Index templates

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 was moved here from WT:MED. Please continue the discussion here. WhatamIdoing (talk) 01:43, 30 September 2015 (UTC)Reply

Original post

edit

Some time ago, I came upon Nervous system, which at the time had an insane collection of navboxes - three containers with a total of 10 navboxes, each of which had a complex structure. With the help of Tryptofish, I narrowed the list down to two. But each of these templates, in addition to a large number of its own links, transcludes two "Indexes" to more templates. This was my first hint of a broader problem.

The articles in this WikiProject are in an advanced state of template creep and need surgery. I have never seen such a maze of templates before. Indeed, medical articles have a monopoly of the so-called index templates, also called subnavs, which are footer templates that provide links between templates instead of articles (see Category:Medicine navigational box footer templates).

I would like to explore a recent example of the supposed usefulness of these subnavs provided by at a recent TfD:

  • Case study 1: A person is diagnosed with heart disease and wants to know more about treatment and relevant anatomy. The subnavs provide easy links.
To see how this works, I go to Heart disease, which redirects to Cardiovascular disease. At the bottom of the page, I see four navboxes. If I expand the second ({{Heart diseases}}), I see a three-layered hierarchy of link groups. Seeing something like that, my inclination would be to abandon the navbox and try a search (see below). But suppose I carry on, wading through dozens of terms like Monckeberg's arteriosclerosis, and eventually make it down to a box titled "Index of the circulatory system" (which is {{Vascular navs}}). And there I see a link to anatomy and a group devoted to treatment. If I click on "Anatomy" I get another template, {{Arteries and veins}}, with another two-level hierarchy and {{Vascular navs}} transcluded at the bottom. At which point only the fear of death would keep me from giving up.

But is that how someone is likely to look for this information? The first thing I would try is searching Cardiovascular disease for "anatomy" and "treatment". If I search "anatomy" and the navboxes are in their normal, hidden state (because there are four of them), I get no hits. So how about "treatment"? Now here is something really strange. The word appears nine times in the article - and none of them are linked to anything, despite the existence of Management of heart failure. There is a section, Cardiovascular disease#Management, with a single sentence in it: "Cardiovascular disease is treatable with initial treatment primarily focused on diet and lifestyle interventions."

Anyone who really wants to help that person diagnosed with heart disease would do something like this: Expand Cardiovascular disease#Management into a decent summary of Management of heart failure and link to it with a {{main}} template. Or, if that is not general enough, create Management of cardiovascular disease and summarize it.

So much for "easy links". But it goes much further than that. Say we click on one of the less general links in {{Heart diseases}}, like Mitral valve. This is an anatomy article, and does not itself transclude {{Heart diseases}}. But it does transclude {{Heart anatomy}}, which transcludes {{Heart navs}}. So this anatomy article has access to several templates for drugs. How likely is it that a user will find this useful?

Wikipedia provides guidelines that are intended to make navboxes useful. It states,

Navigation templates are particularly useful for a small, well-defined group of articles; templates with a large numbers of links are not forbidden, but can appear overly busy and be hard to read and use. Good templates generally follow some of these guidelines:

  1. All articles within a template relate to a single, coherent subject.
  2. The subject of the template should be mentioned in every article.
  3. The articles should refer to each other, to a reasonable extent.
  4. There should be a Wikipedia article on the subject of the template.
  5. If not for the navigation template, an editor would be inclined to link many of these articles in the See also sections of the articles.

But what is the subject of {{Heart navs}} - Heart or Heart navboxes? As for the other guidelines, how would you even check them? Over 750 articles transclude {{Heart navs}}, but it wasn't placed directly on any of them. However, it is pretty obvious that any given article will refer only to a very small fraction of the others. When you add a subnav to a template, you are, in effect, giving up on the guidelines.

The guidelines also suggest, "If the collection of articles does not meet these tests, that indicates that the articles are loosely related, and a list or category may be more appropriate." I am going to make a proposal here rather than at a TfD, because I would like to give editors the chance to cooperatively brainstorm a solution.

Proposal: Delete all of the templates in Category:Medicine navigational box footer templates, or replace them by lists or categories. RockMagnetist(talk) 21:27, 9 September 2015 (UTC)Reply

Clarification: Replace some of the templates in Category:Medicine navigational box footer templates by lists or categories, delete the rest. RockMagnetist(talk) 22:46, 9 September 2015 (UTC)Reply

so..you do seem to have a point, however something "in between" 1. deleting all the templates (in Med[1]) and 2.replace with lists might be more balanced (having said that, I believe more opinions would be best)--Ozzie10aaaa (talk) 21:47, 9 September 2015 (UTC)Reply
If you mean, delete some and replace others by lists, that sounds reasonable (I wasn't really implying an all or nothing scenario). Some templates might convert into good categories or lists, others may not be suitable. RockMagnetist(talk) 22:26, 9 September 2015 (UTC)Reply
exactly what I meant, this might be better,(but we need more opinions)--Ozzie10aaaa (talk) 22:31, 9 September 2015 (UTC)Reply
Rock, this may seem a little indirect, but I can offer you a complete and immediate, if superficial, solution to the problem. Just go to https://en.m.wikipedia.org/wiki/Heart_disease and you will see the the navbox problem is completely removed. Also, if you are reaching the age at which presbyopia becomes an issue, then you may find that you like that display anyway. Quite a few readers using that even on desktop systems. And, on a more serious note, if anyone's hoping that readers will use navboxes to reach important, related articles, you should note that navboxes are completely invisible to 30% of our readers. WhatamIdoing (talk) 22:46, 9 September 2015 (UTC)Reply
Thanks, but I wear glasses. What do you mean by invisible? RockMagnetist(talk) 22:49, 9 September 2015 (UTC)Reply
I mean "not able to be seen". Follow the link, scroll to the bottom, and see what you see. Navboxes are not displayed on the Mobile site, and that's about 30% of our page views. WhatamIdoing (talk) 21:15, 10 September 2015 (UTC)Reply
Naturally, if navboxes are useless, we could delete them all. But I believe they are useful if done right. RockMagnetist(talk) 22:55, 9 September 2015 (UTC)Reply

Ping LT910001 who I think is/ has worked on these. Matthew Ferguson (talk) 02:28, 10 September 2015 (UTC)Reply

RockMagnetist might help if you did a mock up by way of demonstration of what you are proposing. Matthew Ferguson (talk) 02:30, 10 September 2015 (UTC)Reply

O.k., here is a demo where I replace {{Ear navs}} by Outline of the ear. Mostly the links in this demo are to articles, but "Inner ear" links to Outline of the inner ear. There would also be outlines of the outer and middle ears, but the rest of the links would continue to be to articles. This outline cuts across the templates listed by {{Ear navs}}, many of which have a group for the inner ear. Links to these outlines would have to be added to the See also section of articles. This would increase the likelihood of it only being added where it is useful. RockMagnetist(talk) 04:16, 10 September 2015 (UTC)Reply

Thanks for your gentle introduction to this topic, RockMagnetist. Links between medical templates have existed since at least 2008, and recently with some other users I attempted to make these templates somewhat more useable, with a goal to reduce the (huge) amount of navboxes plastered on our many articles by linking them .

Part of your frustration with this template series seems to relate to the incomprehensible jargon we use to describe medical articles and medical diagnosis. This is a frustrating fact of life and with any luck will improve within the next 200-300 years as we move away from Latin and Greek-based names.

For these "indexes", I use these links occasionally myself but having been editing them I am not too impartial about their use. I would love to hear from any editors who have used these to good end. A question that hasn't been asked at any point in the last 6+ years is are these templates actually useful. A few long-term editors agreed last TFD but I am really unsure what other users think and if actual readers use them.

If we get rid of them we should consider removing them completely, as in my experience in the dusty corners of Wikipedia outlines are either forgotten for many years and unusable for that reason, or excessively edited with links from all and sundry that they are also unusable. On the other hand I fear if we remove these links we may end up with users adding more and more navboxes again. Hoping to hear from other users, --Tom (LT) (talk) 09:06, 10 September 2015 (UTC)Reply

@Tom (LT): Thanks for keeping an open mind on this - always hard when something you have worked on is challenged. I hope that anyone who has found them useful will provide some details. In the TfD, you were the only one who did. Some people will go to a deletion debate and more or less equate "navbox" with "useful navigation tool." But is it as useful as it could be? RockMagnetist(talk) 15:52, 10 September 2015 (UTC)Reply
As for delete vs. outline, I lean towards delete; I just provided the demo as an option to consider. Also, I don't think the way to prevent users from adding navboxes is to add meta-navboxes; that just changes the baseline for template creep, and navboxes get added anyway, as the Nervous system example demonstrated. RockMagnetist(talk) 15:57, 10 September 2015 (UTC)Reply

I think that considering specific examples will make it easier to think about these templates. So I will do two more case studies from the TfD. I will describe my experience as I try the exercises:

  • Case study 2: A student has an exam on the nervous system. The subnav provides links to all the relevant anatomical templates for the said area.
    The first thing that occurs to me is, if I were facing an exam, would I rely on Wikipedia instead of a textbook? But, pressing on, I go to Nervous system. I look at the table of contents and see Anatomy in vertebrates. This section has a link at the top to List of nerves of the human body, and it has a discussion in prose of how the nervous system is organized. Possibly good enough already. Now I go to the bottom of the page, and see {{Nervous system}}. This has a nice, compact list of articles on the main parts of the nervous system. Below it are {{Central nervous system navs}} and {{Peripheral nervous system navs}}, with links to templates on many of the same subjects. However, if I want to know more about the meninges, I have already clicked on the article link. I know that meninges is different from meninges; but that would never occur to a first-time user. RockMagnetist(talk) 16:18, 10 September 2015 (UTC)Reply
  • Case study 3: A lay reader wants to know about what bones we have in our body, but don't know their names. The templates provide easy links to said bones by body part.
    I start at bone. A quick scan gives me the impression that this article is mainly about the structure, development and so on of individual bones. I try human anatomy, which redirects me to human body. Browsing this, I see a link to human skeleton (now why didn't I think of that immediately?). This has the section Divisions, with a few lines each for Axial skeleton and Appendicular skeleton. Now, I must admit, this is a bit slow, so I try going to the bottom. I find a box, "Bones of the human skeleton", with 7 templates inside of it, all but one of them massive and multi-leveled, and all of them transcluding {{Bone and cartilage navs}}, which in turn links to all of them. Just think of the reduction in complexity if {{Bone and cartilage navs}} were here instead of all those templates! But {{Bones of skeleton}} without any subnav might be even better.
At this point it occurs to me that I went to the navboxes earlier than usual because I am discussing them. Normally, I would look at See also first. So I go there, and what do I find? List of bones of the human skeleton. RockMagnetist(talk) 16:45, 10 September 2015 (UTC)Reply
I must be crazy, because I don't find this system terribly difficult to navigate, and these little user vignettes read to me like scripts for the wiki-equivalent of those late-night infomercials involving someone who sets their kitchen on fire trying to boil water. I don't object to simplifying it - it's certainly unusual to have meta-templates that navigate between other templates - but converting to outlines doesn't really appeal to me; those things are un-watched and un-edited backwaters, and the vertical-ness makes them harder to read than the condensed navbox format. Opabinia regalis (talk) 17:10, 10 September 2015 (UTC)Reply
No doubt the more you use this system, the better you get at using them - and the less you need them. Each of the indexes has a doc page that says "These are geared towards lay readers". I am an appropriate person to test them on - lots of experience with Wikipedia, but not a lot of medical knowledge. And the thing to note about my case studies is that I can get the information I'm looking for very quickly if I ignore the navboxes. Hardly setting the kitchen on fire. RockMagnetist(talk) 17:24, 10 September 2015 (UTC)Reply
Well, I don't really use this system often - I work on molecular biology articles; anything as big as a bone is beyond me ;) - though I did look into it when one of these was TfD'd awhile ago. No doubt the navbox pileup at the bottom of an article like human skeleton is a problem, but I'm not sure what the best solution is. Separate "outline"-type articles with navbox-y formatting maybe. Opabinia regalis (talk) 17:43, 10 September 2015 (UTC)Reply
The truth is, the above scenarios are how I would search for information if I started within Wikipedia (because otherwise navboxes couldn't possibly help me). But that's not how I really look for information. I would go to Google and do the following (by case number):
  1. Search heart disease treatment. No Wikipedia pages in sight, so maybe I would go to the Mayo Clinic page.
  2. Search anatomy of nervous system. No Wikipedia pages in sight, so just for the heck of it I restrict it to en.wikipedia.org. The first hit is for Nervous system, and I follow "Jump to anatomy in vertebrates", which lands me at the section by that name, where the link to List of nerves of the human body is at the top.
  3. Search bones in the human body. The top link is to List of bones of the human skeleton.
That was 30 seconds or less for each search. I think the moral of this story is, structure the articles well and you won't need navboxes at all. RockMagnetist(talk) 17:44, 10 September 2015 (UTC)Reply
On the question of usefulness: They are doubtless "useful" for de-WP:ORPHANing articles. They may be used significantly by a small number of editors. Whether anyone else uses them is unclear. It seems unlikely that typical readers use them very much. As I said above, they're systematically stripped out of the mobile version. You can't see them, and therefore you can't use them (even if you go to the navbox's page in the template namespace) there. Therefore we can guarantee that 30% of readers aren't using them, and despite these being invisible on the mobile site for a very long time, I have never heard a complaint about their absence. WhatamIdoing (talk) 21:39, 10 September 2015 (UTC)Reply
So what is your recommendation? Get rid of all the templates? Leave them as they are because it doesn't matter? Or something else? RockMagnetist(talk) 21:49, 10 September 2015 (UTC)Reply

@RockMagnetist I think you are conflating three issues, all quite relevant:

  1. We do not name articles in a way that is easy for readers to understand
  2. We use way too many navboxes in our articles
  3. Our navboxes are overburned by these internal link boxes.

Issues (1) and (2) won't be solved in this discussion but will be by editing away -- and I've opened a thread at the talk page of Wikiproject Anatomy to see what we can do there, also. With regard to these internal navboxes, what would you think about moving them to template space or template documentation? You may a good point as to why they may not be used by readers, but they are supremely useful in part because they help keep track of our project's numerous templates, ensuring that we don't start to have duplicate ones. So one option would be to move them all so they only display in template namespace. This will mean readers are no longer confused. An alternate option would be to include all of the 'index' templates as collapsed, so that only a line "Index of..." is displayed and users can open it at will. --Tom (LT) (talk) 22:59, 11 September 2015 (UTC)Reply

Conflating these issues is hard to avoid in a case study. I think the naming is not a big problem because the redirects work well. Article content may be more of an issue (more below). Absolutely you use too many navboxes and the navboxes have too many links. I'm not sure what you mean by moving the internal navboxes to template space, since they are already there; but moving them to documentation pages is a good idea. RockMagnetist(talk) 23:55, 11 September 2015 (UTC)Reply
And now I would like to say something positive about the internal navboxes. They are well organized; I like the division into description, disease and treatment. Some of the articles like Cardiovascular disease would benefit from closer alignment with their content. And, converted into regular templates linking to articles, they would be an improvement over some of the existing regular templates. RockMagnetist(talk) 23:55, 11 September 2015 (UTC)Reply
While a navbox may appear redundant to a category or list, it is not the same as a single category, a group of categories under one parent category, or a list. Each section of a navbox may seem to be a carbon copy of a category or list, but that is not always the case, and in many cases, the navbox "categories" do not reflect Wikipedia's custom for creating and organizing categories or for lists.,,,The success of navboxes can be seen in this (uncontrolled) study of this navbox, which has shown that in the month following its creation, readership of the articles contained within increased by 8.5% (an average of 406 views per article) and editing of these pages increased by 37% from the month prior to its creation [2]--Ozzie10aaaa (talk) 22:44, 16 September 2015 (UTC)Reply
We already agree that some form of navbox is useful. RockMagnetist(talk) 02:40, 17 September 2015 (UTC)Reply

Reading through the above, I'm not convinced you've really shown a convincing reason or consensus for a change. I've minimised the extraneous comments in your case studies to illustrate this. I think unfortunately most of your frustration emerges from the way we've structured our articles. I just don't see a way to solve the problem you pose (how to keep navboxes small and compact in an up-to-date way) without using a system like this to ensure that the templates are compartmentalised. Perhaps "Index of..." could be renamed to be even more clear, or a small (?) question mark could be provided with an explanation for lay users to explain what these subnavs do.--Tom (LT) (talk) 10:21, 23 September 2015 (UTC)Reply

  • Case study 1: A person is diagnosed with heart disease and wants to know more about treatment and relevant anatomy. The subnavs provide easy links.
To see how this works, I go to Heart disease, which redirects to Cardiovascular disease. At the bottom of the page, I see four navboxes. If I expand the second ({{Heart diseases}}), I see a three-layered hierarchy of link groups. Seeing something like that, my inclination would be to abandon the navbox and try a search (see below). But suppose I carry on, wading through dozens of terms like Monckeberg's arteriosclerosis, and eventually make it down to a box titled "Index of the circulatory system" (which is {{Vascular navs}}). And there I see a link to anatomy and a group devoted to treatment. If I click on "Anatomy" I get another template, {{Arteries and veins}}, with another two-level hierarchy and {{Vascular navs}} transcluded at the bottom. At which point only the fear of death would keep me from giving up.
  • Case study 2: A student has an exam on the nervous system. The subnav provides links to all the relevant anatomical templates for the said area.
    The first thing that occurs to me is, if I were facing an exam, would I rely on Wikipedia instead of a textbook? But, pressing on, I go to Nervous system. I look at the table of contents and see Anatomy in vertebrates. This section has a link at the top to List of nerves of the human body, and it has a discussion in prose of how the nervous system is organized. Possibly good enough already. Now I go to the bottom of the page, and see {{Nervous system}}. This has a nice, compact list of articles on the main parts of the nervous system. Below it are {{Central nervous system navs}} and {{Peripheral nervous system navs}}, with links to templates on many of the same subjects. However, if I want to know more about the meninges, I have already clicked on the article link. I know that meninges is different from meninges; but that would never occur to a first-time user. RockMagnetist(talk) 16:18, 10 September 2015 (UTC)Reply
  • Case study 3: A lay reader wants to know about what bones we have in our body, but don't know their names. The templates provide easy links to said bones by body part.
    I start at bone. A quick scan gives me the impression that this article is mainly about the structure, development and so on of individual bones. I try human anatomy, which redirects me to human body. Browsing this, I see a link to human skeleton (now why didn't I think of that immediately?). This has the section Divisions, with a few lines each for Axial skeleton and Appendicular skeleton. Now, I must admit, this is a bit slow, so I try going to the bottom. I find a box, "Bones of the human skeleton", with 7 templates inside of it, all but one of them massive and multi-leveled, and all of them transcluding {{Bone and cartilage navs}}, which in turn links to all of them. Just think of the reduction in complexity if {{Bone and cartilage navs}} were here instead of all those templates! But {{Bones of skeleton}} without any subnav might be even better.
So basically you're saying, skip the article and the regular navboxes and go straight to the index boxes? All the rest is "extraneous"? I think you're having a hard time visualizing how a first-time user might approach this page because you have been developing these indices for so long. One view of how regular users behave is to look at page views, because if you use a link in an index box, you actually visit a navbox. In the last 30 days, 29,049 readers visited Cardiovascular disease, 492 visited {{Arteries and veins}}. And the latter is linked to over 50 other navboxes through {{Vascular navs}}. One of the links in {{Arteries and veins}} is to Venous plexus. This is a stub with no references, yet it got 781 hits in the last 30 days. Unfortunately, I know of no way to determine how often a regular navbox is used because readers can click one of its links without visiting it. But clearly it's the index boxes that are extraneous. RockMagnetist(talk) 16:19, 24 September 2015 (UTC)Reply
On second thought, I can get some idea of how often the regular navboxes are used because a lot of them duplicate lists. For example, {{Heart diseases}} duplicates a lot of List of ICD-9 codes 390–459: diseases of the circulatory system, which got 4430 hits in the last 30 days, and List of ICD-9 codes 760–779: certain conditions originating in the perinatal period, which got 2606 hits. And they got all of these hits even though most of the links to them are through navboxes. RockMagnetist(talk) 16:51, 24 September 2015 (UTC)Reply
Actually I'm equally confused about what you're saying, RockMagnetist. You identify that index boxes are confusing. Yet you also lump together a whole lot of other concerns about they way our articles are linked. My point is that removing index boxes won't fix any of those other concerns. So by fixing index boxes (or removing them) we get rid of any of the problems in clear text. However the preceding problems (which I have struck through) remain. --Tom (LT) (talk) 23:43, 26 September 2015 (UTC)Reply
These things you refer to as "conflating" and "lumping together" are what I call "context". The purpose of all those numbers is to compare the use of an index box with the use of links from the article itself and in regular navboxes. The answer is not definitive, but it does seem to suggest that index boxes are used very little compared to the alternatives. However, I think it is worth exploring the usefulness of certain groups of regular navboxes as well. RockMagnetist(talk) 01:09, 27 September 2015 (UTC)Reply

Here is another statistic that might be easier to understand: {{Vascular navs}} links to 32 navboxes. Collectively, in the past 30 days, they were visited 13,856 times.[1] The index box appears in about 1600 articles. Thus, it was used an average of 9 times per article over 30 days.[2] At the other end, I would estimate the navboxes have a total of about 1000 links, so on average each target article was accessed about 13 times. Is it really necessary to gather more information? RockMagnetist(talk) 02:12, 27 September 2015 (UTC)Reply

References

  1. ^ for an average of 433 per navbox (the min was 142 and the max was 676 - a remarkably tight grouping).
  2. ^ What we don't know is how many of the people who arrived at the navboxes actually made use of them

Note that the above only provides information on access to the navboxes through the index template. The navboxes themselves might be useful because they are transcluded in each of the articles that they are linked to. That is the kind of thing that could be studied by the method described below. RockMagnetist(talk) 07:09, 27 September 2015 (UTC)Reply

  • Comment. RockMagnetist This is quite late, but I'll add my 2cents. I am not into medicine, but a year ago with LT910001 (Tom) I edited those Mednavs (the Index subtemplates) from weird ad hoc illegible abbreviations into readable text. This very step was useful, and at the same time we recognised it was not enough. There was something more about these Mednavs. And RockMagnetist now points to this.
I have not digested this talk section completely because of TL;DR. But my thoughts are:
Bad thing 1: they are Navboxes on an article page, but thay link to template space. In short, links in articles must link to content space (articles, content categories).
Bad thing 2: They deviate way too much from the parent navbox topic at hand. IOW, they menting topics that are not directly linked to the article-topic, and so not the parent navbox topic (as RockMagnetist already noted). Below, I'll add a proposal sketch. -DePiep (talk) 10:17, 16 December 2015 (UTC)Reply

Decision point

edit

Okay, guys: How serious about this are we? I see some comments here that may degenerate into my-link-is-better-than-your-link, but the fact is that nobody has the overall data.

There is a kludge that will let us see how often links in navboxes get clicked on, if we really want the information. It won't let us separate search engine spiders from real humans, but it will let us get page view counts for clicking a link in a navbox vs other ways of reaching the page. It's not elegant, and it's got some small WP:PERF costs, so we shouldn't do this on a grand scale, but we could certainly try it out for a small number of links in one or two navboxes.

Is that interesting to you? More to the point, is it interesting enough to two or three editors to do the work involved? It'll take an hour to pick templates and links, an hour to set it up, a few minutes each day to check the links for problems/potential corruption, and (after a few weeks or so) an hour to revert the whole setup, and an hour or two to collect the data and write it up. Whether anybody uses these navboxes would actually be a useful thing for WPMED to know, and the results would probably be more or less applicable to all subject areas at the English Wikipedia. But it takes sustained work over the course of a month, and there's no point in starting if the project won't be finished. WhatamIdoing (talk) 19:10, 24 September 2015 (UTC)Reply

I like that. It would be interesting to gather some information. I can't guarantee that I could be involved the whole way - next month, real life will be pretty demanding for me. I think the experiment should be carefully designed, with links to closely related and more remote articles. Here is one possible set:
One article that is mentioned frequently in Cardiovascular disease: e.g., Vascular disease
One that is not mentioned as such but is in at least one regular navbox: e.g., Mitral stenosis
One that is not linked in the article or any regular navbox: e.g., Superficial vein
For a bonus, we might want to look at connections between these lower-level articles.
A possible control experiment: replace any templates that duplicate lists by links to the lists in ‘’See also’’:
RockMagnetist(talk) 05:22, 25 September 2015 (UTC)Reply
There's no rush. If somebody else volunteers to help out (always good to have more than one editor involved, because real life happens) and nobody objects, then we could start by picking a couple of navboxes and targets, which will save us some time when we actually want to do it. We could make a list of things to test and how to do it at a page like WP:WikiProject Medicine/Navboxes.
I kind of like the idea of using an actual random number generator to pick which links we'll test in the navbox. We can change some links and not others. Does that sound good to you? WhatamIdoing (talk) 14:25, 25 September 2015 (UTC)Reply
Yes, a separate page sounds like a good idea. I think random choices are good in principle, but some thought still needs to go in the implementation. I think it's best to have controls in the experiment, which is why I suggested comparing different ways of getting between the same pair of articles. Also, since the purpose of index boxes (as I see it) is to provide links that are not in the article or regular navboxes, we should compare links from each of the three groups I described above. Another control experiment is my idea of replacing navboxes by lists and monitoring them for the same length of time as the navboxes. RockMagnetist(talk) 14:57, 25 September 2015 (UTC)Reply
I'd be very interested in a data-driven approach that may have implications for our navboxes in general. Template_talk:Medicine_navs is a place we've been centralising discussion about this and may be a suitable hub while we investigate what impact (if any) these index boxes have. --Tom (LT) (talk) 23:43, 26 September 2015 (UTC)Reply
It's not very active. I tried to ask a question about eponymous categories and got zero response. RockMagnetist(talk) 01:02, 27 September 2015 (UTC)Reply
[3] yes its a little slow--Ozzie10aaaa (talk) 09:15, 27 September 2015 (UTC)Reply
It might be slow, but it'll be higher traffic than a brand-new page. Why don't we copy this whole thing over there, and make some plans? We can always post reminders here about it if we need extra input. WhatamIdoing (talk) 15:23, 27 September 2015 (UTC)Reply
Good point. I think it should be prefaced by a set of proposals, so we know what we're trying to find out. I will do that in a new subsection. RockMagnetist(talk) 15:53, 27 September 2015 (UTC)Reply

Proposals

edit

Here are my suggestions for how to fix the template creep in the articles covered by this project. They go beyond my initial point about index templates, but that is necessary because the index templates themselves were designed to fix template creep. I think the suggestions below are consistent with the principles behind good navbox design.

  1. Delete the index templates or move them to project space.
  2. Before deleting each index template, examine each navbox that they link to, and do two things:
    1. Eliminate excessive levels of hierarchy (for example, in {{Arteries of head and neck}}, go no deeper than the link to Maxillary artery).
    2. Make sure that each navbox is transcluded on the remaining pages that it links to and only those pages.
  3. If there are still a lot of boxes on some pages, for each page see if the article subject is closely related to the other subjects in each navbox (i.e., the articles are likely to link to each other directly). If not, take them off the template.

I would also recommend removing those templates that duplicate lists of ICD-9 codes and ICD-10 codes and replace them by links to those lists in See also. In general, these templates are long and intricate, and the information is better presented as a list. RockMagnetist(talk) 16:09, 27 September 2015 (UTC)Reply

Please do not introduce here an ICD-9 and ICD-10 discussion. Different topicv = diff thread. Complex enough! -DePiep (talk) 10:52, 16 December 2015 (UTC)Reply
RockMagnetist On considering the first point (during my sadly mostly internet-less week :( ) I would support a move of the index templates to project or documentation space, they are simply too useful when editing to get lost in deletion. --Tom (LT) (talk) 09:52, 3 October 2015 (UTC)Reply
As to your other points, I haven't formed an opinion yet. I would point out that this involves something like 700 templates transcluded on 32,000 articles 40,000 (including anatomy and non-med articles) so although I think you make some good points I am not sure, having recently spent a month going through the med templates, how this is going to be done short of a concerted effort by multiple editors over many months.
Something we have slowly been doing is moving the ICD codes in the navbox titles to wikidata, and then removing them from the titles, making it easier for readers to use.
So to summarise I'm having a little trouble visualising how we will actually go about this mammoth task you propose. I do not want to launch into any sweeping action without thinking about it in more detail. --Tom (LT) (talk) 09:52, 3 October 2015 (UTC)Reply
As to cleaning up templates, I fully agree. Please see Wikipedia:WikiProject_Anatomy/Open_Tasks#Cleanup_templates. I have made posts at (from memory) WP:DENISTRY, WP:PHYSIOLOGY, WP:MED and WP:ANATOMY with lists of the templates I believed to be most in need of cleanup in each project space, identified during my editing sweep. So far it has been a mostly lonely process for myself only :(. --Tom (LT) (talk) 09:54, 3 October 2015 (UTC)Reply
Sorry to hear about the loneliness of it. I have had a taste of it myself during this process. The plus side is that if we agree on something, we could just do it and not wait for more consensus.RockMagnetist(talk) 17:09, 3 October 2015 (UTC)Reply
I agree that it looks like a huge job, but I don't see this as something that needs to be done all at once. Each change is good by itself; we could just choose something to do and do it. Here is what I suggest for the first stage: remove links to the index templates and wait a couple of weeks. If no one objects, move them to wikiproject space. If someone objects to that, start a TfD to move them on the basis that they are not used anywhere. RockMagnetist(talk) 17:09, 3 October 2015 (UTC)Reply
The links can probably be removed using WP:AWB. RockMagnetist(talk) 21:52, 4 October 2015 (UTC)Reply
I think that this should wait until we've determined whether anyone clicks on the links. WhatamIdoing (talk) 17:01, 8 October 2015 (UTC)Reply
See below. RockMagnetist(talk) 17:31, 8 October 2015 (UTC)Reply

Proposal: Index articles

edit

My approach. Without fully having read the analysis TL;DR, I just drop a suggestion.

First step: copy all navs (as they are today) to Project space. Sure they will disappear or change heavily, while they are useful for editors in teh background. For this, they are quite stable & OK by now.
  • Intro. An index template gives an overview only of its topic (a metatopic, a higher level that the article itself is; that's why it does not belong on that article page by navbox philosophy). What we could do is:
Create an article for each Index, that gives the overview the Index box wants to show. (sort of example: ATC code A01, within the whole ATC code structure).
  • Think these working steps for bones:
1. Create Draft:Index of bone and cartilage in clinical medicine  Y
2. Add its (old) index template {{Bone and cartilage navs}} as working source  Y
3. Create three main sections: Description, Disease, Treatment  Y
4. Add all of its mentioned templates (navboxes), add a subsection name for each (so there should be by now:
==Desctiption== ===Anatomy=== {{Bone and cartilage navs}}) (partly  Y: level * only, not level **)
  • Milestone: We have a draft overview of all topics related. Enjoy.
But these are still navboxes, and not article content. We can't have an article content being navboxes! So we'll remove the navbox wrapper.
11. Choose a level of detail. MAybe we don't need each article from the navboxes mentioned in the Overview (i.e., maybe not all links in the list).
12. Remove the {{Navbox}} outer template
13. Remove all articles that are too detailed
14. Rebuild the structure per section into plain, indented list ({{Tree list}}?). Add level 4 ==== headers or indents to support the structure, where needed.
  • Milestone: This overview page should basically be fit for Article space.
21. In all the parent templates that now have the child {{Bone and cartilage}}, remove the Index child and add a link to the new Overview page.
In numbers:
- New Clinical Index Overview articles: 41 (One per Index child template)
- Navbox templates to be converted into article treee lists: 774??? (~all those now in the Index navs?)
- Templates to be edited: 774 (replace Index template with link to Overview article).
- Articles affected: 40000 (will show a change). -DePiep (talk) 10:52, 16 December 2015 (UTC)Reply
OK, thanks for your input DePiep, I support at least creating a full draft to see how it will look. --Tom (LT) (talk) 20:16, 16 December 2015 (UTC)Reply
For this 1/41 demo, that's already 40 to 60 templates to convert into tree lists ;-). So please if you see some objecting, object early :-). -DePiep (talk) 08:12, 17 December 2015 (UTC)Reply

Proposal: make index boxes only visible in template namespace

edit

@DePiep, RockMagnetist while we are experimenting with what to do in terms of drafts, I propose concurrently that we move all these index boxes into template space by moving them into the "noinclude" second of each template.

This can be done at the same time as other editing, and means they won't be displayed to end users. They will however retain their useful purpose for template editing as I have previously explained. --Tom (LT) (talk) 20:22, 16 December 2015 (UTC)Reply

"move all these index boxes into template space" - I don't understand. {{Bone and cartilage navs}} is in template space. Anyway, as long as they are in articles (and content categories), they should be in template space.
What we can do without much discussion, is copy these 42 into WP:MED WikiProject space, so MED editors can still access them to have a more complete overview. The 'public' ones (still appearing in articles), may be edited or deleted by this discussion. -DePiep (talk) 08:03, 17 December 2015 (UTC)Reply
I'll make something nice & correct for this. No need to discuss this further (timesaver). -DePiep (talk) 08:08, 17 December 2015 (UTC)Reply
@DePiep my indended meaning is that like template documentation, the index boxes are moved to the "noinclude" section of templates. So any users who go into the template namespace to view the template itself can still see them, but normal users cannot. --Tom (LT) (talk) 22:33, 18 December 2015 (UTC)Reply
Getting it. Then, nothing of the navs information will be available for the reader (in content space) at all ever? Would be an easy solution (for us here), but I thought the info was useful in some situations (but not all 400000 articles). -DePiep (talk) 22:46, 18 December 2015 (UTC)Reply
@RockMagnetist, DePiep. Hope you both had a wonderful Christmas or other festive day recently. Over the past few months with work commitments I've been using Wikipedia again mainly for (self)educational purposes and have to agree with your original point, RockMagnetist, that these navs visually contribute to a lot of clutter and are rarely used. If a user wants to go to a certain article, they can already (as you state) use google, the internal WP search box, nav boxes, and wikilinks. So I have to state (and apologise for all the hoo-hah) but I have come around to your point of view. I fully support removing these subnavboxes from article space. I don't support deleting them though, because for a template editor they are exceedingly useful. So would you support making them visible only to template space? I have left a comment about index articles below so we can separate discussions. --Tom (LT) (talk) 07:31, 27 December 2015 (UTC)Reply
If you, medicine editors, think this info is not needed in content space, that's fine with me. We can leave them visible for WP:MED space only etc. Waiting for RockMagnetist opinion. No need to mention this at WT:MED btw? -DePiep (talk) 08:02, 27 December 2015 (UTC)Reply
I don't think so. We've already mentioned this discussion 3-4 times, so I think if editors were interested they would already be participating. If we all agree on something, we can roll out a change and then alter it according to the reaction. But I don't think it's normal for every peak and trough of a discussion to get mentioned at the relevant project. --Tom (LT) (talk) 22:15, 27 December 2015 (UTC)Reply
@DePiep, thank you for reviving this discussion; I had given up on it and moved on to other things. Sorry it has taken me so long to reply - this is always a very busy season for me. I still haven't managed to read all the recent discussions, but I will reply to this one. @Tom (LT): I'm glad that we're in agreement now. I think it would be fine to remove the subnavboxes from article space and keep them in template space. RockMagnetist(talk) 22:55, 28 December 2015 (UTC)Reply
 Y OK then. Done for {{Bone and cartilage navs}} [4]. So a parent template like {{Bone and cartilage}} looks different now.
Technical note: done correctly, the 42 navs must be removed from all parent templates (those 700+ med navboxes). The construction "show in templete space only" is possible, but too cumbersome. -DePiep (talk) 23:43, 28 December 2015 (UTC)Reply

Proposal: delete all index boxes completely

edit

I am creating this so we can at least centralise discussion. This option proposes that we simply delete all the navboxes. --Tom (LT) (talk) 20:22, 16 December 2015 (UTC)Reply

  • Oppose. I do not favour this option at all. The index boxes may serve a navigational value, although as RockMagnetist points out this is questionable. They definitely serve a very useful role organising our templates and articles, so I believe they should be preserved in some form. --Tom (LT) (talk) 20:22, 16 December 2015 (UTC)Reply
  • Oppose. First discuss which info should be used & retained for articles & categories (content space). -DePiep (talk) 08:10, 17 December 2015 (UTC)Reply
edit

On day one of removed templates, one has already been replaced with another series of links [5]. Would it be worth emulating this on all the navboxes, or providing a series of links (as you two mention above) in navboxes? I am personally not in favour of this in view of the discussions we've had above, but if other editors are going to do this piecemeal we should discuss it. Thoughts? --Tom (LT) (talk) 23:36, 1 January 2016 (UTC)Reply

Not discussed so no acceptable. The 61.62.95.113 IP seems to know about wp, but does not collaborate? Weird. -DePiep (talk) 23:42, 1 January 2016 (UTC)Reply
I reverted [6], new links not specific enough ('medicine' not 'bones' etc). -DePiep (talk) 23:48, 1 January 2016 (UTC)Reply

Move

edit

Pings so that you'll all know that this discussion has moved here. Please be liberal in pinging me (and other editors) for this discussion, since it's not the first page I visit every day (unlike WT:MED).

The description at User:Sebwite/navbox study is interesting, but old (pre-mobile), and User:Sebwite doesn't edit often any more, so we might not be able to get much information about it. Any other ideas? WhatamIdoing (talk) 01:43, 30 September 2015 (UTC)Reply

WhatamIdoing @RockMagnetist two simple ways of getting a rough idea of the utility of the index boxes are:
  1. Retrospectively - we changed the format of the index boxes over a short period (see the archives). Pick some medical articles, see if there are any difference before the date the index box changed & afterwards.
  2. Select two or three index boxes. For one month, every second week show the index box, and every alternate week hide it. Monitor viewership to see if there's any variation. --Tom (LT) (talk) 10:09, 3 October 2015 (UTC)Reply
I like the second idea, but could we do longer changes? It takes a while for changes in templates to propagate out to the articles, and this step is where we create (small) server costs. From what I overhear from the analytics folks, one week (any seven-day period) is the right way to think about changes like this. (At least with people who edit, the people who edit on Sundays are not the same as people who edit on Mondays.) So my suggestion is to do two-week sessions of "on" and "off", even if that means that it takes eight weeks to collect data. Does that sound okay? WhatamIdoing (talk) 19:41, 3 October 2015 (UTC)Reply
Tom, do you have good random number software handy? Maybe User:MastCell does, if you don't. I was thinking that we could use RNG to pick a handful of navboxes out of Category:Templates that use a Medicine navs subtemplate, and split them 50-50 into experimental and control. Does that sound like a good idea to you? WhatamIdoing (talk) 17:05, 8 October 2015 (UTC)Reply
@WhatamIdoing sounds like a good objective way to do this. I say we pick 6 navboxes randomly from the list using a RNG. The list is provided on the "Template" page here. I do not have access to a RNG however if you could generate the numbers I think this experiment will be quite interesting. --Tom (LT) (talk) 21:35, 9 October 2015 (UTC)Reply
Tom, should the six targets be chosen from among the 42 items in Template:Medicine navs#Listed? And is that six to edit, or three to change and three for the control? WhatamIdoing (talk) 22:22, 9 October 2015 (UTC)Reply
Three sounds good (to change and three to control) and quite implementable. Let me know what the results of your random numbers are and we can get started.--Tom (LT) (talk) 09:48, 11 October 2015 (UTC)Reply
I used this, which soupvector suggested at WT:MED. The first six numbers are: 12, 25, 38, 7, 40, 29, in that order.
That's Template:Enzyme navs(edit talk links history), Template:Myeloid navs(edit talk links history), and Template:Tooth navs(edit talk links history) for the first three ("to change") and Template:Congenital disease navs(edit talk links history), Template:Urinary system navs(edit talk links history), and Template:Olfaction navs(edit talk links history) for the second three ("to control").
What's the next step? WhatamIdoing (talk) 01:14, 20 October 2015 (UTC)Reply
It might help if you describe the method you have in mind. RockMagnetist(talk) 03:45, 20 October 2015 (UTC)Reply

Question

edit

Before doing this kind of study, we should clarify what we are trying to find out. I am going to repeat the essence of an earlier post because I think it may have been overlooked:

{{Vascular navs}} links to 32 navboxes. The traffic stats for these pages say that collectively, in the past 30 days, they were visited 13,856 times. The index box appears in about 1600 articles (obtained from What links here). Thus, it was visited an average of 9 times per article over 30 days. At the other end, I estimate the navboxes have a total of about 1000 links, so on average each target article was accessed about 13 times.

This is hard data because any time someone clicks on a link in {{Vascular navs}}, they visit a navbox. And it is an upper limit on their use because some people may have visited the navboxes by other routes, and they didn't necessarily use the links once they got there. This analysis is easy to repeat for any of the other index boxes. Even as an upper limit this is really low usage, so I don't see what more we need to know about index boxes; maybe we should refocus our investigations on the regular navboxes. RockMagnetist(talk) 17:31, 8 October 2015 (UTC)Reply

@RockMagnetist I agree it does not seem like these links are used much. I am however very intrigued about doing an experiment as it may lead us to make one of the few "evidence based" content decisions on Wikipedia. In addition it will provide useful information for the future of this idea (subnavboxes). In the meantime, why don't we start working through some of your other ideas and cleaning up some templates. I'll make a new entry below for this.--Tom (LT) (talk) 21:35, 9 October 2015 (UTC)Reply
I agree with the general idea of making changes based on evidence; but it's important to know what questions you are trying to answer and not ask questions that have already been answered. As far as I am concerned, we already have really good data on {{Vascular navs}}, and I can repeat the analysis for any other index box in about 10 minutes if you want me to. I have an average for 1600 articles over a month - what more do you hope to learn with 6 randomly chosen links? RockMagnetist(talk) 22:58, 11 October 2015 (UTC)Reply
LT910001, WhatamIdoing: Here is an example of something we don't know yet. Which is more effective in providing navigation when there are a lot of links - a multilevel navbox or a link to a list article in See also? RockMagnetist(talk) 22:58, 11 October 2015 (UTC)Reply
Here's what I want to know: "Does anyone actually click on the links in {{vascular navs}}?" NB that this is significantly different from "Does anyone view the page that happens to be linked in {{vascular navs}}?", which is the question that you have answered. WhatamIdoing (talk) 01:01, 20 October 2015 (UTC)Reply
It's different, but we know it has to be lower; and an upper limit of 9 visits a month is the same as zero for all practical purposes. RockMagnetist(talk)
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.

Conclusion: Remove from articles (February 2016)

edit
About Category:Medicine navigational box footer templates (41 index templates)
  • Altogether, this is the outcome:
The 41 templates are not article navigational help (let alone article content). They should be removed from article space.
They can be kept to support Wikiprojet:Medicine etcetera. As main {{Medicine navs}} can be kept for overview.
Practically, they should be removed from navigation boxes that appear in articles. -DePiep (talk) 00:01, 11 February 2016 (UTC)Reply
 Y processing. -DePiep (talk) 00:01, 11 February 2016 (UTC)Reply
Not touched: {{PharmNavFootnote}} (4000 transc's). -DePiep (talk) 00:44, 11 February 2016 (UTC)Reply
Thanks DePiep for your work :). If there were any way to keep them in template namespace only that'd be very useful, maybe including the list in the documentation page so the relevant nav opens up. Otherwise, au revoir metanavs. --Tom (LT) (talk) 00:53, 11 February 2016 (UTC)Reply
The navs templates themselves remain intact by this. They may be used in some WP:MED overview etc. When needed in template space, they should be put in the template documentation page, not in the parent template (with some switch). -DePiep (talk) 08:14, 11 February 2016 (UTC)Reply
I've marked the discussion as 'resolved'. --Tom (LT) (talk) 01:00, 11 February 2016 (UTC)Reply
@DePiep: I'm not clear on what you have done or what you mean by removing the index templates from article space. Mostly they are transcluded in other templates, and based on a small sample, those links seem mostly intact. Going to Category:Medicine navigational box footer templates, I chose the first entry, {{Bacteria navs}}, and clicked on the first three articles that linked to it. These led to two templates - {{Growth media}} and {{Immune sera and immunoglobulins}} - that still transclude this template. I will remove those two and see what happens. RockMagnetist(talk) 17:05, 11 February 2016 (UTC)Reply
As I said, the work is in process. It takes some AWB runs to catch every situation for the 41 templates. -DePiep (talk) 17:14, 11 February 2016 (UTC)Reply
Great! Thanks for doing this. RockMagnetist(talk) 17:19, 11 February 2016 (UTC)Reply
  Done. Todo: promote {{Medicine navs}} in WP:MED :-) -DePiep (talk) 19:24, 13 February 2016 (UTC)Reply
Stray navs (index templates) that happen to end in articles by accident, are listed in Category:Articles that show a Medicine navs template. This category should be empty. -DePiep (talk) 19:24, 13 February 2016 (UTC)Reply
@DePiep: I'm puzzled by this edit. Why is the new version so different from the old? --Redrose64 (talk) 00:09, 2 March 2016 (UTC)Reply
Reverted. A bug in AWBscript. Reported, fixed some days ago [7] (it sometimes mixed up content of two consecuetive (sp?) pages; so what you saw was the content of the page edited right before this one). -DePiep (talk) 06:25, 2 March 2016 (UTC) +@Redrose64: DePiep (talk) 07:05, 2 March 2016 (UTC)Reply
@DePiep: This didn't work, see WP:Echo#Triggering events. Luckily, I have this page watchlisted. --Redrose64 (talk) 10:48, 2 March 2016 (UTC)Reply

Medicine navs - update

edit

A few years ago i did a major update of the medicine navs with a.o @DePiep and @Tom (LT). I see there's been a nice cleanup since then. What's the current status? Anything an editor with a humble academic medical background can contribute for the navboxes at the moment? PizzaMan (♨♨) 20:36, 2 August 2017 (UTC)Reply

They are listed in Category:Medicine navigational box footer templates (0). Point is that they are not fit to be used in articles. That is, because they link to templates, not to other articles. They do not help to navigate articles. The discussion was on this page (see above). This removal from articles was done one year after we had replaced the abbreviations and codes into real words. Anyway, they can only be used outside of articles (Helppage, WikiProjects etc.). -DePiep (talk) 20:43, 2 August 2017 (UTC)Reply
Ah, i see. Sorry i was too lazy to read up on the talk page when i saw all the archives. Too bad for all the effort we put in, but an understandable choice. Thank you for updating me. Any other general medicine or neurology related navboxes in dire need of fixing that you know of? PizzaMan (♨♨) 21:06, 2 August 2017 (UTC)Reply
Good to see you around again, PizzaMan. I'm sure I can dust something off the to-do shelf :). --Tom (LT) (talk) 22:50, 3 August 2017 (UTC)Reply
Likewise :-) Please let me know what you find on the shelf. PizzaMan (♨♨) 19:29, 4 August 2017 (UTC)Reply

Usage instructions for Template:Bone and cartilage navs

edit

Nearly all of the pages that link to Template:Bone and cartilage navs are themselves templates. And they seem to link to it without any problems. One non-Template page that links to it is Draft:Bone Malrotation, where the Wikitext {{Template:Bone and cartilage navs}} generates two link errors, a Stripped tags for </div> and a missing end tag for <div>. The problem doesn't go away by removing "Template:". Template:Bone_and_cartilage_navs/doc says, "This is an embedded medical navbox." If that means that it should not be used anywhere except in another navbox, will someone please edit the doc file and say this, so users and editors will be less likely to use it outside navboxes and more easily diagnose misuse. If it is permitted to use this template outside of navboxes, then why is it causing lint errors at Draft:Bone Malrotation? —Anomalocaris (talk) 05:28, 6 June 2019 (UTC)Reply

Update: I edited Draft:Bone Malrotation, changing

{{Bone and cartilage navs}}

to

<div>{{Bone and cartilage navs}}</div>

and this fixed the lint errors, but my main request remains. —Anomalocaris (talk) 21:54, 7 July 2019 (UTC)Reply