Youtube

Go to The Main Page Add Youtube to favorite!

Wikipedia talk:Summary style 

Contents

Transclusion, an interesting tool.

Earlier, I participated in this talk page to a discussion on how to remove duplication from Wikipedia. Since then, I found a neat tool called transclusion to publish the same text in more than one page. Now, here is my crazy idea: a summary article could be made up of a collection of lead paragraphs transcluded from the corresponding main articles. Thus, it will always be up to date and editors attempting to edit it will be sent to the main article. Emmanuelm (talk) 18:44, 17 April 2008 (UTC)

Can you point to a particular summary page plus main page where that would be a good idea? I like the general idea of not duplicating material, but it would depend on the page. - Dan (talk) 19:03, 17 April 2008 (UTC)
Coming soon. So far, I've only used this tool to host the same discussion in several talk pages. Emmanuelm (talk) 12:50, 18 April 2008 (UTC)
Dan, I just completed my first transclusion. The main article is Indoor bonsai. Its seemingly ordinary lead paragraph is in fact transcluded from the subpage Indoor bonsai/lead. The same text is also transcluded as a section in Bonsai and Houseplant, the two summary articles. All modifications to the transcluded text will be shown in all locations. Do you like it? Emmanuelm (talk) 14:24, 18 April 2008 (UTC)
Update: User:Discospinster bluntly deleted within minutes the subpage Indoor bonsai/lead, judged to be a "test page". There was no warning and she did not discuss it in my talk page. Turns out subpages of main articles are forbidden, for no good reason. I will fight this but I know too well that Gods/editors are immune to attacks from simple mortals. Emmanuelm (talk) 18:15, 18 April 2008 (UTC)
It is not that subpages of articles are forbidden; it is that they quite techincally are not enabled. Could you do this thing by using <onlyinclude>...</onlyinclude>around the lead instead? — the Sidhekin (talk) 18:22, 18 April 2008 (UTC)
Sidhekin, I'll skip the argument about subpages, it is discussed here. I want to thank you for your suggestion of the includeonly tags. For some reason, these tags do not work as advertised but the noinclude tags work well.
I added such tags around the lead paragraph of Indoor bonsai and transcluded the whole page into Bonsai and Houseplant. It seems to work fine and, frankly, is faster than creating a subpage. There is, however, an extra limitation: all recipient pages will show the same text. With subpages, I could have created several transclusion schemes from the same page. Emmanuelm (talk) 19:22, 18 April 2008 (UTC)
Not includeonly — it's onlyinclude (includeonly is something else). I keep making that mistake myself; I reckon they made a poor nomenclature choice for these tags.  :-) — the Sidhekin (talk) 19:29, 18 April 2008 (UTC)
I'll keep an open mind toward this approach for the time being. In the example being offered here, it seems to work reasonably well, at least visually. I think it's not very editor-friendly, however: "What are these 'noincludes' doing here?" "How do I edit the text in this section? All it has is two things inside braces."
It also causes somewhat screwy formatting: in Houseplant#Indoor_bonsai, "bonsai" is bolded because it's bolded per the MoS on the indoor bonsai page, and houseplant ends up getting bolded as well because that's how MediaWiki formats self-links.--Father Goose (talk) 22:44, 18 April 2008 (UTC)
I am in favor of limited deployment of this concept. I think some editors will be concerned that this will lead to laziness (simply sticking your lead into other articles without thinking carefully about rewording as appropriate), but I don't think that's a valid objection; just don't allow that. The most important point, it seems to me, is that more eyeballs make better text; if people are seeing the same text several places, and if edits happen in each place at the same time, then the text will improve faster and degrade less.
There are several reasons that, if it isn't awkward, it's better to have exactly the same text in both places, rather than updating and tweaking the text separately. When the two texts are allowed to contradict, the credibility of both is pulled down. It's also much more work to update the text both places and continually have to check the two texts against each other.
I'm wondering why people didn't try this before. You should bring this up some place that more people watch, like WT:MoS. - Dan Dank55 (talk) 23:31, 18 April 2008 (UTC)
This is an interesting idea and I'm also tentatively in favour (or at least open minded) towards limited deployment: I think it is worth trying it out in a range of articles and see how it goes. The idea also encourages good lead writing: the lead is supposed to summarize the article and provide a self-contained overview, which is exactly what a summary style section is supposed to do.
Concerning technical issues, it is safer to put an "onlyinclude" around the lead, rather than "noinclude" the rest of the article, because someone could add categories or interwikis after the final "noinclude". Concerning "how to edit this section?", this could be handled by using an "includeonly" to transclude the section heading and the {{main}} template. Then "edit this section" would take the editor to the transcluded article. I'll demonstrate that in a moment. Other editing issues could be handled by adding suitable comments to the source. Geometry guy 10:56, 19 April 2008 (UTC)
This turned out to be much harder to do than I thought, essentially because lead sections don't have a section heading to link to. The best I can do can be found at {{spinout}} and I've demonstrated it on Bonsai and Houseplant: the problem is that the edit section link is in the wrong place, and I can't get it in the right place without adding it to the table of contents. I've also created {{spinout/link}} and {{spinout/title}} to deal with the bolding problems noted by Father Goose. They are demonstrated at Indoor bonsai. Geometry guy 14:09, 19 April 2008 (UTC)
GeometryGuy, thanks for the info about spinout; I am learning a lot here! And I agree with those that pointed out that this approach forces us to write good lead section; a well-written lead is a perfect summary. So, who among you will modify the Summary Style guideline to explain this?
Still, I maintain that the subpages approach would be better because it offers more flexibility. Specifically, it would allow the transclusion of different parts of a main article to different summary articles, something the templates cannot do. As I said before, main article subpages are arbitarily forbidden (which, Sidhekin, is closer to the truth than "not enabled"). You may want to discuss this here. Emmanuelm (talk) 14:52, 19 April 2008 (UTC)
You didn't know about {{spinout}}, as I only created it today (the template was previously a redirect to {{subarticle}})! Concerning flexibility, are you suggesting that it would be useful to transclude a section of an article other than the lead section into a summary article? I can't think of very many situations in which that would be a good idea. However, it is possible to achieve this using templates: just wrap the section in "onlyinclude" tags. If different parts are to be transcluded onto different pages, then this can be done by using #ifeq or #switch to test {{PAGENAME}}. Geometry guy 15:10, 19 April 2008 (UTC)
Grin. I was just coming over here to say that if you want to do this right, you should talk with Geometry Guy :) - Dan Dank55 (talk) 16:10, 19 April 2008 (UTC)

A general issue: I can see a case where the lede of a subarticle might be right for that article, but too long for the desired summary in the parent article. I doubt we should try to address cases like those via technical means. But even if the {{spinout}} approach proves to work well only in some cases (it remains to be seen if it is a good idea in any), I at least like the prospect of having it as an option.--Father Goose (talk) 03:12, 20 April 2008 (UTC)

Absolutely. This will crash and burn (at least as a proposal) if there's a pattern of people using it when a word-for-word copy is not appropriate. - Dan Dank55 (talk) 03:17, 20 April 2008 (UTC)
GeometryGuy, when talking about the limitations of the various templates, I had more that WP:SS in mind. There are other potential uses for transclusion. For example, a lexicon explaining jargon words could be transcluded in several articles on the subject. Interestingly, Richard001 gave me another example below, about the vestigiality of wisdom teeth.
Now, imagine I want to i) transclude the vestigiality paragraph of Wisdom teeth to Human vestigiality and ii) transclude the lead paragraph of wisdom teeth to the summary article Teeth. I could do it with subpages, if I was allowed to. Can you do it with templates?
By the way, I brought this discussion to the the Village Pump proposal talk page. Emmanuelm (talk) 13:56, 20 April 2008 (UTC)
I'm not sure that such multiple transclusions are a good idea, but yes, it can be done with templates. If you only wanted to do only two transclusions, as is this case here, then on the Wisdom teeth page you could use:
 <onlyinclude>{{#ifeq:{{PAGENAME}}|Human vestigiality||text of the lead paragraph}}</onlyinclude>
...
<onlyinclude>{{#ifeq:{{PAGENAME}}|Teeth||text of the vestigiality paragraph}}</onlyinclude>
(Note the double bar and the reversal of the pagenames.) However, to be more flexible/transparent, and allow more than two separate transclusions, slightly longer code is needed:
 <onlyinclude>{{#switch:{{PAGENAME}}|Wisdom teeth|Teeth=text of the lead paragraph}}</onlyinclude>
...
<onlyinclude>{{#switch:{{PAGENAME}}|Wisdom teeth|Human vestigiality=text of the vestigiality paragraph}}</onlyinclude>
Now the pagenames are not reversed, but you also have to mention the Wisdom teeth page itself, so that the text actually appears in the Wisdom teeth article too. In either case, you then have to transclude Wisdom teeth into both Teeth and Human vestigiality, using e.g., the {{spinout}} template I created.
The "onlyinclude" tags ensure that only these sections are transcluded, while the Parser functions (#ifeq or #switch) ensure that the right text is transcluded onto the right page. Geometry guy 17:35, 20 April 2008 (UTC)
Wow, GG, you sure know a lot about this. I'll have to find a hour or two to try this. In the meantime, someone (you?) really should edit the WP:Transclusion article to explain this, ideally with working examples. And I am still waiting for someone to modify the WP:SS guideline to mention the transclusion of lead sections. Emmanuelm (talk) 20:00, 20 April 2008 (UTC)

Basic criticisms

The obvious problem, as I have pointed out elsewhere, is that some things (like use of boldface and internal links) will need to be different. For use to be appropriate the section also needs to be an exact summary of the article, which it often isn't. There's also the possibility of a lead containing material that indicates that it is a lead, e.g. 'which will be treated more fully below' or something like that (though we could create a guideline similar to WP:SELF to avoid doing anything like that such that leads can be treated as small articles in themselves, independent of the rest of the article). I thought the main article and {{summary in}} notices were enough to make editors aware they should try to keep such sections in harmony. It could also be confusing for editors who don't understand what's going on, e.g. trying to edit a section and being redirected to editing the lead of a different article. Taking these things into account, even if people do understand not to use it inappropriately (which they don't with {{main}}), it's unlikely to be much use.

The other situation where it could be used is where two sections discuss the exact same thing without there being an article on the overlap of these things (it being too trivial to warrant its own article, for example). An example that comes to mind is the vestigiality of wisdom teeth, discussed both at human vestigiality and wisdom teeth. In these cases I sometimes add a hidden comment to each section alerting editors to the existence of the other and suggesting changes are reflected in both. Richard001 (talk) 07:55, 20 April 2008 (UTC)

Richard, most of your reservations about the quality of lead paragraphs are adressed in the guideline page about them. If need be, we could tweak these guidelines to address this new use of the lead paragraph. Again, a well-written lead paragraph is a perfect summary. Emmanuelm (talk) 13:56, 20 April 2008 (UTC)
I know this; what I was saying with for use to be appropriate the section also needs to be an exact summary of the article, which it often isn't is that the summary (not the lead) is often not a true summary at all. See the example given by Geometry guy below on Al Gore for one of the no doubt thousands of such cases. Richard001 (talk) 00:27, 24 April 2008 (UTC)
I am against using this idea systematically for some of the reasons Richard001 mentions. However, I am open minded to the idea of using it occasionally, where it works well (if it does). To give a silly example (an old favourite of mine), consider List of recurring human characters from Futurama. This has a subsection on Al Gore, and links to it as a main article. However, the lead of Al Gore is completely inappropriate as a summary of the role Al Gore plays in Futurama. Although this is a silly example, it does illustrate (in a rather extreme way) that a lead section might have the wrong emphasis to be used verbatim in another article. For instance, the lead section might not be the right length.
Concerning boldface and internal links, Father Goose already raised this, and I've provided a technical solution: see Indoor bonsai, Bonsai and Houseplant, where the boldface and internal links are correct in all three articles. Geometry guy 17:50, 20 April 2008 (UTC)
Notice in WT:LEAD#Version 1.0 that lead sections might turn out to be the only content on some subjects in WP:Version 1.0. So leads are vitally important for a number of reasons, and that's what might convince some people to try this, in order to get more eyeballs in more places working on leads. You don't have that argument if you're transcluding any old section. The way to make progress with proposals in Wikipedia is generally to do one thing at a time and make your strongest case first, and be prepared to put a lot of work into this if you want it to succeed. That would probably mean regularly visiting "what links here" to that template to see if people have the right idea. - Dan Dank55 (talk)(mistakes) 19:07, 20 April 2008 (UTC)
Dank, I'll put in the time I can afford. Right now, I am waiting for code wizards like GeometryGuy to show me how to do it. Emmanuelm (talk) 20:00, 20 April 2008 (UTC)

Further issues

Okay, while I continue to be open minded towards this idea, I have noticed a couple of further issues by watching the Indoor bonsai example. First, there is a watchlisting issue: when part of one page is transcluded into another, editors watchlisting the parent page also need to watchlist the spinout to keep track of changes. Secondly, the Bonsai#Indoor bonsai section now includes information in addition to the transclusion. Since this change happened so quickly, it is reasonable to expect that such edits will not be a particularly rare occurance. They have the potential to confuse editors and mess up the coordination of the articles. These two issues are related: if someone is watchlisting all of the articles involved in a transclusion, then they can maintain the coordination, but what guarantee is there that such an editor exists? And why is this any better than leaving comments/templates in the article source and on the talk page per Richard001? Geometry guy 19:39, 23 April 2008 (UTC)

Right, I think these are significant concerns. I'm very tied up with style guidelines and other wiki-stuff; this transclusion stuff is interesting, but I'll have to come back to it in a week. - Dan Dank55 (talk)(mistakes) 21:45, 23 April 2008 (UTC)

Summary via transcluded lead paragraphs -- a more complex example

I (a pathologist) have created a complex example of how the summary style guideline can be applied using transclusion of lead paragraphs. Here is the hierarchy so far:

What I've learned: I had to use plenty of self links for the transcluded text to work in both pages. Altough I removed some duplicated content from WP, the summary article contains a lot of repetitions. Overall, I like it a lot. Emmanuelm (talk) 20:54, 23 April 2008 (UTC)

I know we need to do a better job of treating articles as part of a complex web, but using this will make articles look like they were lumped together by a machine. The use of summary style in pathology as a medical specialty looks particularly bad; two summaries there repeat the same sort of information (redundancy) and use boldface where they shouldn't, the result being an amateurish looking article. It's also probably somewhat annoying as a reader to read a summary of an article, then get the exact same text spat back at you when you click on the main article link. Richard001 (talk) 00:54, 24 April 2008 (UTC)
I've fixed the boldface issues. Also, it is best not to do the bolding of titles using self-links because if results in overlinking when the lead is transcluded. For these reasons, I already created {{spinout/title}} and {{spinout/link}} to deal with titles and links respectively. I've reduced the overlinking, but there is still some in the main pathology article as a result of all these transclusions. I also agree with Richard001 that there is some redundancy in the prose. Again, I don't want to rush to judgement: it maybe that with some rewriting the leads can be made to read well as summary sections when they are transcluded. My point of view is that we should be trying to see if this idea can be made to work, rather than declaring that it doesn't, or conversely backing it with premature enthusiasm. Geometry guy 15:58, 24 April 2008 (UTC)
Thanks GG for the formatting, it indeed works better now.
Let me remind you all the main advantages of this approach: it prevents the creation of a POV fork between the main article and the summary article and also reduces the arguments over the wording of controversial texts. If you look at my user page, you will see that I edit very "hot" pages where extreme POVs are the norm. In that environment, the "lead-as-summary" approach saves lots of time and nerves. Emmanuelm (talk) 17:39, 24 April 2008 (UTC)
I might remind you that the 'lead as summary' approach is how all leads are required to be written, transclusion or no. The fact that they aren't just reflects on the present quality of Wikipedia. Richard001 (talk) 05:03, 25 April 2008 (UTC)
Btw, there's additional information at Help:Section#Sections vs. separate pages vs. transclusion. - Dan Dank55 (talk)(mistakes) 13:24, 25 April 2008 (UTC)
This information is somewhat out-of-date and opaque: help pages are one of the least glamorous parts of Wikipedia, so they don't get the attention they deserve. Everytime I want to do something new with a parser function, I struggle to find the relevant information... Geometry guy 18:47, 25 April 2008 (UTC)
Just wanted to affirm everyone going down this path, especially coder Geometry guy, because it is a great need on WP and a community of people who are learning to do it right will help greatly to redirect those who do it wrong later. I experimented with this last fall but did not have the tools to make it stick, but it looks like we are getting there now. There is transclusion in use at 2008 straw polls, but not due to summary style and not fully smoothly either. Without having looked at the examples, some of the keys are (1) ease of using slightly different text in the main and sub sections (e.g. bolding vs. not, included vs. excluded clauses) without encouraging totally different sections; and (2) ease of editing text from both articles (e.g. getting users to know that editing the one requires thinking of editing the other, and minimizing extra clicks necessary). I'm very pleased with this progress! At some point I will reconsider its use at Ron Paul, Political positions of Ron Paul, and the rest. Keep it up. JJB 20:56, 27 April 2008 (UTC)

Can we be consistent in template use please?

We need to define the specific role of the templates used in summary style and summary-like fashions. It remains unclear to me what the exact niche of {{main}}, {{further}}, {{see}}, {{details}} (and others?) is. If they overlap, they should be merged. Each should have a specific and unique function otherwise it only creates confusion about the whole process and inconsistency among and even within articles. I have long ago proposed {{details}} and {{further}} be merged, but there has been no real response. This article is similarly vague, e.g. "—see {{Main}}, {{Details}},...)" - and what is the difference? Richard001 (talk) 23:26, 26 May 2008 (UTC)

Keep summary and main synchronized using transcluded leads

I moved this discussion from my user talk page to here.

Hi Emmanuel, I reverted, but only because I don't see any evidence that any trial projects have started or that we have the results from them yet...do we? - Dan Dank55 (talk)(mistakes) 14:24, 27 May 2008 (UTC)

The Pathology summary was a trial by me, with the invaluable technical help of GG and, in my opinion, it works well. The technique and this example were discussed here a month ago under Transclusion, an interesting tool. and its sub-headings. Then, you wrote you were in favor of this tool.
Dank, I searched for "Wikipedia:Trial project" and "Trial project" but found nothing. What is a "trial project"? What constitutes the "results" of a trial project? Emmanuelm (talk) 14:47, 27 May 2008 (UTC)
When a suggested change in guidelines involves a kind of templating that may not work as predicted and that people aren't familiar with, then WP:BRD is not the way to proceed; that is, we don't just change the guideline to say "this is okay now" and then watch what happens. I'll ask Geometry Guy if he feels the technical issues are settled, then the next step is to talk about this at some wider forum and see what the reaction is; WP:VPP would work. I'm still in favor of using the tool, but getting a number of people to try it out first and then report back is an important part of the process. There's no guideline that says you have to do this, but if you want it to succeed, it's a good idea. - Dan Dank55 (talk)(mistakes) 14:59, 27 May 2008 (UTC)
I'd like to see this tried on an article with a high edit rate such as Ron Paul (mentioned previously). Raising the topic at the Village Pump is also a good idea. The technical issues are not completely resolved, but I'm willing to help think up further improvements. Pathology works quite well, but not extremely well. My own tentative assessment of the transclusion idea is that it might better be regarded as a stage in the development of a family of articles rather than a final product. Of course there is no final product on Wikipedia, but at some point one may need to reconsider whether transclusion is helping or holding back the articles from getting better. Geometry guy 19:39, 27 May 2008 (UTC)
Okay, I've had some discussions with G-guy and mentioned this to Bulten. I was acting mostly in my role as watchdog here, and I don't have time to lead this discussion at WP:VPP or elsewhere myself; I'm hoping we can con persuade Bulten to take this on. - Dan Dank55 (talk)(mistakes) 22:06, 27 May 2008 (UTC)
P.S. I didn't want to give this any thought, but too late, I already did. If what we want is to enforce certain lead sections to be the same as certain sections in survey articles, this approach may make sense. If we're looking for a more flexible approach, one that would allow differences under some circumstances, then we're not looking for transclusion, we're looking for a tool that rapidly generates the desired diffs, so that people can easily keep track of discrepancies. - Dan Dank55 (talk)(mistakes) 22:10, 27 May 2008 (UTC)

Everything looks good, i.e., ready for beta. Emmanuel's insertion here looked fine for a first draft of the appropriate explanatory text, though of course the spinout family of templates should be mentioned here, and more documentation would be useful there. The flexibility exists in judicious use of the noinclude and includeonly templates (and perhaps onlyinclude), which needs documenting. I'll put this on watchlist and see how it can be incorporated and built upon here and there. Thanks again for everyone working on this idea. JJB 22:51, 27 May 2008 (UTC)

Thanks JJB for your kind words but my insertion was reverted and still is. How can a new idea be tried if it is deleted? I've done my part with Pathology, its your turn now. Emmanuelm (talk) 01:19, 28 May 2008 (UTC)
Okay, let me try again. Emmanuel, I very much support the idea of giving people better tools to keep lead sections in some kind of sync with sections in other articles. But we're nowhere near a consensus on whether it's better to do this with transclusion or some other tool, such as a diff-generator, and G-Guy is not going to swear that his template will work as desired, especially since he doesn't know what the desire is yet. The next step is to ask in some fairly public way what people want. And btw, this could cause real headaches with the proposed WP:Flagged revisions feature. It wouldn't be appropriate to say in this guideline "it's okay to do it", nor would it be appropriate to say here that it's proposed. I wouldn't be opposed to starting a new page where we talk about the proposal, but my guess is that's overkill; just discuss it on WP:VPP and do an RFC, and see what the reaction is. - Dan Dank55 (talk)(mistakes) 01:51, 28 May 2008 (UTC)
Another criticism that comes to mind is that summaries of articles are of various lengths; some will be shorter than the lead and others will be much longer. Transclusion makes it seem like summaries should always be lead-clones, which certainly isn't the case. It also makes changing the length more inflexible, as people are forced to keep the lead and summary the same and may not realize that abandoning the transclusion approach is a possibility. Richard001 (talk) 04:06, 28 May 2008 (UTC)
Yes, transclusion is definitely not a panacea. It is something that can be used sometimes in some articles, and perhaps later removed. This needs to be made clear. The only way to explore its scope and limitations is to develop examples. I'm willing to contribute template-based solutions to issues which arise. Geometry guy 20:35, 28 May 2008 (UTC)

Added shortcut

I added a shortcut called WP:AVOIDSPLIT because I think certain merge / split discussions need to keep it in mind. If someone needs to reformat it or rename it, I would support that. Randomran (talk) 06:34, 29 May 2008 (UTC)

Judith Butler reception: Summary style or POV fork?

See Talk:Judith Butler and Talk:Influence of Judith Butler's concepts. Thanks! Hyacinth (talk) 04:20, 16 June 2008 (UTC)

Transwikification

I removed a reference to transwikification but was later reverted. As the standard next step, I'd like to explain my rationale in more detail.

The clause in question is this: "If information can be trimmed, merged, moved to another wiki, or removed, these steps should be undertaken first before the new article is created."

User:Zappernapper is right in saying that transwikifiction is a fairly common practice. The problem with the current wording is that it describes transwikification as a competitive alternative to on-wiki forking; it essentially says that we cannot .

Clearly this is problematic. The chance that a reader will see transwikied content on any particular subject is relatively low. External links can help, but since WP:SS deals with subtopics within broader articles, bottom-of-the-page links are generally inappropriate (as too narrow), and using them in {{see}}-like notices goes against various guidelines and norms. Transwikification is a Good Thing, but it shouldn't be regarded as a replacement for housing content in-wiki, especially in the context of WP:SS.

xDanielx T/C\R 19:26, 2 July 2008 (UTC)

Could not update stat
UP