the data isn't on Wikidata yet. --Izno (talk) 11:39, 11 August 2016 (UTC)

This would require one of two things. More drastically, a complete rewrite of the template into a module so that pulling of values can be tallied and determined if nothing was pulled. Less drastically, a new method in Module:Wikidata that takes a list of properties and returns the total count of populated items. This would be somewhat simple but becomes complicated if we do sub-templates for some of the rules around publisher, developer, etc. For example, what if Developer is in Wikidata, but our sub-template excludes the only value because it didn't match our region rules? Wikidata would say there's "1" value and display link. However, we're talking weird edge cases there. -- ferret (talk) 12:56, 11 August 2016 (UTC)
Secondary thought: If there's no data on Wikidata yet, isn't that a case where we want the link? I.e. "Go populate it here!". If you drop an empty infobox on a new article, that might be your clue to go populate the values for your new article. -- ferret (talk) 13:31, 11 August 2016 (UTC)
I realized after you responded the first time, and I wasn't sure if you had covered it in your response, that I may not have been clear. My suggestion is more pointed at: if all the fields in the infobox are locally filled (even if there is Wikidata), we shouldn't provide the edit link, as that's misleading to the readers/editors. --Izno (talk) 14:01, 11 August 2016 (UTC)
Third option then is possible, if ugly. Just chain checks for parameters. I.e. {{#if:{{{director|{{{designer|{{{programmer|<etcetc>}}}}}}}}}||Edit on Wikidata link}} -- ferret (talk) 14:09, 11 August 2016 (UTC)
Another alternative is to place the edit-on-Wikidata icon after each value that is called from Wikidata, as is done in Module:WikidataIB - cribbed from the way the French implemented it. It's built into the call so only shows when a value is returned. You then don't need a separate link at the bottom. Here's a demo adopted from {{Infobox book}}. I'm pretty sure it's the most functional solution; I'm equally sure that it won't be liked by those who are used to doing things the way they always have done. --RexxS (talk) 15:56, 11 August 2016 (UTC)

Due to the discussion at Wikipedia:Village pump (technical)#"edit on wikidata", other adminstemplate editors have looked to revert the link, so I've removed it while discussion continues. I disagree with removing it, since it follows the basic convention established by other Wikidata-enabled infoboxes. -- ferret (talk) 17:42, 11 August 2016 (UTC)

I don't see any admin looking to revert the link, just GeoffreyT2000, a template editor since 17 July 2016, taking three attempts to blindly revert the wrong change. I'm happy to wait a while for further discussion here, but I'll give notice that unless there is consensus not to have the link, I'll restore it in the near future. --RexxS (talk) 18:45, 11 August 2016 (UTC)
Noted, adjusted above. -- ferret (talk) 19:28, 11 August 2016 (UTC)
Why are we discouraging visitors from editing things on Wikidata? SharkD  Talk  01:48, 26 August 2016 (UTC)

Title[edit]

There is a title attribute. We should use it, where the language is English. For most video games, we might even able to fall back to "any claim made" before needing to fallback either to local definition or to the PAGENAME. Suggested order of priority:

  • Wikidata, in English
  • Wikidata, not in English
  • User definition
  • Pagename.

--Izno (talk) 13:29, 23 August 2016 (UTC)

Should we use a title property? Or just use the item label? Since Wikidata doesn't do "Fallout (video game)" as the label. Would need to add a "getLabel" to Module:Wikidata first. I'll craft one on the sandbox in a little while. -- ferret (talk) 13:38, 23 August 2016 (UTC)
Title is a bit safer/less prone to vandalism right now. Also, sometimes, the label doesn't match the English article (though I haven't noticed any cases where that's true of a video game, it's sometimes or even especially true w.r.t. editions of written works). --Izno (talk) 14:10, 23 August 2016 (UTC)
Ok will let this sit for a couple days to gather comment but seems uncontroversial. One problem as proposed, I believe Module:Wikidata will only return English strings. So it won't do the second bullet.. (Need to check...) Otherwise will add this weekend. -- ferret (talk) 14:13, 23 August 2016 (UTC)

@Izno: Module:Wikidata does not have an appropriate function for this logic. It would output all values of P1476 regardless of language. I will have to build something that will filter to just EN. Will be a bit longer on this one. I am not planning to try to do the second bullet, because I don't have a way to determine the best language to return if EN is missing but multiple others exist, other than "the first one found". -- ferret (talk) 18:50, 26 August 2016 (UTC)

Oh, no hurries, and not worries about the 'best language'. --Izno (talk) 19:10, 26 August 2016 (UTC)
When Module:Wikidata was written, there was no monolingual text, so there's no function to fetch it. Monolingual text is only useful if you have text available in different languages, so I'd have to ask why you'd want the title of a video game in some language other than English? Do we have infoboxes in articles for videogames that have no English title? Anyway, the function getImageLegend has the code to get whatever language is desired for the property media legend (P2096), so it should be simple to adapt for what you want, if there really is a need. --RexxS (talk) 21:23, 26 August 2016 (UTC)
Do we have infoboxes in articles for videogames that have no English title? Actually, yes, come to think of it. There are older titles and even the occasional new title that see release in e.g. Japan first and don't receive an English name (except either as a literal translation or an unreliable-sourceable not-quite-so-literal translation) that aren't the actual titles of the games. I'm thinking Japanese predominantly... --Izno (talk) 21:43, 26 August 2016 (UTC)
Okay, what are our articles called? Do you have an example? --RexxS (talk) 02:01, 27 August 2016 (UTC)

Links[edit]

Is there a point to "wikilink" people that has no page? In the case of Dishonored 2, It seems unhelpful to do so, if the wikidata page is only going to tell me that he is a game designer. It looks rather clumsy, having [*] all over the place. There is another issue with it. In the case of Tom Clancy's The Division, it is essentially cheating to list the game engine, since we decided not to list them unless they have a page. In addition, is there a way we can apply plainlist or something similar into this wikidata stuff? It looks terrible, since there are so many names mentioned together in one line. As a whole these wikidata stuff does not look good, especially with that giant icon (Edit on wikidata). I would prefer having a small [edit] icon right next to the top of the infobox, or a small "E" in the top left or top right corner, similar to how the navigation template talk page is handled. AdrianGamer (talk) 12:58, 25 August 2016 (UTC)

The output is generated by Module:Wikidata. Please use the talk page there to discuss changes to the formatting. It affects all Wikidata enabled infoboxes on enwiki. There is already a discussion to change the [*] to a link icon for clarity. -- ferret (talk) 13:02, 25 August 2016 (UTC)
Regarding an "Edit on Wikidata" link, it is currently removed entirely. I am opposed to the small "E" at the top because the typical convention is used to link to editing the template itself. Infoboxes suppress the display of it currently. Regarding the engine, specify |engine= to suppress the Wikidata display. This is the accepted RFC backed convention on enwiki. -- ferret (talk) 13:05, 25 August 2016 (UTC)
Didn't know it was so simple. Thanks for replying. AdrianGamer (talk) 13:16, 25 August 2016 (UTC)
@AdrianGamer: Please let me know if WP:VG/WD needs further work. It's new and I'm trying to make sure it gets seen. Izno did a pass over it, but if there's any details you or anyone think is lacking, please let me know. I added an example of suppressing engine. -- ferret (talk) 13:19, 25 August 2016 (UTC)

Other templates[edit]

Will be taking a look at the complexity for Template:Infobox video game series soon. There are several others that need work as well. -- ferret (talk) 12:29, 25 August 2016 (UTC)

Remove wikidata edit links[edit]

I can't imagine how a discussion on this tiny corner of the wiki was allowed to make changes to almost every VG article without some sort of announcement that contributors like myself would see. Instead, the echo chamber went ahead with this change with no input from the hoi polo, with a total of what, four people involved? It went from proposal to in-flight in a total of five days. It was only mentioned to the wider VG audience after it was turned on.

In any event, the change sucks. It adds whitespace to every VG infobox. That would be fine if it did something useful, but it doesn't. Quite the opposite, I can't imagine why anyone would want to visit another project to edit data that is actually primarily on this project. If this were information that was being transcluded back, like images from the Commons, that would be one thing. But in this case it's the opposite direction, there's literally no reason to ever want to edit there. And even then, Commons transclusion first links to a page here on en.wiki, specifically to avoid problems like this.

Turn it off and send it into the VG for a wider vote. This is a major violation of the trust of your fellow editors, and flies in the face of every policy and ad hoc rule we have specifically to prevent small groups from wielding undue power by hiding in corners and having private discussions.

Maury Markowitz (talk) 12:00, 11 August 2016 (UTC)

@Maury Markowitz: Here are some replies:
  1. This has been advertised multiple times prior to turning on, and been socialized with (at least) WT:VG since 2013. That that's the only instance you were able to find doesn't mean other instances don't exist.
  2. Adding whitespace is not categorically an issue, much less is it, in this particular case, an instance where I agree with you. However, I have just added a class to the template ("wikidata-link") which you can use to hide the edit link; you can add
    tr.wikidata-link {display:none}
    
    to your User:Maury Markowitz/common.css to remove the link, everywhere on Wikipedia.
  3. I have suggested, just above, that the link to edit on Wikidata should be removed where the infobox is pulling no data from Wikidata. Does that satisfy your 'usefulness' concern?
  4. The point is to get us to the Commons model. I can't point to a "big" article that uses the infobox in this way currently, but ferret can. Air Control (video game) uses it on a small level.
  5. Commons transclusion first links to a page here on en.wiki, specifically to avoid problems like this. is not quite the true state of things, but that's an aside. There is work being done to make it easier to edit Wikidata from the client wikis (such as en.WP).
  6. Please see the rather large RFC at WP:Requests for comment/Wikidata Phase 2. In no way is this a case of a small group wielding undue power by hiding in corners and having private discussions. Please do not make personal attacks on the editors who have contributed to this discussion.
--Izno (talk) 12:43, 11 August 2016 (UTC)

The discussion you link to has nothing to do with VG, and is three years old. I don't think it's a stretch to suggest that this does not count as a public airing of the discussion taking place here, three years later. Adding to the issue, that discussion concludes "this modification should be done carefully and deliberately, at least at first." Is this an example of careful and deliberate? I would argue no.

You suggest that there has been consensus, or at least reasonable discussion, in VG. As far as I can tell this is simply not the case. Over the last six months or so the topic has only come up in passing, with perhaps a dozen short posts in total, and nothing remotely like a full discussion. A somewhat lengthier discussion took place in VGR, but that's another lesser-traveled area and the discussion is by the same people, discussing technical issues. In total there's perhaps a half dozen people that have posted more than a couple of sentences on the topic, and the most recent of those is someone asking how to turn it off.

Sadly, finding any such discussion is extremely difficult, because the header now includes the post-facto notification string, so every single page in the talk area shows up if you search. So if I'm missing any sort of lengthy RFC decision making process that took place, mea culpa, please point me to anything recent on the topic in the main pages.

As to the suggestions: no, editing my display template is not a suitable answer for a problem that others are also complaining about, and no, removing the link if it's empty is not a solution to the problem. This entire concept is problematic, from transclusion of data the editor does not want and can't control, to all sorts of V issues, to the fact that there is little confidence in the data.

Maury Markowitz (talk) 16:12, 11 August 2016 (UTC)

Required to mention: I'm leaving for two weeks starting this afternoon. Maury Markowitz (talk) 16:13, 11 August 2016 (UTC)
I think I'll just disengage with you if your opinion is that the entire concept is problematic. --Izno (talk) 16:19, 11 August 2016 (UTC)
Have to agree with Izno. It does not seem your issue is about this infobox in particular, which follows established convention of other wikidata enabled infoboxes, but with Wikidata as a concept in general. The RFC Izno linked established the consensus as far as that goes. -- ferret (talk) 17:40, 11 August 2016 (UTC)
Izno, I tried using the above code on my css page to remove the link, and it hasn't worked. The link is still appearing. Any suggestions? Cheers. Bertaut (talk) 22:00, 11 August 2016 (UTC)
Bertaut The link is currently removed entirely pending discussion. If you're still seeing it, try purging the page in question. -- ferret (talk) 22:05, 11 August 2016 (UTC)
Thanks Ferret, that's done the job. Much appreciated. Bertaut (talk) 22:20, 11 August 2016 (UTC)

Add abbreviation template to WW label for world wide release date[edit]

Would it be possible to add the {{Abbreviation}} template to the {{Video game release}} WW so that it appears as WW? Maybe even do this for all the country abbreviations, kind of like what {{Drugbox}} does for the legal status of drugs in different countries (see the infobox at Amphetamine for example). I'm not sure everyone reading the article is going to instantly understand what the abbreviations mean. Sizeofint (talk) 05:26, 16 August 2016 (UTC)

I've added it for "worldwide" to Template:Video game release and Template:Video game release new as that is uncontroversial.

I'm not sure the links need them, as each of those carries a title attribute providing the title of the page linked. --Izno (talk) 11:25, 16 August 2016 (UTC)

Thanks! Somehow I completely missed that the other country names were linked. Sizeofint (talk) 19:06, 16 August 2016 (UTC)

ISBN and OCLC numbers[edit]

I am using WorldCat, and I am seeing that video games also have OCLC numbers as well as ISBN numbers (but why ISBN? They are not even books!). That information may be useful to add to the template, or it may not be, but who knows? Gamingforfun365 (talk) 18:53, 17 August 2016 (UTC)

This information is available from Wikidata, should this field be added to the infobox. However I do not think it has any real value to us and should be left out. -- ferret (talk) 14:13, 19 August 2016 (UTC)
Not in the infobox. OCLC numbers are interesting, I suppose, but should be in the external links (either using {{authority control}} or otherwise). --Izno (talk) 13:13, 23 August 2016 (UTC)