Template talk:Asbox/Archive 3

This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 1

Archive 2

Archive 3

Archive 4

Microformats in stubs

Latest comment: 14 years ago70 comments11 people in discussion

To facilitate the inclusion of microformats in instances of this template, please change:

<table class="metadata plainlinks stub"

to:

<table class="metadata plainlinks stub {{bodyclass}}"

allowing an extra class value to be specified when the template is deployed. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 22:41, 9 August 2009 (UTC)

What sort of class name would be passed in? --- RockMFR 23:21, 9 August 2009 (UTC)

The parent class-name of a microformat; for example, vcard if the hCard microformat is to be used; vevent for hCalendar and (most likely based on recent manual conversions) adr for an address or place-name - for example, in Template: Ghana-stub; removing the need for one pair of span tags. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 23:28, 9 August 2009 (UTC)

I don't know much about microformats, but is it useful to place them inside the stub templates where you have to assume that it will be removed without replacement at some point? I'd think it would be better to have a bot go through all of Category:Ghana stubs and place a permanent {{Locationdata}} or something (similar to {{Persondata}}). Amalthea 11:14, 10 August 2009 (UTC)

Yes, it will be useful, with or without Locationdata; conversely, it will do no harm. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 13:34, 10 August 2009 (UTC)

I don't want to be an ass, but can you try to explain how it is useful? As I said, I don't know much about them, but I think that a) it's much more logical to add more permanent meta data that doesn't need to be re-added in a different way once an article advances to Start class, and b) it's a bad idea to hardcode the span ids "" in the first place, just as it was a bad idea to hardcode the stub template wikicode with classes and ids in thousands of stub templates in the first place. Amalthea 15:31, 10 August 2009 (UTC)

It will be useful because it will add semantic meaning to instances of the template about the types of subject I listed above (hCards for people, events, etc.); at no extra overhead for editors or readers. Can you say why you think it's a bad idea? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 19:59, 10 August 2009 (UTC)

Because when the article graduates from stub, the class will be lost, and the editor who removes the stub tag might not realize he has to add it to the article proper. The same goes for hardcoded non-stub categories in place on a handful of stub templates. –xeno^talk 20:04, 10 August 2009 (UTC)

The editor who removes the stub notice may never need to replace the class - it may, for example, be included in an infobox. But if not, how is thart more harmful than it never having been there at all? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 20:36, 10 August 2009 (UTC)

Can that create problems if there are several, possibly conflicting countries? Or is the format designed to handle that?
But of course, the names of the classes should in any case be abstracted away in a dedicated template, so there's no need to pass the class as a parameter. Amalthea 20:42, 10 August 2009 (UTC)

I foresee no problems around "conflicting" countries; and I regret taht I don't understand the rest of your post, Remember, though, that microformats are used un this way in hundreds of thoisands (possiblfy millions) of instances of thousands of infoboxes and navboxes. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 21:20, 10 August 2009 (UTC)

(ec)I believe I have:

Don't hardcode class names in hundreds or thousands of stub templates, it's tedious to update
Don't place information that isn't related to the stub status inside the stub templates since it will need to be re-added once it's advanced to Start or above.

Instead, have a bot monitor Category:Ghana stubs et al to add a permanent {{Locationdata}} template. Or add this process to the default fixes of AWB.
I can see that it's convenient to automagically add the desired meta information to thousands of articles by (ab)using the stub templates, but I think it's a short-lived joy, and should be done properly instead. Amalthea 20:42, 10 August 2009 (UTC)

You have not explained how including class names in templates will be "tedious to update"; nor even why you think they will need to be updated. You assert, without basis, that "information that isn't related to the stub status […] will need to be re-added once it's advanced to Start or above". Your hypothetical Locationdata template will not impart semantic meaning, nor serve our users, in the same way that the currently-available microformats would do so; and does not apply to the types of article I list above, which are not about locations. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 20:55, 10 August 2009 (UTC)

No answer? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 22:48, 31 August 2009 (UTC)

No. I don't particularly appreciate a discussion style where it's challenged that updating a classname in thousands of templates might be considered more tedious (wikt:tedious: boring, monotonous, time consuming) than updating one. Amalthea 12:14, 10 September 2009 (UTC)

I haven't challenged that. I have challenged that that will ever need to be done. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 15:21, 10 September 2009 (UTC)

I agree that adding permanent metadata to a transient template is a Very Bad Idea; it will just result in data either being duplicated, or lost. (also)Happy‑melon 19:47, 10 August 2009 (UTC)

If the template is transitory, then so will be the metadata it emits. Can you say why you think it's a bad idea; and how you think metadata will either be lost? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 19:59, 10 August 2009 (UTC)

Dunkwa-on-Offin is a city in Ghana. Dunkwa-on-Offin will still be a city in Ghana whether or not it has {{Ghana-stub}} on its article. If we add location metadata to the stub template, then there will be metadata on the Dunkwa-on-Offin article for the time being. 99.99% of readers will not be aware that the metadata is present; 99% of the remaining 0.01% will not understand where it is being added from. When the article is improved beyond Stub class, the stub template will be removed. The metadata will be lost, because even of the tiny number of people who were aware that it was ever there, who is going to suspect that it was being added by a completely unrelated template? That is a Bad Thing. To prevent that happening, the metadata must be encoded in a different template, either a permanent one like an infobox, or in a bespoke template such as the {{locationdata}} that has been proposed above. To avoid loss of data, that separate template needs to be added before the stub template is removed. At which point there will be duplication of the metadata. So why bother adding it to the stub template in the first place? (also)Happy‑melon 12:16, 11 August 2009 (UTC)

Can you explain how this will be used, in layman's terms? –xeno^talk 20:57, 10 August 2009 (UTC)

Suppose we have two stub notices (or navboxes, or whatever), "xxxx about Foobar". By using either Foobar or Foobar, a parser (a bot, or a tool operated by a human user) can tell whether the word "Foobar" refers to the vernacular name of a living thing, or a place name. This, for example, improves searchability. It also allows the term to be passed to a relevant search engine (such as a biological database or a gazetteer) thus making it easier for the editor improving the stub to find relevant information. For further information I suggest reading microformat and WP:UF. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 21:12, 10 August 2009 (UTC)

I too think this is a bad idea. A stub is basically a maintenance template with a categorization scheme, that just happens to have some contextual information. It is however not part of the article proper (unlike an infobox). It also adds further complexity to the stub templates, which might be confusing to some of our users. —TheDJ (talk • contribs) 21:56, 10 August 2009 (UTC)

Even if a stub is not, as you say, "an article proper" (that's debatable) that is immaterial; and no reason not to semantically identify the meaning of key terms used in a stub template. Since this will be a simple, optional parameter, your "confusion" claim is unsubstantiated, and unjustified. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 22:11, 10 August 2009 (UTC)

I know that the WSS project expressed concerns that the metatemplate would become too parameter-laden. Thank you for the links you provided above, I'll take a look. –xeno^talk 23:56, 10 August 2009 (UTC)

That's a slippery slope argument. Aren't such arguments discounted by policy? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 14:57, 11 August 2009 (UTC)

No answer? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 22:48, 31 August 2009 (UTC)

The example that seems to be being thrown around here is {{Ghana-stub}}, and the proposal to add locationdata for Ghana to it. If you look at Category:Ghana stubs, we see that the use of this template is by no means restricted to geographical articles: is such a tag appropriate on Elections in Ghana?? How about Abron?? These situations are not very clear-cut. (also)Happy‑melon 12:16, 11 August 2009 (UTC)

On every page on which {{Ghana-stub}} appears, the word "Ghana" means "the county Ghana"; it is semantically correct and meaningful to mark it up as such with an ADR microformat. How coudl it not be so? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 14:57, 11 August 2009 (UTC)

But the purpose of the metadata is to apply to the article, no? Of course "the word Ghana" means the country; that's not AFAICT the point of the microformat. The purpose of the code Ghana is to say "the subject of this article is located in Ghana", no?? That is simply not consistently the case with all Ghana stubs, as many of them do not have a definable geographic location. What am I missing? (also)Happy‑melon 19:52, 11 August 2009 (UTC)

No. I don't know where you get the impression that the purpose of the code is to indicate that the subject of the article; but that's not the case. It's perfectly valid to have multiple microformats on one page, describing different things. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 22:56, 11 August 2009 (UTC)

Disabling edit request since there is not consensus for this change. --- RockMFR 23:13, 13 August 2009 (UTC)

I have addressed every concern raised; many of which were founded in - often admitted - ignorance of the purpose and nature of microformats. What do you think I should do to satisfy any remaining concerns? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 23:27, 13 August 2009 (UTC)

Have you run this by the stub sorting project? I still don't see how this microformat business is supposed to work; do we need to tag every instance of a country name with the microformat? Is it just search engines that read this microformats? Is it worth the effort? –xeno^talk 23:31, 13 August 2009 (UTC)

This won't affect the work of the stub-sorting project at all. Do you really think it worth bothering them? Yes, I think every instance of a country name (at least, using this and other similar templates) should be marked up, semantically, with a microformat. no, it is not just search engines which use microformats, as I've already said, above. And yes, this is worth the - very tiny amount of - effort (far, far, less effort, in fact, than this debate has taken!). Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 15:38, 14 August 2009 (UTC)

As they are the ones that maintain the over 10,000 stub templates that may require these microformats, yes, I think you should run it by them. –xeno^talk 15:44, 14 August 2009 (UTC)

I might also suggest a discussion at a wider venue to determine if Wikipedia is going to embrace these microformats. –xeno^talk 15:49, 14 August 2009 (UTC)

That discussion took place about two-three years ago; Wikipedia now has thousands of templates emitting hundreds of thousands, if not millions, of microformats. Wikipedia's microformats are already parsed by Google and Yahoo, among others. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 16:33, 14 August 2009 (UTC)

I've reinstated {{editprotected}} since, so far as I can see, discussion has ceased and all concerns have either been addressed or refuted. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 18:43, 17 August 2009 (UTC)

I still think it's pointless. I'll ask for feedback on the Village Pump. —TheDJ (talk • contribs) 19:51, 17 August 2009 (UTC)

Either that, or people have just given up discussing it with you.
FWIW, I'm still strongly against hardcoding any class names in thousands of templates. I have no strong opinion about placing microformat templates into the stub templates, but find that this too is the wrong approach. Amalthea 21:20, 17 August 2009 (UTC)

If people choose to absent themselves from a discussion, then they can hardly expect their views to be taken into account. You say you're " strongly against hardcoding any class names", but without reasons, such a statement (and things such as "I still think it's pointless") sound like WP:IDONTLIKEIT. If you can give reasons, I'd be happy to either address them or concede. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 09:47, 18 August 2009 (UTC)

It looks as though people have already expressed their reasons further up. When so many people oppose your idea it's time to start questioning it yourself. — Martin (MSGJ · talk) 09:52, 18 August 2009 (UTC)

Which reason do you think has been given, which has not already been addressed or refuted? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 13:15, 18 August 2009 (UTC)

I liked this proposal more when I didn't fully understand it. What is the point in doubling the bytecount of the sentence to add 'semantic meaning' to a word that only has one meaning in the first place? And you propose to (eventually) do this to every single word that represents a location on wikipedia? Yikes. But even if this process has consensus, why would this particular request facilitate this? You don't want to add the microformat to the whole template. The {{Ghana-stub}} template is not located in Ghana, nor does it represent a location. The word "Ghana" represents a location; that's what you need to wrap in the microformat. Why can't you just do that in the |text= parameter? (also)Happy‑melon 12:39, 19 August 2009 (UTC)

I can and do do it in the |text= parameter; the change requested above will streamline this and will make it easier for others to subsequently edit the template. The requested change does not equate the whole template in a microformat; it adds a parent class so that the specific word may be defined. Many words which are place names ("Bath", "Telford", "Wall", "Barry") do not uniquely mean that; and microformats are not just for placenames, but also for people, venues, events, species, products, recipes/ food items and more. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 13:33, 19 August 2009 (UTC)

To summarise the summary, stub templates are things we aren't supposed to have. Their (purported) function is self destructive - to get the article made non-stubby. They attempt this by two methods - a visible presence which is an invitation to expand the article (or to remove the template), and categorization. Because of the nature of stubs, stub templates (and categories) are being created and destroyed all the time - often "out of process" tsk-tsk. In particular they are made by comparative newbies to fill a ral or perceived need - generally by cut and paste. Adding a MF parameter will result in many many incorrectly MF'd stubs. If you want to make the meaning from the stub template available in MF a much stronger idea is to file a bug to have categories emit the MF, and make it heritable, therefore you would simply need to label

Africa geography stubs
Airport stubs
Americas geography stubs
Antarctica geography stubs
Asia geography stubs
Europe geography stubs
Fictional location stubs
Marine geography stubs
Oceania geography stubs
Protected area stubs

and you would have thousands of categories emitting location codes. Moreover the same would work on the main category tree. Rich Farmbrough, 19:04, 25 August 2009 (UTC).

The fact that stubs are intended to be transitory is no reason not to have them emit a microformat. Adding a microformat will increase the visibility of a stub's presence (and thereby increase the probability of it being improved). Raising a bug to have categories emit microformat might work (though I can see flaws, such as needing to add classes to parts of words in anchor links), but will take months if not years to achieve. Adding microformats via stub templates will work now. You say that "Adding a MF parameter will result in many many incorrectly MF'd stubs", but I don't see why that would be, and you don't substantiate the assertion. If it is the case, then surely categories would be similarly corrupted? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 20:29, 25 August 2009 (UTC)

Actually the categories are probably worse, because they need a massive overhaul. But that said the stub tree is relatively clean. Rich Farmbrough, 21:13, 31 August 2009 (UTC).

So where does this leave us? I note some of my questions, above, remain unanswered. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 22:48, 31 August 2009 (UTC)

Section break

Discussion has stalled; and there has been no apparent input from VP. Unless anyone has any unresolved concerns, I intend to reactivate the edit request. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 11:23, 10 September 2009 (UTC)

I don't believe you have to any completeness explained why implementing these microformats requires increasing the complexity of the templates by adding extra parameters, when the classes can just as easily (and arguably more transparently) be applied directly to the text input. Happy‑melon 11:51, 10 September 2009 (UTC)

Yes, my concerns are unresolved. We've only just managed to clean all stub templates up, by adding the {{asbox}} abstraction layer, and removed the hardcoded inconsistencies from thousands of templates. You want to change the stub template with the express intent to add hardcoded meta class names back in. I've stated two alternatives that wouldn't require that. Using a bot to add persistent meta data has no drawback I can see except for a higher initial effort, and the advantage to retain the meta data past stub status. Using a template has no drawback I can see except for a slight performance hit, and the advantage to be more flexible. The latter alternative would have the exact same end result you're looking for.
No, I don't immediately know when a {{locationdata}} template might be useful. That's part of the point of abstraction: you leave yourselves options that you wouldn't have if you hardcoded or substed everything. During the Asbox Stub Unification Drive nobody had all potential benefits in mind. You are currently trying to make use of that added abstraction layer, yet refuse to see that abstraction with the location meta data might have future benefits as well. Talk about WP:IDONTLIKEIT.
Amalthea 12:14, 10 September 2009 (UTC)

←[reply to both] Addressing your points in turn:

"increasing the complexity": Adding one optional parameter (singular, not plural as suggested above) to this template, to reduce the need for adding inline markup to many, is hardly unnecessary complication. Surely that's what templates are for?
"classes can just as easily … be applied directly to the text input. I've often been told, elsewhere on WP, that including classes as template parameters (e.g. bodyclass="vcard"), rather than in-line HTML (Foobar)is preferable, both technically, and for ease of updating by users who are not HTML proficient. Why do you think the latter is preferable in this case, but not others?
"…arguably more transparently: I fail to see how what I propose is any less transparent; indeed I think it more so (because it will be included in this template's documentation), than in-line mark-up.
"hardcoded inconsistencies: I fail to see how those inconsistencies are relevant here; indeed, adding a parameter centrally ensures greater consistency. Perhaps you care confusing the straightforward change I propose, with such inconsistencies, and that is giving you needless cause for concern?
"Using a bot to add persistent meta data: Where have you initiated this programme? What text do you think the bot will mark up as a microformat? Can we see some examples in use? Without such answers, this seems to me a hypothetical red herring.
"{{locationdata}} Another hypothetical red herring. And as I have already pointed out, my requested change does not apply only to lcoations, but also to taxa, organisations (including bands and other such groups of people) venues, events, foodstuffs and more. You likened {{locationdata}} to {{persondata}}; Note that {{persondata}} does not emit microformats and is often used alongside templates which do.
"IDONTLIKEIT: not IDONTLIKEIT; it's NOTPOSSIBLENOW and quite possibly NEVERGOINGTOHAPPEN. On the other hand, once the requested change is made, it will be possible to make existing instances of stub templates emit microformat metadata with no "higher initial effort" of edits to every single one of the pages using them (again, that's what templates are for), and no performance hit. In the unlikely event that one of your hypothetical solutions does come to pass, then this edit can be undone at a stroke thereby disabling all the microformat emitted by all such stubs, without the need to manually remove them from hundreds.

I really don't see what you are worried about. Have either of you used microformats emitted by stub templates, or examined any of those currently configured, manually, to emit them? Perhaps we need additional outside opinions? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 15:12, 10 September 2009 (UTC)

I'm developing an increasing sense of "meh" towards this proposal; I find I'm increasingly objecting to the principle of using this convoluted method of attaching semantics to the phrases than I am to the actual implementation of them. What happens when you want to identify different objects within the same template? "This article about Ghanan plants in the taxon Quok", for instance? Happy‑melon 16:26, 10 September 2009 (UTC)

I don't think we actually have any stubs like that, do we? But in that case, bodyclass=hcard biota. And this proposal is less convoluted than the pure-markup method you advocated above. I asked some Qs of you, above, BTW. But let me put it anther way: what would it take, to satisfy you that this us a worthwhile change? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 17:36, 10 September 2009 (UTC)

I think you have convinced me that, taking the assumption that it is constructive and worthwhile to add these microformats in this convoluted fashion, the extra parameter is the way to go about it. I don't take that assumption; I'm sure that there is a better way of achieving these ends, but I really don't have the enthusiasm to pursue it. I'm glad that you do. Happy‑melon 18:53, 10 September 2009 (UTC)

This proposal is a bad idea and this has been extensively explained above. Continuing this discussion beyond this point and actively asking more people to participate sounds like a big waste of time and efforts. Cacycle (talk) 22:18, 10 September 2009 (UTC)

There may have been discussion, but it's been neither extensive (by WP standards) nor conclusive. If you have a specific objection, not yet addressed, please state it. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 11:16, 12 September 2009 (UTC)

It seems that there is already consensus for "not adding" the new parameter, except obviously Andy. Locos epraix ~ Beastepraix 02:51, 12 September 2009 (UTC)

Please address my specific responses to your concern, just above. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 11:16, 12 September 2009 (UTC)

I'm mostly concerned about Amalthea stated above. The purpose of this template is to remove inconsistencies for stub templates, just as {{ambox}}, {{tmbox}} and brothers did when created. So with the new parameter, there will be a big amount of work to add microformats to stub templates that require them. Why? because people who use microformats will realize that these templates at the bottom of the pages exist to add "x" microformat to the page instead of tagging the page as "an article deemed too short to provide encyclopedic coverage of a subject" (from WP:STUB). Logically, there will be stub templates with microformats and others without them, that oddness will only create confusion. Locos epraix ~ Beastepraix 14:07, 12 September 2009 (UTC)

Well, of course I can't explicitly refute your concerns: because they're hypothetical, and based on a worse-case-scenario at that. So why don't we make the change, and revert it in the unlikely event that your concerns are realised? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 14:13, 12 September 2009 (UTC)

So may we now proceed? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 12:57, 18 September 2009 (UTC)

But, I haven't seen anybody else defending this proposal, so... are you (Andy) going to add microformats to the zillions of geographical stub templates that we have? If that's the case, the new parameter is (again) unnecessary because you stated that it is going to help other editors to add microformats and I don't see anybody else adding them. Is it too hard to use a tag? Locos epraix ~ Beastepraix 18:54, 18 September 2009 (UTC)

I'm going to add them where I can and invite others, not least through proper documentation and by asking projects to do so. The batteries in my crystal ball are flat, so I can't speak for what other people will do; but that's exactly the point that was made when I asked for the necessary edits be made to infoboxes to make it possible to edit them to them emit microformats; now plenty of other editors do that too , and we emit a million or more. No, it's not "too hard" to use a tag, but as I've already said above, I've been told that it's better to use template mark-up than to do so. It is too hard to make the requested change? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 19:17, 18 September 2009 (UTC)

Is it better to use template markup? that will only complicate the use of this template for everyone. The change won't be fulfilled until there is consensus and obviously this is not the case. Locos epraix ~ Beastepraix 00:07, 20 September 2009 (UTC)

You assert that using template mark-up will "complicate the use of this template for everyone". How so? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 19:15, 20 September 2009 (UTC)

Well? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 10:19, 26 September 2009 (UTC)

No response, again, so edit protected request reactivated. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 19:57, 3 October 2009 (UTC)

Not done: please establish a consensus for this alteration before using the {{edit protected}} template. — Jake Wartenberg 00:54, 4 October 2009 (UTC)