Template talk:Sfn/Archive 2

Archive 1

Archive 2

Archive 3

Archive 4

Archive 5

Sfnp

As we add features here, we need to remember to update the variants: {{sfnp}} and {{sfnm}}. I added |paren= to sandbox to add parenthesis for the year:

Markup	Renders as
{{sfn/sandbox|Smith|2005|p=25}} {{sfn/sandbox|Smith|2006|p=25|paren=y}} {{reflist}} {{refbegin}} * {{cite book |last=Smith |title=Book |year=2005 |ref=harv}} * {{cite book |last=Smith |title=Book |year=2006 |ref=harv}} {{refend}}	Smith (2005). Book. {{cite book}}: Invalid |ref=harv (help) Smith (2006). Book. {{cite book}}: Invalid |ref=harv (help)

What we really need to do is create {{sfn/core}} with named parameters so we can keep all of the variants in sync. ---— Gadget850 (Ed) ^talk 07:51, 20 March 2012 (UTC)

Obsoleted by #Core. ---— Gadget850 (Ed) ^talk 09:10, 3 April 2012 (UTC)

question on Harvb error in All Saints' Church, Shuart

I can't figure out why ref 38 in All Saints' Church, Shuart throws an error: "Harv error: link to #CITEREFLewis1732 doesn't point to any citation." Thanks, MathewTownsend (talk) 18:06, 4 April 2012 (UTC)

It was the wrong year, but the problem has already been fixed. Ucucha (talk) 20:18, 4 April 2012 (UTC)

Thanks! (I thought I checked for that - guess not.) MathewTownsend (talk) 20:23, 4 April 2012 (UTC)

Centralized talk page

Propose centralizing these talk pages here, as this is the most active talk page.

• Template talk:Harvard citation
• Template talk:Harvard citation no brackets
• Template talk:Harvard citation text
• Template talk:Harvcol
• Template talk:Harvcolnb
• Template talk:Harvcoltxt
• Template talk:Sfn
• Template talk:Sfnm
• Template talk:Sfnp
• Template talk:Harvard citation documentation

See WP:TALKCENT for details on the process. ---— Gadget850 (Ed) ^talk 09:39, 3 April 2012 (UTC)

• Support Blue Rasberry (talk) 13:25, 3 April 2012 (UTC)

• Oppose. Not a good idea. If talk activity is to be the criterion of where to move a discussion then why not merge with Wikipedia:Editor assistance/Requests? Better that discussions be organized by topic. While there might be some justification for pulling Sfnm and Sfnp here, and while this topic (Sfn) links with the various aspects of Harv, it is different enough to not warrant sucking in all the other less related topics into one confusing tar ball. ~ J. Johnson (JJ) (talk) 19:22, 4 April 2012 (UTC)

The only difference between {{sfn}} and {{harvnb}} is that the former is wrapped in a <ref>...</ref> tag. Indeed, I proposed a #Core for the nine templates. ---— Gadget850 (Ed) ^talk 23:30, 4 April 2012 (UTC)

Indeed. But this is like saying nine kinds of bolts may – but note: need not – be put into a plastic bag. That doesn't make "plastic bag" the preferred reference term for "bolts" generally. Sfn mixes two different kinds of tools (should discussion of <ref> also be merged here?); that does not subordinate those tools, or their discussion, to this tool. And while Sfn uses Harv, it is not itself any form of Harv. Blurring these concepts would lead to more of the confusion of concept has made citation on Wikipedia unnecessarily difficult. We don't need to compound that. ~ J. Johnson (JJ) (talk) 23:16, 5 April 2012 (UTC)

{{sfn}} doesn't mix <ref>...</ref> with {{harv}} - it encloses {{harvnb}} in <ref>...</ref>. The construct <ref name=id>{{harvnb|author|year|p=n}}</ref> is a common one, and {{sfn|author|year|p=n}} does the same job in at least 21 fewer characters. This makes the wikicode significantly shorter, with the added benefit of ensuring the uniqueness of id. --Redrose64 (talk) 09:55, 6 April 2012 (UTC)

These templates are very similar in both use and in markup. Enough that changes to one template should be discussed and applied to all, and enough that queries about use of one applies to all. ---— Gadget850 (Ed) ^talk 11:26, 6 April 2012 (UTC)

But sfn is an outlier in naming, at least. It would make more sense to me to centralize on one with a more central name, like {{harv}}. —David Eppstein (talk) 16:13, 6 April 2012 (UTC)

 Request withdrawn ---— Gadget850 (Ed) ^talk 16:25, 6 April 2012 (UTC)

Strong Support (although I'm sorry it's belated). These templates are nearly identical. If there is an issue with one of them, chances are the issue applies to all of them. If a change is made to one of them, the change should be made to all of them. Gadget: why the rush to withdraw? ---- CharlesGillingham (talk) 07:27, 7 April 2012 (UTC)

Pretty obvious there is no consensus to do this nor where, so there is no sense in trying to bag these bolts. Will re-propose in six months. ---— Gadget850 (Ed) ^talk 12:12, 7 April 2012 (UTC)

Also obvious that the inner working of these templates is not understood. So, even though they use a common doc page, we will just have to do some education before going to a central talk page. ---— Gadget850 (Ed) ^talk 17:20, 8 April 2012 (UTC)

I suspect there would be no objection to merging the Sfn* talk pages. But dissimilarity between Sfn* and Harv* is not something that is likely to change with simply the passage of time. ~ J. Johnson (JJ) (talk) 19:52, 8 April 2012 (UTC)

Harvard citations

Was {{Harvard citations}} a deliberate omission or an oversight? It would seem to be a more natural choice of where to centralize than this one to me, as being in some sense the most general of the lot rather than the most specific. —David Eppstein (talk) 15:22, 3 April 2012 (UTC)

I am pondering that one. It is very different from the listed templates and does not have the potential to update to a common core. ---— Gadget850 (Ed) ^talk 15:34, 3 April 2012 (UTC)

{{Harvard citations}} should not be included. It is structured very differently than the other templates, and there is unlikely to be an overlap of issues. ---- CharlesGillingham (talk) 07:27, 7 April 2012 (UTC)

Core update

Given the commonality in markup for the following templates, I have developed a meta-template at {{Harvard citation/core}}. A version using this core is in each sandbox.

Note that {{Harvard citations}} is very different in implementation, thus is not supported by the meta-template.

In-text templates

Templates	Aliases	Sandbox
{{Harvard citation no brackets}}	{{harvnb}}	{{Harvard citation no brackets/sandbox}}
{{Harvard citation}}	{{harv}}	{{Harvard citation/sandbox}}
{{Harvard citation text}}	{{harvtxt}}	{{Harvard citation text/sandbox}}
{{Harvcoltxt}}		{{Harvcoltxt/sandbox}}
{{Harvcol}}		{{Harvcol/sandbox}}
{{Harvcolnb}}		{{Harvcolnb/sandbox}}
{{sfn}}		{{sfn/sandbox}}
{{Sfnp}}		{{Sfnp/sandbox}}
{{Sfnm}}		{{Sfnm/sandbox}}

Here are samples of the current implementations:

Current

Markup	Renders as
{{Harvard citation no brackets |Smith |2015 |p=25}}	Smith 2015, p. 25
{{Harvard citation |Smith |2015 |p=25}}	(Smith 2015, p. 25)
{{Harvard citation text |Smith |2015 |p=25}}	Smith (2015, p. 25)
{{Harvcolnb |Smith |2015 |p=25}}	Smith 2015:25
{{Harvcol |Smith |2015 |p=25}}	(Smith 2015:25)
{{Harvcoltxt |Smith |2015 |p=25}}	Smith (2015:25)
{{sfn |Smith |2015 |p=25}}
{{sfnp |Smith |2020 |p=25}}
{{sfnm |1a1=Smith |1a2=Jones |1a3=Johnson |1y=2005 |1p=15 |2a1=Jones |2a2=Johnson |2a3=Smith |2y=2004 |2p=50}}
{{Harvard citation no brackets |Smith |p=25 |ref=smithref}}	Smith, p. 25

And samples using the new core:

Harvard citation/core

Markup	Renders as
{{Harvard citation no brackets/sandbox |Smith |2015 |p=25}}
{{Harvard citation/sandbox |Smith |2015 |p=25}}
{{Harvard citation text/sandbox |Smith |2015 |p=25}}	Smith (2015, p. 25)
{{Harvcolnb/sandbox |Smith |2015 |p=25}}
{{Harvcol/sandbox |Smith |2015 |p=25}}
{{Harvcoltxt/sandbox |Smith |2015 |p=25}}
{{sfn/sandbox |Smith |2015 |p=25}}
{{sfnp/sandbox |Smith |2020 |p=25}}
{{sfnm/sandbox |1a1=Smith |1a2=Jones |1a3=Johnson |1y=2005 |1p=15 |2a1=Jones |2a2=Johnson |2a3=Smith |2y=2004 |2p=50}}
{{Harvard citation no brackets/sandbox |Smith |p=25 |ref=smithref}}

• Smith (2015). Book. {{cite book}}: Invalid |ref=harv (help)
• Smith (2020). Book. {{cite book}}: Invalid |ref=harv (help)
• Smith. Book.
• Jones, John; Johnson, John; Smith, John (2004). Our First Book. {{cite book}}: Invalid |ref=harv (help)
• Smith, John; Jones, John; Johnson, John (2005). Our Second Book. {{cite book}}: Invalid |ref=harv (help)

I propose updating to this core, as it would keep the related templates synchronized, thus simplifying upkeep. ---— Gadget850 (Ed) ^talk 01:32, 9 April 2012 (UTC)

• Strongly support this update being made. It will make no difference to users of these templates, but will make maintenance much easier. As an example, last month changes to allow both a page number/range and a location to be displayed were made to one of the "sfn*" templates, but not to the others. This kind of inconsistency can be removed entirely by having a central core template which is used by all the others but which is never directly used in articles. The approach works very well elsewhere (e.g. in the templates which create the taxoboxes visible in almost all articles about animals, plants and other organisms). Peter coxhead (talk) 08:05, 9 April 2012 (UTC)

Looks reasonable enough. ~ J. Johnson (JJ) (talk) 22:22, 9 April 2012 (UTC)

• Support. I note that the new version also fixes an issue in sfnp where leading and trailing spaces in the year are significant. —David Eppstein (talk) 23:39, 9 April 2012 (UTC)

Done except for {{sfnm}}. Need to dig into that some more to figure out what is wrong. ---— Gadget850 (Ed) ^talk 16:09, 28 April 2012 (UTC)

Can this be used with comic books?

I'm trying to upgrade some comic related articles and I've used this template a lot with film articles, but Template:Cite comic does not use a separate surname parameter and multiple issues of a comic can be released in a single year so I'm not clear how SFN differentiates between which I'm referencing. For example if I cite two comics in the same series by a single writer in a single year. Darkwarriorblake (talk) 14:35, 17 July 2012 (UTC)

{{Cite comic}} does not support an anchor, see Template:Sfn#Citation format does not support anchors: .7B.7Bwikicite.7D.7D. See Template:Sfn#More than one work in a year. ---— Gadget850 (Ed) ^talk 15:14, 17 July 2012 (UTC)

I added |ref= to {{cite comic/sandbox}}. Review and discuss at Template talk:Cite comic#Anchor. ---— Gadget850 (Ed) ^talk 15:16, 17 July 2012 (UTC)

Thanks for the input and change Gadget850, should prove useful. Darkwarriorblake (talk) 16:51, 17 July 2012 (UTC)

 Done ---— Gadget850 (Ed) ^talk 12:55, 31 July 2012 (UTC)

Template documentation

The template documentation should include a simple list or table for the available parameters.
For instance, I am attempting to cite a not-so-unusual source: a book with different authors for each chapter. Since the book itself only has editors, and not an author, per se, I was wondering if 'editors' (plural) was a viable parameter, and if 'chapter' had a parameter, and if there was some means of including a 'chapter title'. Ideally, there would be a 'chapter author' for multiple citations. Unfortunately, the examples given did not provide anything appropriate to this situation. My point is: a parameter list would provide guidance for essentially any situation. ~Eric F 74.60.29.141 (talk) 15:07, 27 October 2012 (UTC)

I found it here:Template:Cite_book -- perhaps a link "For usage information..." at the top of the page? ~E 74.60.29.141 (talk) 15:19, 27 October 2012 (UTC)

Can I have a little help?

On the article Throffer, the template is struggling to link one of my sources (Steiner 1974-75) to its entry in the bibliography. I assume this is because the citation has two years. Is there a way around this? J Milburn (talk) 17:28, 23 February 2013 (UTC)

No, it was that you were using |author= and a wikilink instead of |first=, |last= and |authorlink=. Kanguole 17:53, 23 February 2013 (UTC)

Thanks, it's appreciated. J Milburn (talk) 21:02, 23 February 2013 (UTC)

Lua version is broken

Tracked in Phabricator
Task T48815

Using {{sfn}}: One. Two. Three. Four.

Using {{sfn/sandbox}}: One. Two. Three. Four.

As shown above, the Lua version omits the small-letter backlinks which were present before. This is a bug. --Redrose64 (talk) 22:41, 2 April 2013 (UTC)

Filed as bugzilla:46815. Dragons flight (talk) 23:24, 2 April 2013 (UTC)

I'm now seeing some strange behavior on Adriaen van der Donck. Namely, number 15 (O'Donnell 1968 xxxviii.) is using the small letter backlinks, whereas other multiple citations aren't, but it's rendering it as O'Donnell & 1968 xxxviii, which is breaking the link to the full reference. Laura Scudder | talk 21:13, 3 April 2013 (UTC)

Two problems there. One is the {{sfn}} bug that I described above; the other is that the article contained {{sfn|O'Donnell|1968|xxxviii}} which is misuse of the third positional parameter - it should have been a named parameter as in {{sfn|O'Donnell|1968|p=xxxviii}}.

Since bugzilla:46815 doesn't yet have a solution, I've reverted {{sfn}} to the last working version (this means that Module:Footnotes is not used). I've also put the version of {{sfn}} that had been live into {{sfn/sandbox}} so that the above demo continues to exhibit a difference. --Redrose64 (talk) 16:13, 4 April 2013 (UTC)

The next update for Lua should include the frame:extensionTag function which is a functional work-around for this issue. I was told about 2 weeks ago that the relevant incremental release for Mediawiki would probably be this week, so we may have a solution soon, but I don't know for sure. Dragons flight (talk) 16:37, 4 April 2013 (UTC)

I've restored the Lua version. The underlying ref bug is not actually fixed, but the recent update to Mediawiki provided a work-around so that I could make the Lua module work anyway. Dragons flight (talk) 23:37, 8 April 2013 (UTC)

Multiple Volume in same year

How do you differentiated volume 1 of a work form volume 2 of another work with the same title, same author, and same publication year.--KAVEBEAR (talk) 09:14, 5 June 2013 (UTC)

When the same author/authors have more than one publication in a year the usual practice (not just on Wikipedia) is to add a lowercase letter to the year - eg 1997a, 1997b etc. (both in the cite/citation template and in the sfn template) Aa77zz (talk) 09:48, 5 June 2013 (UTC)

Template:Sfn#More than one work in a year --  Gadget850 ^talk 10:28, 5 June 2013 (UTC)

SFN link

Resolved

Can someone help me get the Coplans link to the references section working at Whaam!.--TonyTheTiger (T/C/BIO/WP:CHICAGO/WP:FOUR) 15:56, 21 July 2013 (UTC)

 Done, see here --Redrose64 (talk) 16:23, 21 July 2013 (UTC)

 Done as well! I've had a go at the references section and tidied it a bit. The harvard references for the four books are now:

• id="CITEREFAlloway1983"
• id="CITEREFBader2009"
• id="CITEREFCoplans1972"
• id="CITEREFWaldman1993"

You just need to use {{sfn|Coplans|1983|p=?}}. Even the Waldman (1993) link ought to work satisfactorily. HTH, Martin of Sheffield (talk) 16:49, 21 July 2013 (UTC)

Thanks.--TonyTheTiger (T/C/BIO/WP:CHICAGO/WP:FOUR) 16:53, 21 July 2013 (UTC)

Two authors and no year

I occasionally come across a web page where there is no easily definable year, see for instance [1]. Since there is no year, the reference ({{sfn|Lane|Singh}}) correctly produces #CITEREFLaneSingh but displays as Lane & Singh not Lane & Singh. Is there any way to force the correct display whilst keeping the automatic linking from text through citation to bibliography? Martin of Sheffield (talk) 22:43, 7 July 2013 (UTC)

The last value is being parsed as the year, not the author, thus the lack of the ampersand. I don't see any current way to fix this. The best way would be to add |year=none to the module to explicitly suppress the year. --  Gadget850 ^talk 18:22, 22 July 2013 (UTC)

Thanks for the suggestion. I've used year=unknown for Lane & Singh on the Richard Watts page. Martin of Sheffield (talk) 22:18, 22 July 2013 (UTC)

Having no year of publication seems so gravely deficient I wonder if it ought to be considered an error condition. But on the assumption that there are some cases where we have to finesse not having a year: couldn't Harvid be used for this? ~ J. Johnson (JJ) (talk) 22:32, 22 July 2013 (UTC)

The problem is not in the references, template:citation generates them quite correctly. The problem lies with how the inline reference is displayed. The inline reference correctly links using CITEREF, but incorecctly displays the last author as if the author were a year. I can't see that template:harvid would help this, but if you know better please explain. The only alternative I know to using year=unknown is to use <ref>{{harv|Lane & Singh}}</ref> which of course does not link through. As regards the lack of a year, please have a look at [2]. The provenance is excellent, the sponsors being the charity themselves, and some conclusions are noteworthy but I have not been able to find a year. Should I include apparently good information or reject it due to a fault in the web site? Martin of Sheffield (talk) 23:06, 22 July 2013 (UTC)

Using the wget utility on the the above link to retrieve the raw HTML, and then using the MS-DOS dir command on the retrieved file shows me

16/04/2013  11:45             9,164 richardwatts1.html

- that's BST, therefore the true datestamp is 10:45, 16 April 2013 (UTC) --Redrose64 (talk) 14:30, 23 July 2013 (UTC)

Excellent lateral thinking. I know for a fact that the page is significantly older than that (accessdate = 21 June 2012) but is the best alternative there is. I checked to make sure that MS wasn't "doing its own thing" and confirmed it using ls -l

-rw-r--r-- 1 jmr users 9164 Apr 16 11:45 richardwatts1.html

I even grep'ed for the string "20", but could only see the visible "middle of the 20th Century".

More generally though, there needs to be a way of dealing with multiple authors without a date. wget and checking the date is probably a bit geeky for many editors. Martin of Sheffield (talk) 17:47, 23 July 2013 (UTC)

Yes! I thought I had done something like this before. Try this:

*[[#{{Harvid|Smith|Jones}}|Smith & Jones]]

Which with something like this:

*{{citation
 |ref={{harvid|Smith|Jones}}
 |year= ''undated'' 
 |last1= Smith |first1= A.
 |last2= Jones |first2= B.
 |title= Some source with no date
}}.

gives us "Smith & Jones" linking to:

• Smith, A.; Jones, B. (undated), Some source with no date {{citation}}: Check date values in: |year= (help)CS1 maint: ref duplicates default (link) CS1 maint: year (link).

Which implies a harv link can be made to display any way we want it. ~ J. Johnson (JJ) (talk) 20:08, 24 July 2013 (UTC)

I altered the |ref= to pull in {{harvid|Smith|Jones}} - an exact match of the one used inside the <ref>...</ref>. If they match, somebody who arrives later on will have a better idea of the method (CITEREF... is somewhat obscure) --Redrose64 (talk) 21:42, 24 July 2013 (UTC)

Thanks to both of you. I've altered Richard Watts to use

<ref>[[#{{harvid|Lane|Singh}}|Lane & Singh]]</ref>

in the text. As J. Johnson suggested, this generates the correctly formatted entry in the {{references}} section and the appropriate CITEREFLaneSingh. I simply removed the "year=unknown" parameter from the citation in the bibliography, there is no need for an explicit ref parameter, the default is correct. Harvid is indeed flexible, see template:NHLE for an example of hijacking the year. I'll have a go at including this work around in the documentation for sfn. Once again, thanks. Martin of Sheffield (talk) 22:00, 24 July 2013 (UTC)

I have generally preferred using the explicit CITEREF as I know exactly what I'm getting. But perhaps using parallel harv templates is more straightforward? I wouldn't mind if we could determine a preferred practice here. ~ J. Johnson (JJ) (talk) 22:01, 24 July 2013 (UTC)

Self-referencing link?

Don't see what the use of having a citation that self-references the article you are in, and does not actually point a new reader to the actual cite. Why not have a named citation that the reader can see and doesn't have to "look for" in a list of citations.

There's an irony here inasmuch as we stop newbies from using self-referencing links. Student7 (talk) 14:50, 30 August 2013 (UTC)

Well, it doesn't reference the article you are in – it just requires a second step to find the full reference.

But consider the alternative when a book is used many times in an article, and page references are given, as for example the use of Anderson (2001) in Cactus#References. The alternative to the use of the Harvard style links in the list of references is to repeat the full book citation every time – there are more than 25 such uses. What's the point of 25+ repetitions of "Anderson, Edward F. (2001), The Cactus Family, Pentland, Oregon: Timber Press, ISBN 978-0-88192-498-5"? Peter coxhead (talk) 15:38, 30 August 2013 (UTC)

What does the #invoke magic word do?

Because the Ringo Starr page on the Beatles Wiki was in a terrible mess (caused by somebody copying a previous version of the Wikipedia equivalent in a totally clueless fashion, resulting amongst other things in two different tables of contents, the Wikipedia one and the local one), I have just copied the current version over; unfortunately this involves copying the templates used, if they don't exist there.

I have hit a major roadblock copyng this (sfn) template; the #invoke magic word isn't recognised by Wikia's version of the cite parser, and neither the magic-words documentation nor the cite-parser documentation explains this either. I need to know so that I can recreate this functionality by some other means. Please reply on my userpage there, User:RobertATfm. — 188.29.16.101 (talk) 14:21, 13 October 2013 (UTC)

It's listed at Help:Magic words#Other and that directs you to Wikipedia:Lua which in turn links to mw:Extension:Scribunto; that extension needs to be installed on the wiki concerned. If you don't want to do that, you can use a pre-Lua version of {{sfn}} - the last one is here. --Redrose64 (talk) 18:25, 13 October 2013 (UTC)

Several page numbers to same book

I have used the sfn template on Wells Cathedral which we are working towards an FA nomination. One book (Smith 1975 - currently ref No: 101) has been used to support several claims. Another editor who has a copy of the book says that all the text is included on pages 1-2 and 25. What would be the page number syntax to show this within the sfn template as I can't see it in any of the examples given?— Rod ^talk 13:19, 20 November 2013 (UTC)

You can say {{sfn|Smith|1975|pp=1–2, 25}} in each place, but you might also consider being more specific, say if one of the statements were based on page 25 only. Kanguole 13:57, 20 November 2013 (UTC)

Thanks for your help - now done.— Rod ^talk 14:05, 20 November 2013 (UTC)

Proposed addition of a 'url' parameter

For me the sfn template is an elegant approach to citing an article. Unfortunately, the use of a separate reference appears to limit the use of links to individual online book web pages. Since the sfn template carries the page information for the reference, I decided I would like to be able to add a 'url' to a specific online book page. Currently it appears possible to do this by passing the link directly in the page field. However, I would like to propose making this a separate parameter, such as 'url='. Here's an example:

Referencing this statement.

Notes:

References:

• Ryan, Craig (2003), The Pre-astronauts: Manned Ballooning on the Threshold of Space, Naval Institute Press, ISBN 1591147484.

• example sfn entry: {{sfn|Ryan|2003|pp=[http://books.google.com/books?id=1QS38bu9iTwC&pg=PA51 51–54]}}
• proposed sfn entry: {{sfn|Ryan|2003|pp=51–54|url=http://books.google.com/books?id=1QS38bu9iTwC&pg=PA51}}

Does this seem reasonable? Praemonitus (talk) 18:53, 20 April 2013 (UTC)

Agreed. --SlothMcCarty (talk) 23:28, 10 January 2014 (UTC)

Multi-year ranges produce in {{sfn}} produce different results that those produced by {{cite Journal}}

When the publication date for a journal consists of a multi-year range, the span id produced by {{sfn}} differs from those generated by {{cite Journal}}. An example is:

{{Cite journal |last=Farlow |first=Archa Malcolm | title = Arizona's Admission to Statehood | journal = Annual Publication of the Historical Society of Southern California | volume = 9 | issue = 1/2 | pages = 132–153 | publisher = University of California Press | date = 1912–1913 | jstor = 41168902 |ref=harv}}

which produces a span id of "CITEREFFarlow1912.E2.80.931913", while

{{sfn|Farlow|1912-1913|p=152}}

produces a corresponding span id of "CITEREFFarlow1912-1913". Things would work better if the two matched. --Allen3 ^talk 01:35, 16 April 2014 (UTC)

Using an en-dash character instead of "&ndash;" fixed it. It seems that {{sfn}} handles &ndash; properly, but Module:Citation/CS1 doesn't – the {{cite journal}} produces "CITEREFFarlow", while {{sfn|Farlow|1912&ndash;1913|p=152}} produces "CITEREFFarlow1912.E2.80.931913". This is mentioned in the table at Module talk:Citation/CS1/Archive 10#Update to the live CS1 module week of 2014-03-23. Kanguole 07:39, 16 April 2014 (UTC)

In the example given by Allen3, the |date=1912–1913 in {{cite journal}} does indeed use an en-dash - the character, not the entity – whereas the {{sfn}} doesn't use a dash of any form - it uses a hyphen-minus (the character typed on a normal keyboard). If you alter {{sfn|Farlow|1912-1913|p=152}} to {{sfn|Farlow|1912–1913|p=152}} it will work, because the {{sfn}} will now generate the URL fragment #CITEREFFarlow1912.E2.80.931913 to match the anchor in the {{cite journal}}. --Redrose64 (talk) 08:22, 16 April 2014 (UTC)

True, but in the article Allen3 was having trouble with, both used "&ndash;". Kanguole 08:30, 16 April 2014 (UTC)

Citation has multiple authors and no date

In section citation has multiple authors and no date it's suggested to use <ref>[[#{{harvid|Lane|Singh}}|Lane & Singh]]</ref> instead of {{sfn|Lane|Singh}}. Is it also OK to add a |date=n.d. to citation and use {{sfn|Lane|Singh|n.d.}} instead?

Example.

References

• Lane, Kieran; Singh, Karun (n.d.), "Richard Watts", Richard Watts Charities, retrieved 21 June 2012 — Preceding unsigned comment added by دالبا (talk • contribs) 19:28, 21 June 2014‎

I have never seen the point of "n.d." The {{sfn|Lane|Singh}} form works perfectly well if the |year= and |date= parameters are omitted from the {{citation}}. --Redrose64 (talk) 19:32, 21 June 2014 (UTC)

<ref>[[#{{harvid|Lane|Singh}}|Lane & Singh]]</ref> produces a footnote "Lane & Singh" with a CITEREFLaneSingh which is what is normally required. {{sfn|Lane|Singh}} produces a footnote of "Lane Singh" with a CITEREFLaneSingh. The linkage works fine but the displayed footnote is inadequate. For example:

References

Personally I'd be worried about using "n.d." Remember WP:RF, what would a reader who is interested in the citations but doesn't understand Wiki markup think? Martin of Sheffield (talk) 22:24, 1 July 2014 (UTC)

Although it may seem unusual for some readers but it's recommended by Harvard style guide for situations where there is no known date and no approximate date can be estimated. (Maybe it's OK to leave it out, but I hope this discussion won't go so far as prohibiting its usage.) Dalba 05:32, 2 July 2014 (UTC)

Perhaps; but {{sfn}} is not Harvard and was never intended to be. Wikipedia does not enforce any one style, save that in order to qualify for featured article status, the citation style must be consistent within the article. --Redrose64 (talk) 11:35, 2 July 2014 (UTC)

By a curious coincidence I was recently wondering about dropping the year from a standard author-date (but not necessarily Harvard reference!) type of citation. Not for lack of a year, just for not being necessary. It seems to me that inclusion of the year is so standard that anyone with any familiarity with citations would find such a citation odd. Getting back to the case considered here, where no date is available, that does happen. And I wonder if the lack of a year is odd enough that that something (e.g., "n.d.") ought to be supplied in it place. In the print environment this is recommended as a placeholder where a data might be forthcoming. In our environment it could indicate that the absence of the year is not due to inadvertence, or some editor's goofy "style". ~ J. Johnson (JJ) (talk) 18:54, 2 July 2014 (UTC)

The problem as I see is that a reader who is unfamiliar with academic works but is intelligent, as WP:RF says a high school pupil, won't know what "n.d." means. If they see "Lane & Singh p. 3" it is pretty clear. "Lane & Singh (2009) p. 3" is also pretty self explanatory if you see an entry such as "Lane, Keiran & Singh, Karun (2009) ..." in the bibliography. Likewise constructs such as "2009?" or possibly even "c. 2009" work. However,, without consulting the Harvard manual of style, what does "n.d." mean? "Nunc dimittus?", "nearly done" or possibly Singh has the initials N.D. but doesn't like to use them, or alternatively they might be Singh's nickname. If you must use a placeholder then avoid technical jargon: "Lane & Singh (date unknown)" is a clear as a bell. Martin of Sheffield (talk) 22:36, 2 July 2014 (UTC)

The advantage of "n.d." in the print environment is that it could be replaced by a four digit date without upsetting any of the formatting. We, of course, are not concerned about that, and I think "date unknown" is quite satisfactory. The question is whether, in cases where the date (year) is unknown, it is satisfactory to omit the date entirely, without any explanation, or whether some indication be given ("n.d.", "date unknown", "????" or whatever) lest a reader think the omission is due to inadvertence. ~ J. Johnson (JJ) (talk) 19:13, 3 July 2014 (UTC)

"Script Error" in Statue of Liberty

Statue of Liberty is very messed up right now, as it has "Script Error" in big red bold all over it, apparently where sfn is used. Is someone looking at this?

Oddly it seems OK in Opera and Firefox, but fails in Chrome "Version 36.0.1985.125 m" on Windows 8. The first instance is on the last words of "Origin":

which Bartholdi would later bring to the Statue of Liberty.Script error

A quick survey of a few other pages which use sfn shows no problems for me. Opening the article with Chrome in an incognito window gives me the same error, though. johantheghost (talk) 16:51, 10 August 2014 (UTC)

Well, it was happening consistently through restarts etc., but now it's OK... maybe a server glitch? johantheghost (talk) 03:33, 11 August 2014 (UTC)

Source in source

Is there a good way to cite an article in an anthology (book or otherwise), with just a short listing for the book which links to the full entry? This would be like sfn, but would include the author and title of the article. There is now a Template:Source in source to do this. But it is now limited in its capability. Is there a better way to do this, perhaps with sfn? DougHill (talk) 23:53, 22 August 2014 (UTC)

Could someone check my syntax?

I'd tried to implement {{sfn}} at Cakewalk to stop the References section looking too cluttered with replicated citations of the same thing, just with different page numbers, but it threw up errors each time. Could someone please check my code? Thanks! It Is Me Here ^{t / c} 19:59, 1 September 2014 (UTC)

@It Is Me Here: The {{sfn}} template cannot be used inside <ref>...</ref>. Either remove the <ref> and </ref> wherever {{sfn}} is used, or use {{harvnb}} instead of {{sfn}} (but still inside <ref>...</ref>). --Redrose64 (talk) 21:01, 1 September 2014 (UTC)

That is correct. For the diff you have provided, with the callout consisting of just the template, I would suggest removing the ref tags and leaving just sfn as that will result in simpler source. --Mirokado (talk) 22:07, 1 September 2014 (UTC)

To editor Redrose64: OK, thanks, I've gone for the {{harvnb}} option. Now, the internal links Fletcher 1984 and Stearns & Stearns 1994 at Cakewalk#Notes work, but Baldwin 1981 doesn't. Any ideas? It Is Me Here ^{t / c} 22:55, 1 September 2014 (UTC)

{{cite jstor}} doesn't recognise |ref=harv, nor does it pass it along the chain. Even if it did, the other links in the chain don't recognise it or pass it along either; so rather than amend each one, it's easier to go right to the end and do this. --Redrose64 (talk) 23:06, 1 September 2014 (UTC)

Is that a bug, or intentional? Would it make sense to start a thread about this at Template talk:Cite jstor? It Is Me Here ^{t / c} 12:20, 2 September 2014 (UTC)

I don't know; jstor is outside my field (the identifiers that I typically use are isbn, lccn and oclc), so I never use {{cite jstor}}. --Redrose64 (talk) 14:15, 2 September 2014 (UTC)

Along with some other templates, {{cite jstor}} shouldn't be used for any new sources at present, as it relied on software on the old tool server, now defunct. See its documentation. Peter coxhead (talk) 14:28, 2 September 2014 (UTC)

Broken wrt {cite doi}

Recently, there and edit was made (not here, but I don't know where it was), that breaks {sfn} behaviour wrt {{cite doi}} (possibly because {cite doi} was deprecated recvently). That breaking edit is unacceptable. It is up to the closing process (actors) to cleanup any undesired (newly introduced) usage. That could be a bot in this case. -DePiep (talk) 09:11, 13 October 2014 (UTC)

@DePiep: The {{sfn}} template hasn't been edited in six months, and {{cite doi}} not in eight months. What is this breakage, and on what pages does it manifest itself? --Redrose64 (talk) 12:17, 13 October 2014 (UTC)

Indeed, I don't know where the edit was made. This question is evolving at Module_talk:Citation/CS1#Where_dos_this_break_come_from.3F (I'm sorry, bad I did multi-forum this). -DePiep (talk) 13:02, 13 October 2014 (UTC)

Postscript

"The postscript is only effective the first time sfn is used for a particular author, year and location." Is this deliberate or is it a bug? I can't see any obvious reason why this would be a good thing, but equally it's presumably quite easy to change with the addition of an extra conditional? Unless I'm missing something there could well be occasions when a two refs would want to refer to the same page, but not quote the same text or make the same point. JimmyGuano (talk) 07:55, 1 January 2015 (UTC)

In that, probably rare, case consider using |loc= to disambiguate the quotations: {{sfn|Jones|1984|loc= p. 34, para 1|ps= (but he contradicts himself below)}} and {{sfn|Jones|1984|p=34}}.

Martin of Sheffield (talk) 13:51, 1 January 2015 (UTC)

Thanks - that does look like a viable workaround. I'm still not clear why this situation is the case though? JimmyGuano (talk) 07:26, 3 January 2015 (UTC)

I think because virtually always editors are content to simply cite pages (indeed getting them to do that can be an issue). {{sfn}} is explicitly aimed to collect citations, and the tuple (author, date, page) is nearly always adequate. Off hand, except for references to multicolumn works (newspapers, physically large encyclopaedias), I can't recall seeing detail finer than that. As an alternative to using |ps= to add a note, you can always use the {{efn}} mechanism. See Hartley Colliery Disaster for an example of footnotes. HTH, Martin of Sheffield (talk) 13:44, 3 January 2015 (UTC)

@JimmyGuano:When two instances of {{sfn}} have identical authors, year, |p= |pp= |loc= they are treated as entirely identical, regardless of any other parameters that may be present. This is because it is only those eight parameters which are used to construct the name= attribute of a <ref name=...> tag, and when the MediaWiki software finds two such tags with identical names, it assumes that they are identical in all other respects, and only displays the information pertaining to the first instance. You can get a similar effect with regular <ref>...</ref> tags:

First claim;<ref name=Smith2015p1>Smith (2015), p. 1 "The Government announced that some taxes were to be increased"</ref> Second claim.<ref name=Smith2015p1>Smith (2015), p. 1 "The Government announced that some restrictions would be imposed"</ref>

gives

First claim; Second claim.

Notice how the two refs are merged, with the text of the second being ignored. --Redrose64 (talk) 16:17, 3 January 2015 (UTC)

Thanks Redrose, very good point about the similar <ref></ref> behaviour. Just to make it absolutely clear, in your example the whole of the text is being ignored, even the repeated citation:

One<ref name=Smith2015p1>Smith (2015), p. 1 "Comment 1"</ref> Two<ref name=Smith2015p1>Text with no relation to Smith</ref> Three<ref name=Smith2015p1/> The third instance shows the more typical way of doing this, instead of a <ref></ref> pair you can close the tag off internally.

One Two Three

Personally I prefer the {{sfn}} method. With ref tags you need to invent a name which is unique and which you reuse as appropriate. Too much work for late night sessions, {{sfn}} does all the work for you! Martin of Sheffield (talk) 20:12, 3 January 2015 (UTC)

In the (rare) cases that I use {{sfn}} to cite two different items on the same page, I would use the |loc= parameter instead of |p=, as in {{sfn|Smith|2015|loc=p. 1, paragraph 2}} {{sfn|Smith|2015|loc=p. 1, paragraph 4}} or for a newspaper {{sfn|Smith|2015|loc=p. 1, col. 2}} {{sfn|Smith|2015|loc=p. 1, col. 4}} - these are treated as distinct refs because the |loc= parameters differ. --Redrose64 (talk) 20:25, 3 January 2015 (UTC)

Thanks both - I understand now. I had assumed that there was a piece of logic that separated out the footnotes where the author, date and page were not the same, and it would therefore be quite straightforward to add "and if ps is different too". Now you've explained how it works I appreciate that it isn't that simple. The workaround you've explained seems the way forward when this is needed. I'm a big fan of sfn in general too - a very elegant footnoting solution. Apologies for the awkward question! JimmyGuano (talk) 20:40, 3 January 2015 (UTC)

No author name in citation template

In section No author name in citation template it lists examples of referencing by book title or periodical title rather than author, but there are two issues.

1. Book and periodical titles are ordinarily in italics. The examples show italics in the main footnote but not in the short footnote. It should say definitively whether short footnotes should have book and periodical title in italics, and the examples should be consistent with this.
2. Some Wikipedia articles may reference multiple articles in the same periodical, some with authors and some without. For consistency, it may be desirable to list all such articles by publication. For example, 1992 European Community Monitor Mission helicopter downing#News reports lists 5 articles in The New York Times, one of which comes from Reuters and does not have a listed author, and the article references all news stories using short footnotes referencing the periodical name even if the author is listed in the long footnote, which I think is stylistically more pleasing than using authors for some news stories. I edited this article to italicize periodical names —Anomalocaris (talk) 21:32, 1 July 2014 (UTC)

Please see Template talk:Sfn#Title-Date with wiki markup in link above, which discusses work titles in sfn. Clearly the preference is for italics for work titles and I would agree with that. If you are working a lot with sfn etc it would be a good idea to install User:Ucucha/HarvErrors.js. If you do that you will see several red messages reporting problems in the article you mentioned. It is rather difficult to catch these without some automated help. --Mirokado (talk) 01:08, 2 July 2014 (UTC)

Mirokado: Thank you for your comment. It's nice that there has been some discussion of book or periodical titles in italics on this talk page, but I am talking about the lack of discussion of this topic at Template:Sfn#No_author_name_in_citation_template, where it is needed. —Anomalocaris (talk) 10:20, 4 July 2014 (UTC)

I posted this under {{citation}}, but perhaps the experts here could have a think?

I've recently been working on Tir national and am trying to tie up the citations. One problem that has surfaced is the lack of authors for web sites, particularly those in a foreign (to me) language. In order to get an anchor for short form footnotes, I have simply forced the ref to be the title: ref = CITEREFTir_National for example. In the absence of an author the title displays first and so both the electronic linkage and the human's view agree: ^Tir National looks right.

The problem comes with long names. A title such as: "Ancienne caserne Prince Baudouin, dite également caserne Dailly" can hardly be called SHORT form! Basing the anchor on a subset works electronically but is confusing to the reader. What I'm trying is putting (in this case) "History:" before the citation, then using ref = CITEREFHistory and {{sfn|History}}. I did look at the display options, specifically author-mask, but that only works when there is a real author. Advice/comments please! Martin of Sheffield (talk) 14:32, 22 January 2015 (UTC)

@Martin of Sheffield: {{Sfn}} is for use when different pages or sections in a work are referred to. To use it, you only give the full citation to the work once (either the first time it appears in the article or in a separate Bibliography section) and then use expressions like {{sfn|Smith|2010|p=100}}, {{sfn|Smith|2010|p=110}}, {{sfn|Smith|2010|p=220}}, etc. to refer to different pages in the work. You don't need to do this for web citations as they don't have pages or chapters, so you don't need {{sfn}}. Just use <ref name=SOMENAME>FULL WEB CITATION</ref> the first time and then <ref name=SOMENAME/> thereafter. Peter coxhead (talk) 17:29, 22 January 2015 (UTC)

@Peter coxhead: We seem to have two threads on this; it's also under discussion at Template talk:Citation#Anchors. --Redrose64 (talk) 17:40, 22 January 2015 (UTC)

translate page ranges to ndash

I use sfn as my primary cite markup. Recently I was admonished because I used "minus signs" in page ranges instead of ndash. Is there any reason sfn couldn't translate these for me? Maury Markowitz (talk) 22:12, 18 February 2015 (UTC)

My guess is that you actually used hyphens ("-"). The minus sign ("−") is different from the en-dash ("–") and em-dash ("—"). Anyway, one reason is that sometimes you want to refer to a single page that has a hyphen in its number. There is no good way of distinguishing those correctly formed but hyphenated pages from mistyped page ranges. —David Eppstein (talk) 00:46, 19 February 2015 (UTC)

Sure there is, pp vs. p. The alternative requires significant work for basically zero benefit to the end user. Maury Markowitz (talk) 15:55, 19 February 2015 (UTC)

User:Ucucha/HarvErrors

I think it would be helpful to to add to the template documenation that the tool at User:Ucucha/HarvErrors will give an error message when the coding is wrong. Dudley Miles (talk) 18:57, 3 March 2015 (UTC)

I don't think having a citeref that is never linked to should be classified as an error. That would prevent many/most uses of {{citation}}, for instance, which automatically makes citerefs. But the warning about links that don't go anywhere does look useful. —David Eppstein (talk) 19:23, 3 March 2015 (UTC)

I agree that it is annoying that the tool gives false error messages when citation is used without harv refs, but it is extremely useful in picking up errors. Dudley Miles (talk) 20:03, 3 March 2015 (UTC)

I made a copy in my userspace at User:David Eppstein/HarvErrors.js with only the messages about missing link targets, and not about unused link targets. (The other difference is that I write-protected it so that vandals can't inject bad javascript into your Wikipedia page views.) I'm already finding it very useful. —David Eppstein (talk) 00:16, 4 March 2015 (UTC)

You don't need a customised script. You can do it using the normal script: all you need to do is to add

span.citation strong.error { display: none; }

to Special:MyPage/common.css. --Redrose64 (talk) 11:31, 4 March 2015 (UTC)

Thanks very much Redrose64. Brilliant. Dudley Miles (talk) 12:34, 4 March 2015 (UTC)

Request TemplateData

Would someone who knows what they're doing be willing to add TemplateData to sfn? I'd try it myself but I don't like experimenting with such a widely used template. As far as I can see there's no TemplateData currently; no parameters descriptions appear in the VE template dialog. Mike Christie (talk - contribs - library) 10:44, 16 April 2015 (UTC)

The documentation is mostly in {{Harvard citation documentation}} which is shared with several others. --Redrose64 (talk) 14:33, 16 April 2015 (UTC)

The reason I asked is that in adding an sfn earlier I saw no template data come up in VE. Would that mean I just have to copy the relevant portions of the Harvard citation documentation over to sfn's TemplateData? Mike Christie (talk - contribs - library) 14:50, 16 April 2015 (UTC)

What relevant portions would those be? --Redrose64 (talk) 16:49, 16 April 2015 (UTC)

I'm not familiar with sfn (I'm trying to learn how to use it on an article I'm working on) so I don't know what the most commonly used fields would be. To see what I'm seeing, edit an article in VE, choose Insert -> Template and put "sfn" in the template name field, then click "Add template". The resulting dialog has no fields displayed, so I've no idea what to put in. If you do the same with "cite book" instead of "sfn", you'll see numerous fields pop up for the fields in cite book that are most often entered. That's what I want to have happen with sfn. So I guess the answer to your question is "whatever fields a user of sfn would expect to be defined"; I'm not really sure which they are because I don't know my way around sfn yet. Mike Christie (talk - contribs - library) 17:21, 16 April 2015 (UTC)

I never bother with "Add template" - to see what parameters are valid, I use the template's own documentation. In brief: {{sfn}} recognises eight parameters, but three are mutually exclusive, so a maximum of six may be used simultaneously. Five of the six are positional (unnamed) parameters: the surnames of up to four authors, plus a year. The remaining parameter may be either |p= |pp= or |loc= and these last three correspond directly with the |page= |pages= or |at= parameters in {{cite book}}. For working examples see NBR 224 and 420 Classes, but you've got the right idea here. Are you experiencing some difficulty with the usage? --Redrose64 (talk) 19:44, 16 April 2015 (UTC)

That edit was basically copied from another one, but I had to do it in the wikitext editor. I'd really like to use the visual editor if I can; I find it more productive, but I don't think it's possible to use a template in the visual editor if there's no templatedata for it. You're right, I can figure it out the old-fashioned way; I was just hoping to use VE. Mike Christie (talk - contribs - library) 19:52, 16 April 2015 (UTC)

Is it the case that parameter 1 is the author name, parameter 2 is the year if no parameter 3 is given, but is another author if there is a parameter 3, and so on up to parameter 5? I understand there are named parameters mentioned in the documentation; I'm just trying to get a clear definition of the interpretation of the unnamed ones so I can create the TemplateData for the basic ones. Mike Christie (talk - contribs - library) 10:04, 17 April 2015 (UTC)

Yes. The unnamed parameters are between 1 and 4 author last names (which must match the corresponding author last names in the citation to which it links) followed by a year (matching the year of the date in the citation to which it links). So the formats can be {{sfn|last1|year|...}}, {{sfn|last1|last2|year|...}}, {{sfn|last1|last2|last3|year|...}} or {{sfn|last1|last2|last3|last4|year|...}}, where ... represents the optional named parameters. Note that there are two other possible named parameters: |ref= and |ps= or |postscript=. {{Sfnp}} has the same parameters, but puts parentheses round the year on output. Peter coxhead (talk) 10:43, 17 April 2015 (UTC)

Thanks; that's what I thought. I'll see if I can build the TemplateData with that. Mike Christie (talk - contribs - library) 10:48, 17 April 2015 (UTC)

Sfn nested

I'd like to use {{sfn}} within a footnote (e.g., {{refn|Blah blah blah footnote side conversation.{{sfn|Author|2004}} }} but the sfn doesn't stack with other instances used throughout the article. Instead it dumps the ref in its own automatic section at the bottom of the page. What are my options for handling this? I'd prefer if this footnote's footnote pointed to the "Author 2004" short entry (in reflist) alongside the others. – czar 17:15, 24 May 2015 (UTC)

Have you tried {{efn|blah}} instead of refn? It does not seem to cause the problem. Dudley Miles (talk) 17:20, 24 May 2015 (UTC)

I have, and it solves the issue, but I'd prefer to have the footnote appear as a normal ref in the {{reflist}} rather than having to use {{notelist}} as well. I thought {{refn}} is supposed to be specifically for nested refs? It might have something to do with how refn is processed. The example article is Albany Free School. No matter where my refn rests in the article, it won't place the nested sfn in the main reflist. Is there any way to have the same effect as sfn (with the short harv style, direct link, and stacking instances) with just the <ref> tags? – czar 17:37, 24 May 2015 (UTC)

There is a related discussion at Template talk:Refn#Refn references getting sent to the bottom of the page?, it's not a problem with either {{sfn}} or {{refn}}. Have a look how I did it at LB&SCR A1X Class W8 Freshwater#Notes. --Redrose64 (talk) 18:15, 24 May 2015 (UTC)

Was just going to post that it is more of a refn than a sfn thing. Will follow up there. – czar 18:20, 24 May 2015 (UTC)

Shortcuts

Would it be useful to reserve {{s}} and/or {{sf}} as {{sfn}} shortcuts? They would be easy to repurpose (not in high use) and would save us a few keystrokes and screen real estate on a high use template. – czar 18:39, 24 May 2015 (UTC)

Save one or two characters? On a template that is already considered over-cryptic by its use of a three letter name? No thanks. --Redrose64 (talk) 19:07, 24 May 2015 (UTC)

Protected edit request on 26 May 2015

We need the code for Be.wiki.org. Where could we see it? Belarus2578 (talk) 10:31, 26 May 2015 (UTC)

@Belarus2578: Not done: this is the talk page for discussing improvements to the template {{Sfn}}; it's not clear what changes you want to be made. Please mention the specific changes in a "change X to Y" format. --Redrose64 (talk) 11:48, 26 May 2015 (UTC)

Documentation for pp=

The template page doesn't explain or provide examples of the pp parameter. How do you specify a page range? With a dash? ndash? comma? This needs to be documented, and, ideally, one of the examples should use it. — J D (talk) 14:51, 24 July 2015 (UTC)

Group as namespace vs. visual style selector

This is probably the wrong place to ask, but since I'm at a loss as to the correct venue and Sfn was what I was trying to use to solve my problem, and I expect at least some interested editors might be watching here, I'm taking a chance y'all can at least point me in the right direction.

The case is this: I have a navbox that contains encyclopedic content (a family tree), and thus needs to be verifiable. Since the navbox is transcluded onto several articles in mainspace, just stuffing Sfn+Reflist in the template will interfere with references on the page it is transcluded into. My approach to solving this was to try to put the navbox's refs into a separate group. However, the first problem I ran across is that Sfn doesn't appear to support a |group parameter. The second was that there's no apparent way to supress wikilinking the short footnote (the links form the navbox will also conflict with those of the page it's transcluded onto).

So my next attempt was to use {{Efn}}, which does support a |group parameter. However it seems to overload it so that setting it affects both which "namespace" (in CS terms, not Wiki terms) the refs are in and how the reference markers are displayed. And, crucially, for that reason restricts the valid arguments to "lower-alpha", "lower-greek" etc. (the predefined ones). Which means the navbox can't invent its own namespace to reduce the chance of collisions with the pages it's used on.

And finally, I've resorted to just plain <ref group="foo"> (which I, incidentally, would prefer not to for various reasons anyway), but find that this ends up creating footnote markers of the form "foo 1" (or "navboxwsft" in my first attempt). That is, the plain ref tag overloads the |group parameter to mean both which namespace to use and how to display the marker.

In summary: *sigh*

So… What my wishlist contains just now is the following:

• A way to suppress generating the wikilink
• A way to place the ref in a separate (CS) namespace, that has no side effects (not overloaded; does not affect list-style-type or prefix the footnote marker)
• A way to choose a list-style-type that matches {{Reflist}}, i.e. such that Sfn and Reflist can be set to display lower-greek (or whatever)
• A way to tune the footnote marker prefix / display distinct from the namespace, for instance |group=randomstring and |note-prefix=nb. (should be able to combine with list-style-type, e.g. "nb. 1" or "fn. α")

Am I hopelessly confused here? Did I miss anything? And is this the wrong venue for this question? If so, can anyone point me at the right place?

PS. For those who are curious (or just want to point and laugh), the template in question is {{William Shakespeare family tree}} (transcluded on several high-visibility articles, so edit with care!). --Xover (talk) 06:14, 2 September 2015 (UTC)

I don't really understand why you think you need groups. This is intended to be used as a navbox at the end of an article, after the references section of that article, right? So if you just use <ref>/{{sfn}} and {{reflist}} with no group name, all the previous references should be cleared by the reference section, and you can get your own set of footnote numbers that don't interfere with anything. That is, if you have an article containing

Article text with references

Followed by a references section with {{reflist}}

References

Your navbox with its own references and references section

References

then the two sets of references should have their own numbers and not interfere with each other, automatically. Note that even though they get overlapping numbers, they are assigned distinct html href tags, so there is no ambiguity about what your browser should show you when you click on one. —David Eppstein (talk) 06:39, 2 September 2015 (UTC)

Hmm. Well, the reason I noticed the problem in the first place was that I tried to transclude it in the middle of the page Shakespeare's life, which currently contains a subst'ed copy of a (very) old version of the template (in #Shakespeare genealogy). There's also the case of (the right name escapes me) the infobox-style navboxes used in the top right of articles. And in general engineering terms, depending on element order in the page seems very brittle (high probability of errors occurring, with a bad failure mode, and with disproportionate consequences).
However, there is also the distinct possibility that I'm just hopelessly confused about something here. If so, I would much appreciate elucidation. --Xover (talk) 07:00, 2 September 2015 (UTC)

Why can you not simply use a normal {{sfn}} in the template? In the article, the refs will show up in the main reflist. --Redrose64 (talk) 10:32, 2 September 2015 (UTC)

Well, in that case they will show up in the main reflist in some cases (when the template is used before the References section), but not in other cases (when the template is used after the References section). In addition, having a navbox insert references in the main article's reflist is surprising and is likely to not be popular with the editors of the transcluding article (which may or may not overlap with the template's editors). --Xover (talk) 11:03, 2 September 2015 (UTC)

You said that this was at Shakespeare's life, where the family tree is before the refs, so there's no problem that I can see. If you want a referenced family tree, just add them. If you want the family tree to be moved after the {{reflist}}, don't put references in it. --Redrose64 (talk) 11:12, 2 September 2015 (UTC)

Well, no, the navbox will have multiple uses. In some cases it will be at the end of the article (after the invocation of {{Reflist}}) and in some cases it will be in the middle of the article (before the invocation of {{Reflist}}). And I didn't particularly want to bother with refs in this particular family tree (for various resons), but some editors have (recently, I think) started to tag family trees in particular as unreferenced. And their position is, to a degree, persuasive, and, as best I can tell, in line with policy (WP:V). i probably wouldn't have bothered on my own initiative, but under the circumstances I felt adding refs was the right option. --Xover (talk) 11:34, 2 September 2015 (UTC)

Navboxes are not referenced. If you want a family tree to be treated as a navbox, don't put refs into it. If you want it to be treated as article content, move it before the references section. --Redrose64 (talk) 11:41, 2 September 2015 (UTC)

Well, the problem would apply to any template that includes arguably encyclopedic content, not just navboxes, and, I would argue, having a template insert references into the transcluding article's References section (possibly, maybe even probably, in a divergent citation style) would be unwanted behaviour that is unlikely to be popular with any random set of editors. There are, I'm sure, various ways to work around it with various levels of cludgeiness, but at least to my mind it would be much prefereable if {{Sfn}} or a sibling contained the necessary scaffolding to enable this kind of use case when needed. --Xover (talk) 12:21, 2 September 2015 (UTC)

Interaction with new cite error detection

The discussion at Wikipedia:Village pump (technical)/Archive 141#Citation error is relevant to editors and users of this template. Please join in. DES ^(talk) 23:19, 23 October 2015 (UTC)

Redrose64 suggested, in the above linked thread, that the use of the |ps= parameter for quotes should be deprecated. For one thing if the order of text is changed so that the call to sfn with the quote is no longer first in the article, the quote will cease to be rendered, that is it will silently disappear from the article. Use of {{harvnb}} does not have this problem. Does anyone have any views on the point? DES ^(talk) 00:28, 26 October 2015 (UTC)

@Redrose64, Czar, Xover, David Eppstein, J D, Dudley Miles, Maury Markowitz, Peter coxhead, JimmyGuano, Martin of Sheffield, and Mike Christie: You have all commented in this talk page within the last year. Please take a look at the issue described above, and give your views on whether the use of ps= for quotes and other such notes should be deprecated, or some different action taken. Thank you. DES ^(talk) 15:25, 27 October 2015 (UTC)

Not much I can add I'm afraid, I have not use PS, and looking over the dox I'm not sure I understand where it would be used - aside from the "off label" use mentioned above. It does not effect me personally, so as far as that goes I cannot offer an argument against change. Maury Markowitz (talk) 15:32, 27 October 2015 (UTC)

It would have been nicer to have spelled out the whole story in the summary, but I skimmed the thread anyway. If circumstances have changed such that |ps= will no longer (and never again) perform as it has been expected to function, certainly those citations should be reformatted to work as needed. I know that I've used that parameter many, many times, so I'm curious what the game plan would be for reformatting those extant template calls (namely who would be doing the updating).  czar 15:40, 27 October 2015 (UTC)

Sorry, czar the WP:VPT thread was rather long. I have been doing some AWB edits, replacing calls to {{sfn}} that include ps= with more than just a period with calls to {{harvnb}} inside explicit ref tags. I've done perhaps 15-20 so far, and that covered more than 2000 of the not quite 25,000 articles that transclude {{sfn}}. (i've only been looking in mainspace, I suppose a run through draft: would be appropriate.) I would welcome review of these edits, they are plainly indicated by edit summary in my recent contributions. It looks to me as if this usage is actually rather rare. I found one article with a call to {{sfnm}} including multiple quotes, that I'm not quite sure how to deal with: Argentina. Anyone who cared to edit that to fix the issue there would be doing a good thing. DES ^(talk) 16:42, 27 October 2015 (UTC) @Czar: DES ^(talk) 16:44, 27 October 2015 (UTC)

To clarify a bit, {{sfn}} generates and outputs a ref with a refname, the same refname for all calls with the same author, year, and page. However, if ps= is used to include a quote (as the doc currently recommends) the new cite error detection logic will see that multiple instances of the same refname have different content (i.e one has the quote, one doesn't) and will output a large red error message. Even before this change, such use was risky, because if an edit moved sentences or paragraphs so that the sfn call with the quote was no longer the FIRST sfn for that author/year/page combo, the quote would simply and silently vanish from the rendered article, with no warning. ({{harvnb}} doesn't have these issues, because it doesn't generate a refname, the editor must put it inside ref tags.) I hope that is clearer. DES ^(talk) 16:52, 27 October 2015 (UTC)

Ahhh, I see the problem now. But I have to ask, is this the purpose of PS, or simply a behaviour that people were using because it worked? I've never been a fan of in-ref quotes, so I've never run across this. Maury Markowitz (talk) 17:51, 27 October 2015 (UTC)

My .2¢… The |ps= parameter strikes me as badly overloaded. It tries to control CS1 vs. CS2 style, control or supress trailing punctuation entirely, and providing a place to stash a quote. The citation styles issue has only two cases, both of which are fixed in terms of the necessary output; which suggests this switch should be controlled by a boolean parameter. If really needed this could be a tri-state switch to allow for an explicit "none" case. Personally I would just as soon not have to care about that issue. The quote case strikes me as a misfeature: any quotes either belong in the full citation to which the shortened footnote refers, or in a separate Footnotes section (i.e. using {{Efn}}, something like what was done on William Shakespeare). In other words, my position, based on the rather quick spin through the VP thread and the docs just now, is that |ps= should be dropped, to be replaced, if necessary, with a separate parameter that does the CS1/CS2/trailing-punct stuff; and the documentation should refer people to {{Efn}}/{{Notelist}} for quote-like uses. It'll inconvenience those (few, I would guess) who like to include quotes with their cites, but not worse than that they can work around it using another citation template if needed. If one felt the need was far greater than I believe it to be, one might explore some structured way to connect a footnote with a cite (logically) rather than make editors use recursive templates (Sfn inside Efn, or Efn inside <ref>Harvnb</ref>). I've wished for that a few times (I like to cite my footnotes, don't judge me) and would, IMO, be a far better option than adding a quote to a short footnote. One could even envision a wrapper template that takes a quote and author/year/page in and spits out the equivalent of {{Efn|The relevant quote.{{Sfn|Smith|2015|pp=25–6}}}}. Maybe something like {{Mlfn|Smith|2015|pp=25–6|The relevant quote.}} --Xover (talk) 18:08, 27 October 2015 (UTC)

I am not technically expert, so apologies if I have got it wrong, but I would deprecate the use of ps= for quotes in {{sfn}}. As I understand it, you should use {{efn}} for notes - including quotes which you do not want to put in the main text - and {{sfn}} for citations, which should not include quotes.

BTW ps= should be useful in some templates, and I tried to use it in {{cite book}} to meet a situation which the template did not cover - quoting a new introduction to a book by a deceased author, but the problem is that it leaves no space after the previous field. An example is Keynes 2003 in Æthelwulf, a new intro to a book by Blair, which I now have at FAC. I tried putting in a hard space, but that did not solve the problem, and a reviewer changed it so that it now appears as if Keynes wrote a chapter in a book edited by Blair, which does not seem satisfactory to me. Dudley Miles (talk) 18:39, 27 October 2015 (UTC)

This is an interesting case. To give both Blair and Keynes as authors (|lastN=|firstN=) would be misleading; Keynes only contributed an introduction, which only barely counts as a chapter much less co-authorship of the whole work. However, the only available way to make this present sensibly is to add the original author as if he were the editor (|editorN-last=/|editorN-first=), and this generates incorrect metadata (granted the current Z39.88 data is a bit limited, but...). For most compilations or compound works this isn't a problem as it's either a straight co-authorship, or a editor vs. author relationship. But introductions are far more common and does not fit into either model.
I think that presentation wise, the right thing would be for Blair to be given as author but presented as the editor is (the intro by Keynes. In Blair, Peter Hunter etc.). That is, either a new |somethingN-last=/|somethingN-first= parameter that captures the sense of "the real author of a compound work", or a switch that treats a specific |editorN-* as if he were an author in the metadata (|editorN-isauthor=true, or something like that). All these options look kinda ugly to me, but would at least be less wrong than what's necessary currently.
Anyways, that's an issue for a different talk page (or maybe the technical Village Pump?), I suspect. --Xover (talk) 19:29, 27 October 2015 (UTC)

An alternative would be introduction-first and introduction-last parameters, which would be shown at the beginning of the citation, and insert "Introduction to " before the author of the book. Dudley Miles (talk) 00:05, 28 October 2015 (UTC)

I was pinged, but also don't have much to say, as I haven't used quotes in sfn myself. This is a sad situation, in which documented and useful existing functionality (the ability to supply quotes in short footnotes and then reuse the footnote without repeating the quote) depended on generating invalid html, and cannot be made to work correctly in any reasonable way. It's tempting to suggest "fixing" the error by making the quote modify the refname (so that the sfn's that don't have the quote become separate from the one with the quote) but that would be wrong because it would lose the association between what were originally intended to be copies of the same reference. I think the current behavior, of throwing an error and making an editor clean it up (for instance by switching to the more traditional <ref name...> syntax), is probably the best way forward. —David Eppstein (talk) 20:13, 27 October 2015 (UTC)

No way that a "short cite" (as done with {{Harv}}) should contain a quotation. (Or, for that matter, any other commentary.) But as David says, it can done within the <ref>...</ref> tags, which define a note, and can contain short cites and quotations. If editors need sfn to wrap these together then perhaps a |note= parameter could be used to include all the quotations, comments, etc., that don't belong in the citation itself. ~ J. Johnson (JJ) (talk) 23:49, 27 October 2015 (UTC)

Another parameter would presumably be for something that is intended for display, but that would cause exactly the problems that |ps= is causing, for the simple reason that it won't be part of the ref name. See my comments of 08:25, 25 October 2015 at Wikipedia:Village pump (technical)#Usage of ps= in .7B.7Bsfn.7D.7D. --Redrose64 (talk) 00:24, 28 October 2015 (UTC)

JJ's misguided crusade to separate quotes from citations aside, I agree that deprecating |ps= (and only allowing parameters like |p= and |pp= that modify the refname) seems to be the correct way forward. —David Eppstein (talk) 01:01, 28 October 2015 (UTC)

What is misguided is including quotations as part of the citation ("citation" being strictly construed). I believe the reason for enabling such was so editors could provide a basis for paraphrasing done in the text. The goal was laudable, but splicing it into any of the citation templates was quite misguided. As to how refnames should be generated, that really needs a broader consideration. ~ J. Johnson (JJ) (talk) 22:22, 28 October 2015 (UTC)

I strongly disagree. Indeed in helping editors at the Teahouse with new drafts, I frequently advise them to use the |quote= parameter defined in the various Cite XXX tempaltes, largely for offline sources and for non-English sources (In which case the quote will be on contain a translation. The purpose is to allow a reader to verify that the source content supports the statement(s) in the article, and how it does so. It is not perfect, of course, but it can be very helpful. Of course, i never advise a new editor to use sfn or any form of Harvard referencing. DES ^(talk) 22:39, 28 October 2015 (UTC)

Please note: I entirely agree with the purpose of providing quotations, and favor doing so. No argument on that! But bastardizing citations (using any template, or none) for that purpose is misguided, as it confuses the concept and practice of citation, and leads to these complications as we are working on here. Especially as there is no need to merge a citation with a quotation. Quotations in support of paraphrases, like any other commentary about a source, can be included in the note with the citation. ~ J. Johnson (JJ) (talk) 23:09, 28 October 2015 (UTC)

Huge lists of errors reported

In some articles (well, one that I've looked at) there are enormous lists of Cite error: Invalid <ref> tag; name "GoogleIO2011" defined multiple times with different content (see the help page)., associated with {{sfn}}. Specifically look at article List of plain English words and phrases; as of now it has a colossal list of sfn errors. Entries are of the form {{sfn|SEC|1998|p=27|ps=: "negative compound"}}. This may well be an instance of #Interaction with new cite error detection, but I'm not familiar enough with the template and its changes to say. But I would say that examples like this are significantly detrimental to Wikipedia as seen by a casual reader; either the template or the articles need to be fixed; I think the template is the obvious candidate. I did a Wikipedia search for "Invalid <ref> tag; name" and got 34,907 hits; many of them are from copy-and pasted duplicates like <ref name="Any name">[http://wibble...]</ref>, but I expect sfn contributes plenty of cases. If this is due to recent changes in sfn I'd suggest putting it back the way it was, and not changing until proposed changes have been tested. [Added later: if this is new, it seems reasonable that it's due to improved detection of duplicated references rather than a change to {{tl:sfn}} as such, so reverting to an older sfn won't help. I won't speculate further about solutions, but just point out the problem.]

This might be something that's just gone wrong and is actively being corrected that I've just stumbled on at the worst time, in which case ignore this. But I would suggest that the spurious error messages need to be got rid of ASAP, even if this delays a real improvement. Best wishes Pol098 (talk) 13:41, 30 November 2015 (UTC)

Something similar may have occurred at All Saints Church, Frindsbury. The three refs {{sfn|Hope|1883|ps=: quoted by Wade.{{sfn|Wade|1990|pp=6–7}}}} were cut-and-pasted, so were identical and yet the reflist gives Cite error: Invalid <ref> tag; name "FOOTNOTEHope1883" defined multiple times with different content (see the help page). I've had to change them to a footnote and two refs to Hope - which is not WP:SAYWHEREYOUGOTIT Martin of Sheffield (talk) 14:19, 30 November 2015 (UTC)

Rechecking the current List of plain English words and phrases [the permanent link to the old version in my comment above was wrong, now corrected], I see that User:Kanguole has usefully fixed all the references by changing {{sfn|Garner|2009a|p=371|ps=: "formal word"}} to <ref>{{harvnb|Garner|2009a|p=371}}: "formal word"</ref>, etc. This would seem to need doing throughout Wikipedia, if the duplicate reference name checking and {{sfn}} are not themselves changed. Pol098 (talk) 16:27, 30 November 2015 (UTC)

@Pol098: the point is that they were errors, but weren't previously flagged as such. {{sfn|Smith|2015|p=12|ps=Text1}} and {{sfn|Smith|2015|p=12|ps=Text2}} link to the same footnote; two different values for |ps= don't generate two different footnotes. Peter coxhead (talk) 16:51, 30 November 2015 (UTC)

@Peter coxhead: Thanks, I had an idea something of the sort was happening (there's been a lot of discussion, which I haven't followed, on the use of "ps=" with sfn). I don't know what to suggest, but continue to maintain that articles with lots of red error messages (however justifiable they may be) are not good for the status of Wikipedia as a source of useful information. Is there any ready-made tool to change from {{sfn}} to {{harvnb}}, as done very recently to List of plain English words and phrases? It's a fairly simple macro for any text editor. Pol098 (talk) 17:08, 30 November 2015 (UTC)

@Pol098: This is the same as #Interaction with new cite error detection above, and the (archived) VPT thread linked from that. --Redrose64 (talk) 23:42, 30 November 2015 (UTC)

Is there any update on this? I've just run into the problem on an article where multiple {{sfn}} references to different pages of the same book are generating the same error, and would obviously like to fix it.  — Scott • talk 15:36, 2 January 2016 (UTC)

@Scott: Shouldn't happen if the pages are different. What is the article? --Redrose64 (talk) 21:47, 2 January 2016 (UTC)

@Redrose64: Interesting, maybe I'm reading the source wrong. It's M11 link road protest#References. Thanks for any suggestions!  — Scott • talk 22:32, 2 January 2016 (UTC)

(i) Don't add an empty |ps= (ii) if you must use |ps= do it sparingly, and never do it in {{sfn}}. This edit should have sorted it. --Redrose64 (talk) 23:37, 2 January 2016 (UTC)

Great - thanks very much. Somebody else had added those since the last time I looked at the article, for some reason.  — Scott • talk 04:26, 3 January 2016 (UTC)

Misleading error message

There's an odd error message at this old version of "Botany" – see References, item 173. The error message is "Cite error: Invalid <ref> tag; name "FOOTNOTECampbellReeceUrryCain2008739" defined multiple times with different content". It seems to be caused by the use of both |p= and |pp= with the same single page number (or at least it was fixed when Materialscientist corrected the "pp" to "p"). What's odd is that I can't reproduce this error message. Extracting the three uses of {{sfn}} with page 379 from the article gives:

• {{sfn|Campbell|Reece|Urry|Cain|2008|pp = 739}}
• {{sfn|Campbell|Reece|Urry|Cain|2008|pp = 739}}
• {{sfn|Campbell|Reece|Urry|Cain|2008|p = 739}}
• Campbell, Neil A.; Reece, Jane B.; Urry, Lisa Andrea; Cain, Michael L.; Wasserman, Steven Alexander; Minorsky, Peter V.; Jackson, Robert Bradley (2008). Biology (8 ed.). San Francisco: Pearson – Benjamin Cummings. ISBN 978-0-321-54325-7. {{cite book}}: Invalid |ref=harv (help)

References

Any ideas? Peter coxhead (talk) 12:08, 12 January 2016 (UTC)

I see the message here. It is indeed caused by using both |p= and |pp= for the same single page. There's a lengthy thread at Wikipedia:Village pump (technical)/Archive 141#Citation error, but what it boils down to is: always use |p= for single pages; use |pp= only when multiple pages are being specified. --Redrose64 (talk) 12:18, 12 January 2016 (UTC)

I don't see the message on this page, but it is in the raw HTML, so presumably Peter and I have CSS settings that hide these messages in some namespaces. Kanguole 12:32, 12 January 2016 (UTC)

See H:SHOWCITEERROR. --Redrose64 (talk) 12:49, 12 January 2016 (UTC)

Ah, right; I see why I didn't see the error here. But I still think the error message is misleading and should be fixed; it's a separate case from different content in |ps=. Peter coxhead (talk) 13:57, 12 January 2016 (UTC)

As explained at that VPT thread, it's a message from the MediaWiki software. Nothing we can do about it except use {{sfn}} in accordance with its own documentation, which is "For single pages, use |p=; and for multiple pages (such as ranges), use |pp=. Use of |p= for multiple pages or |pp= for single pages can cause cite errors." --Redrose64 (talk) 14:10, 12 January 2016 (UTC)

I understand that a template/module cannot know about its previous invocations or look ahead, so that {{sfn}} can't trap this particular error, but we could, for example, issue a warning when |pp= is used without a dash or hyphen in its argument, which would be more helpful than the bare MediaWiki error message. Peter coxhead (talk) 15:02, 12 January 2016 (UTC)

Muiltiple sfns with a single "ps:" field

Subhas Chandra Bose has cite errors due to it stating "defined multiple times with different content " because of one of the cites having the ps field. Anybody knows what needs to be done here? There are four such instances. Ugog Nizdast (talk) 15:28, 19 January 2016 (UTC)

I've just had a quick look at the page and I'm afraid to say that the referencing needs a thorough overhaul. In general I would advise lengthy quotations to be entered as a separate note, and only once. Use {{efn}}. Are all those quotes needed? If they are relevant then perhaps they should be incorporated in the text, if not then merely allowing the reader to find the text ought to be enough. The citations list (from {{sfn}} is then just a list of cross references giving author, date and page to enable the reader to find the work in the bibliography.

I'll happily go through and knock the referencing into shape if you would like, but may not be able to until next week. Of course if someone else wants to jump in - go ahead. Martin of Sheffield (talk) 16:01, 19 January 2016 (UTC)

Those quotations were added by the major contributor due to it being a controversial topic. I agree they may be excessive but have no idea what's relevant or what can be trimmed. Yes, please do separate them into efns and sfns. Ugog Nizdast (talk) 16:27, 19 January 2016 (UTC)

Thanks for the "heads up" about the quotations and the controversy. Have a look at Emilie Schenkl to see how I've changed it. Martin of Sheffield (talk) 17:06, 19 January 2016 (UTC)

Don't cite quotations for verifiability sake need to paired with their respective citations? as opposed to explanatory footnotes which don't have to be? Afterall for what is that quote field present in citation templates? Ugog Nizdast (talk) 12:50, 20 January 2016 (UTC)

Good point, I had accidentally left off the citation at the end of the quote. It is now fixed. There are a number of issues here:

1. The most important issue is WP:RF. What does a high school reader in, say, Chile need and expect to find.
2. Next consider WP:VERIFY. Controversial statements need to be referenced to a source. Using Harvard referencing the source's bibliographic information is displayed once, the references guide a researcher to the source material and usually only should be the name, date and page.
3. Now consider quotations. My feeling is that:
   1. If the quotation is relevant to the article and highlights an important controversy then it should be summarised and moved into the main text. Use {{harvtxt}} to set up lines such as: Jones the Sage (1097) claimed that the sun was a red hot lump of coal, but Newton (1668) showed it was much larger.
   2. If the quotation adds to the article, not just justifies it, but should not be in the main text then a footnote is a mechanism for displaying it and keeping the main cross-reference section clean.
   3. If the quotation merely justifies the statement, then just cite it and be done. In most articles quotations seem to be from a phrase up to a line or two. Much more than that and the only people who will read it are the ones who will go back to the source material anyhow.
4. Finally, vast swathes of text embedded in a paragraph makes things harder for editors. So does repeating the same text in multiple places. Both stretch the markup and in some cases will break it. All of this makes editing and maintaining the page harder, particularly for inexperienced editors.

HTH, Martin of Sheffield (talk) 13:41, 20 January 2016 (UTC)

Thanks for your informative reply. I guess putting quotes along with the citation should be used sparingly when you can just use notes and put the citation in it, especially seeing that it makes the article hard to navigate for verifiability and hard to edit too. Look forward to your edits at Bose. Ugog Nizdast (talk) 16:56, 20 January 2016 (UTC)

The various cite templates do provide a |quote= parameter, but {{sfn}} intentionally does not, since it is intended for shortened footnotes, hence its name. Using the |ps= param for a quote is stretching the purpose of both that param and the template itself. --Redrose64 (talk) 20:54, 20 January 2016 (UTC)

Need someone's help in citing two authors

It is not working for me, in my Sandbox8, specifically the third reference, which uses "Riedlmayer & Ostapchuk" , my second reference worked very well and I have used it on Wikipedia. Alexis Ivanov (talk) 04:30, 21 January 2016 (UTC)

It just needs |ref=harv. Kanguole 10:00, 21 January 2016 (UTC)

Thank you very VERY much Alexis Ivanov (talk) 10:20, 21 January 2016 (UTC)

It does raise the question as to why |ref=harv isn't the default in the CS1 templates as it is in {{citation}}. I'm often caught by this. Peter coxhead (talk) 16:19, 21 January 2016 (UTC)

If it was, then we might get cases like Workington Central railway station where there is a ref to Webb 1964, and there are two candidates for that - the relevant one is the October one, listed under "Sources" - this one has |ref=harv, which is absent from the September one listed under "Further reading". --Redrose64 (talk) 17:07, 21 January 2016 (UTC)

More precisely, only one "Webb 1964" is used as a source, so you don't want to use the "Webb 1964a" convention. I don't think there is any problem, as the link goes to the first anchor found, which is the correct one. But if there is a problem, just give the unwanted citation a dummy ref. ~ J. Johnson (JJ) (talk) 22:40, 18 February 2016 (UTC)

Cite templates emit the reference text wrapped in a <cite>...</cite> element. The |ref= parameter, when non-blank, adds an id= attribute to that element. When present, an id= must be unique within a document - it is a breach of HTML standards for two elements to have the same id. --Redrose64 (talk) 00:37, 19 February 2016 (UTC)

|ref=none prevents generation of the element id. Kanguole 11:44, 19 February 2016 (UTC)

So does omission of the |ref= parameter, if you're using anything except {{citation}}. --Redrose64 (talk) 20:35, 19 February 2016 (UTC)

I believe what we are discussing is having {{cite}} behave like {{citation}}. Non-uniqueness of ids doesn't seem like a serious breach, as WP seems quite tolerant of that. But to the same end one can always use |ref=none, as suggested. ~ J. Johnson (JJ) (talk) 22:58, 19 February 2016 (UTC)

I hate to bother you guys again, but I need another help in reference number 9, when I lick on Lowry it doesn't go to the Bibliography unlike the other books, feel free to edit Alexis Ivanov (talk) 05:41, 20 February 2016 (UTC)

|ref=harv again. If you install User:Ucucha/HarvErrors, it will flag these problems for you. Kanguole 10:26, 20 February 2016 (UTC)

Yup. One of the reasons I prefer {{citation}} is so I don't have to keep adding |ref=harv. And avoiding that blush of embarrassment when I am momentarily flustered as to why a link has failed. ~ J. Johnson (JJ) (talk) 20:07, 20 February 2016 (UTC)

Thank you very much, I will install that. Alexis Ivanov (talk) 04:36, 21 February 2016 (UTC)

date=n.d.e

|date=n.d.e – no date (n.d.) What is the "e" for in the documentation on "n.d."? (Also see the archive) czar 18:27, 8 March 2016 (UTC)

This is in the section "More than one work in a year". When there are several citations with the same authors and year (or n.d.), one can add a lowercase letter to distinguish them. Normally "n.d." should suffice. Kanguole 18:43, 8 March 2016 (UTC)

Didn't see the a-f pattern there—thanks (would make sense to put it in alphabetical order)  czar 03:02, 9 March 2016 (UTC)

Statement of prevalence of use

Hi. I don't understand the objection to "This template is used in less than 1% of articles on the English Wikipedia.". Please explain. Thanks. Jytdog (talk) 19:48, 8 March 2016 (UTC)

It's underhand if this is the only citation template picked out in this way, since it's clearly designed to create prejudice against this template. It's also not very meaningful if there's no comparison. The tool tells me that {{Cite journal}} has, as of now, 439,574 transclusions, so is used in less than 8% of articles on the English Wikipedia. {{Cite book}} on the other hand is used almost 15% of articles. What does that mean? Should I conclude that using any kind of citation template is a minority activity, so I shouldn't do it? Peter coxhead (talk) 20:25, 8 March 2016 (UTC)

Agree with PC. The 1% figure makes it sound like sfn is for weirdos and geeks, when in fact it should get wider user -- where appropriate. (I say this as someone who personally dislikes its complexity and clutter -- personally I prefer {{r}} with its p= feature, giving callouts like ^[1]:15, where :15 is the page #.) EEng 21:59, 8 March 2016 (UTC)

how are facts "underhand"? That is like saying NPOV means things must be "balanced" or inexperienced editors saying "but that article says X!" if folks want to add similar facts to other templates, knock yourselves out. I really don't understand this. Jytdog (talk) 22:33, 8 March 2016 (UTC)

I made two relevant comments at User talk:WhatamIdoing#Template:Sfn/doc:

• in relation to this edit: How many other templates "have information about how unpopular their choice is"? Why should this one be singled out?
• in relation to this edit: Some editors do indeed dislike {{sfn}}; to me, your comment give the impression that you are one of those who dislikes it. I don't like {{rp}}, but I don't alter its documentation to indicate my dislike.

and one at User talk:Jytdog#Template:Sfn/doc:

• in relation to this edit: Please revert your edit. Per WP:BRD, WhatamIdoing (talk · contribs) was bold, I reverted, it should now be discussed. Instead, we have three users (the third is Izno (talk · contribs)) going against WP:EW.

the last of these was soon acted upon by Jytdog (talk · contribs) ( Thank you). --Redrose64 (talk) 00:44, 9 March 2016 (UTC)

I find the removed statement misleading for the followig reasons:

• I analysed the article reference styles used by featured articles promoted in September 2014 (see Current referencing practice) and in that month ten out of 23 articles promoted used Sfn (about 43%). Including another four articles which used short callouts created differently, about 60% of the articles used short footnotes in some way. Clearly the percentages will vary from month to month: it is equally clear that experienced editors involved with articles which "exemplify Wikipedia's very best work" (FAC) often choose to use Sfn.
• Sfn is most relevant for articles having multiple callouts with different page numbers per cited work. I imagine that most editors who choose to use Sfn in that case don't use it in articles not fulfilling that criterion, so to gauge their preference accurately we would need to sample articles with the relevant disposition of callouts and citations.

Finally, Sfn will often become a relevant choice some time after an article was created, so it is natural that editors will add it to some articles they are expanding. Not every one doing so should be regarded as a zealot. --Mirokado (talk) 02:11, 9 March 2016 (UTC)

• I agree that sfn is appropriate for some articles, that other forms of short footnotes are appropriate for even more articles (IMO far more than where they get used), and that a minority of very good editors use this specific template enthusiastically (and almost always appropriately, in my experience). However, the mere fact that it's "unpopular" (NB "unpopular" does not mean "bad" in any sense) in typical articles means that editors who are accustomed to this template and who best understand its virtues may benefit from being reminded that the majority of editors – the "typical" editors rather than the folks who write FAs – are not used to this template and do not understand its virtues. I hope that reminding the proponents of the relative lack of use in typical articles will help them figure out ways to talk about and educate "typical" editors, rather than, say, assuming that anyone who doesn't jump up and down at the prospect of a template they've never encountered is merely objecting because of unimportant personal preferences. WhatamIdoing (talk) 02:59, 9 March 2016 (UTC)

  That rationale is underwhelming specific to this template, especially since the statement was buried somewhere in the documentation. Calling it 'unpopular' when many power users are using it is disingenuous. If you think that editors should do more to educate others on the use of a particular template, I'm sure that sentiment could be written up in an essay, or better yet, is IMO already-reflected in WP:BITE. --Izno (talk) 12:34, 9 March 2016 (UTC)

  Can you name any other citation-related templates that are equally unfamiliar to the average editor and cause as many complaints when someone unilaterally changes the article's citation style (e.g., the one that prompted a long discussion at WT:CITE recently)? WhatamIdoing (talk) 07:55, 11 March 2016 (UTC)

  (a) People shouldn't be unilaterally changing cite style; (b) people can discuss choice of styles on the merits and on their own experience. Tyranny of the majority doesn't help. EEng 10:34, 11 March 2016 (UTC)

  Broadly, your argument still isn't relevant. All templates are unfamiliar at some point. When we find something unfamiliar we either give up, decide to use some other method, or look up how to use the one in question. I know you've said exactly that in other discussions, so I am puzzled why you think this one is special in that regard. (Aside: I have issues with this template regarding the wikitext that must be in the page to support the template, because it is a different method of "referencing" in that the ref tags are wrapped inside this template, but I also understand that the way this template works is the way it needs to work pending some MediaWiki work.)

  Problems with editors are problems with editors. You need to resolve that using dispute resolution, which I know you know about. Again I am puzzled why you think this template needs to be singled out.

  Which template started the discussion at WT:CITE is also irrelevant to this discussion (or indeed, whether any template started that discussion at all). --Izno (talk) 12:50, 11 March 2016 (UTC)

  I don't know why you keep saying that this template is "being singled out". Template:Rp/doc has an entire section titled ===Warning===, which goes far beyond a simple statement of relative under-use, and {{rp}} doesn't produce as many disputes as {{sfn}} (Rp is used on ~40% as many articles). If you are correct that this one is "being singled out", then why are there similar templates with even stronger statements, despite fewer problems?
  I have already told you why it deserves a very mild comment: There is a slow but steady stream of disputes about this, which means that this template's use causes a disproportionate burden on dispute resolution systems. It should be obvious that preventing disputes, e.g., by (very) gently reminding proponents to be a bit more communicative than usual, is better than resolving them after the fact. Preventing disputes is one of the main purposes of most documentation, from major content policies down to minor template doc pages. WhatamIdoing (talk) 22:55, 11 March 2016 (UTC)

  It might well be a good idea to remind people of WP:CITEVAR in all citation template documentation (especially given the strengthening of the text in November 2015), but it would be quite wrong to single out this template. The case of {{Rp}} is a little different, since it can only be used in addition to other referencing methods, and its documentation says it "may end up being a temporary solution to these problems". Peter coxhead (talk) 09:00, 12 March 2016 (UTC)

• I've added the {{high-use}} template to the documentation, since it is used widely-enough to merit that template. --Izno (talk) 12:34, 9 March 2016 (UTC)

Multiple uses of sfn with name=

Unless I have missed something, sfn differs from efn in not allowing name=, as in {{efn|name=fn1|Very long footnote 1}}. This feature would be particularly useful where the use of loc=[] creates a long sfn and clutters the wikitext. In that case, the long sfn would be defined in reflist=. --Michael Goodyear (talk) 15:04, 21 March 2016 (UTC)

That's right. Why are you having excessively long locations though? loc is only needed to find the citation within the op. cit., not to define it. Remeber that the "S" of sfn means "shortened"! Martin of Sheffield (talk) 15:26, 21 March 2016 (UTC)

• Maybe you want ref=, but sfn is so very, very complicated I can't remember if it does what you want. EEng 18:24, 21 March 2016 (UTC)

@Martin of Sheffield: one reason for uses of {{sfn}} inline to become very long is when URLs are provided to link to the page. For example {{sfn|Gray|1821|loc=[http://www.biodiversitylibrary.org/item/95185#page/10/mode/1up p.vi]}}.

Another problem has been created by reduced flexibility in citing parts of a book; once you could leave out the book title and have citations only to the chapters, with something like "In {{Harvtxt|Author|Year}}" after the citation template to link to the book in the Bibliography. Now this throws a "no title parameter" error. So editors have begun to expand their use of {{sfn}} to include the chapter: have a look at the wikitext of Amaryllidaceae, for example. Peter coxhead (talk) 08:50, 5 April 2016 (UTC)

|ps=none

Invalid with {{sfnm}}.

Markup	Renders as
{{sfn|Author|Year|ps=none}}{{sfnm|1a1=Author1|1y=Year1|1ps=none}}

72.43.99.130 (talk) 17:13, 4 April 2016 (UTC)

It's not documented as working in the way you seem to expect. Look at the {{sfnm}} page. It might be better to raise the issue there. Martin of Sheffield (talk) 20:59, 4 April 2016 (UTC)

Nowhere in the doc is it shown that {{sfnm}} (a) is based on a different code base (b) the postscript parameter should not be expected to work in {{sfnm}} as it works in {{sfn}} or {{sfnp}}. The doc at {{sfnm}} is even poorer. The doc of {{sfnm/core}} is non-existent. Thank you for your suggestion, but I don't have time right now to chase another talk page. 72.43.99.146 (talk) 23:49, 4 April 2016 (UTC)

I've fixed the documentation of {{sfnm}} to match its behaviour. It needs updating to use Module:Footnotes as the other "sfn templates" do. Peter coxhead (talk) 09:06, 5 April 2016 (UTC)

Help with errors on Amor Prohibido (song) article

I am not sure why there are errors on this article seeing as the surnames and years match the books used in this article. Can anyone help me understand why is it reading as "doesn't point to any citation"? Thanks, – jona ✉ 17:02, 20 April 2016 (UTC)

Missing |ref=harv – fixed now. Kanguole 17:28, 20 April 2016 (UTC)

@AJona1992: to clarify: if you use the CS1 templates, i.e. {{cite ...}}, then you have to add |ref=harv to make them generate the anchor to which the sfn template links. The CS2 template {{citation}} adds this automatically. Peter coxhead (talk) 18:12, 20 April 2016 (UTC)

Thanks for your help guys. Greatly appreciated, – jona ✉ 18:26, 20 April 2016 (UTC)

Pages interval

Is there a way to force the template to show a n-dash when using pp=number-number, even if one puts a hyphen in the template?--Carnby (talk) 21:25, 8 June 2016 (UTC)

I guess so, since {{cite book}} manages to do it. But it would involve a change to Module:Footnotes, and Lua edits are beyond my capabilities. --Redrose64 (talk) 22:21, 8 June 2016 (UTC)

Remove null edit

There's currently a <!--null edit --> string right at the end of the template code and it's outside of the <noinclude> tags. Could someone remove it if it doesn't serve any purpose? Uanfala (talk) 18:51, 21 August 2016 (UTC)

Done. Jo-Jo Eumerus (talk, contributions) 19:01, 21 August 2016 (UTC)

Fault when date is "n.d."

When the correct abbreviation for a source with no date is used, "n.d.", two periods appear in the short footnote. Example

Smith says "blah blah blah". Smith's book is about trees.

Notes

References

Works cited

Smith, John (n.d.). Musings. Buels Gore: Jabberwocky. {{cite book}}: Invalid |ref=harv (help)

Jc3s5h (talk) 15:44, 23 September 2016 (UTC)

I'm seeing n dot d dot comma ... which is what I think I would expect. --Mirokado (talk) 15:54, 23 September 2016 (UTC)

I missed that. I changed the example to show it's OK if a page number is given, but there are two dots if the page number is not given. Jc3s5h (talk) 16:15, 23 September 2016 (UTC)

Well, it's because "n.d." is treated the same as "2016", so the last "." is added just as it would be for a numerical date. @Trappist the monk: Module:Footnotes needs a small fix. Peter coxhead (talk) 16:25, 23 September 2016 (UTC)

Not sure that a 'fix' is necessary:

{{sfn|Smith|n.d.|ps=none}}

—Trappist the monk (talk) 16:27, 23 September 2016 (UTC)

No, but in the interest of user-friendliness, it should be done nonetheless. Headbomb {talk / contribs / physics / books} 16:39, 23 September 2016 (UTC)

I agree with Headbomb – yes, there's an obvious work-around, but it's a work-around and like all such should be replaced by a fix for the convenience of editors. It's hardly a major piece of Lua programming compared to all the other great work you've done on the citation templates, Trappist. Peter coxhead (talk) 16:51, 23 September 2016 (UTC)

Strong push-back against use of this template.

I've been trying to add this into articles to improve the referencing and tidy-up the overall appearance, but I am finding strong objections and reversion of edits. Is this template deprecated? KirksKeyKard (talk) 18:08, 9 December 2016 (UTC)

There are those who like it and there are those who don't. The template is not deprecated.

—Trappist the monk (talk) 18:19, 9 December 2016 (UTC)

• In Pennsylvania-class battleship you made changes so that the cites were now linked to the books to which they referred, and allowed those refs to generate controlled metadata, rather than being unparseable plain text. I would certainly support such changes.

But WP:OWN is an uphill struggle 8-( Andy Dingley (talk) 19:01, 9 December 2016 (UTC)

I'm right behind the use of this template. Along with the {{citation}} template it makes referencing much easier and neater. Martin of Sheffield (talk) 19:06, 9 December 2016 (UTC)

• I like this template (although I usually prefer the sfnp variant) but changing the formatting of references in an article (when there is already an established consistent format) should involve getting the consensus of other editors on that article's talk page; see WP:CITEVAR. —David Eppstein (talk) 19:22, 9 December 2016 (UTC)

Agreed on {{sfnp}}, where there's a useful publication year to be had.

As to "consensus" though, we're supposed to improve articles and all too often (i.e. every time) this change is complained of, it's simple OWNership of an existing, poorer state for an article, not a desire to have the best article. Often, which is a shameful excuse, there's then a complaint that the previous editor simply doesn't understand this "complicated" new template, and is happy to keep it that way. Andy Dingley (talk) 20:15, 9 December 2016 (UTC)

• Is it possible that people are pushing-back because of how you are going about this? Perhaps you would provide some examples? ~ J. Johnson (JJ) (talk) 00:08, 10 December 2016 (UTC)

group=

It is not clear from the documentation whether the parameter |group= is supported. By contrast {{efn}} documentation specifically supports this function. --Michael Goodyear (talk) 04:38, 10 December 2016 (UTC)

There is no |group= parameter. It is difficult to name all of the possible things that a template does not do so limiting the template documentation to only those things that a template does do makes some sense. I presume that there is no |group= parameter because you can use <ref name="Name" group="Group">{{harvnb|...}}</ref> when grouping is necessary.

—Trappist the monk (talk) 10:52, 10 December 2016 (UTC)

Your logic is irrefutable! And yes, that would do as a workaround, thanks. But a built in parameter would be preferable, I will modify the documentation and put in a request --Michael Goodyear (talk) 04:19, 11 December 2016 (UTC)

Enhancement request: new paramer: name

reference info for Philippine–American War

unnamed refs	71
named refs	102
self closed	82
Refn templates	2
cs1 refs	113
cs1 templates	146
cs2 refs	8
cs2 templates	19
harv refs	48
harv templates	49
sfn templates	81
rp templates	1
uses ldr	yes
refbegin templates	2
cleanup templates	3
dead link templates	4
webarchive templates	10
use xxx dates	mdy
cs1|2 df mdy	3
cs1|2 dmy dates	2
cs1|2 mdy dates	37
cs1|2 ymd dates	7
cs1|2 dmy access dates	4
cs1|2 mdy access dates	34
cs1|2 ymd access dates	14
cs1|2 dmy archive dates	1
cs1|2 mdy archive dates	16
cs1|2 last/first	99
cs1|2 author	12
List of cs1 templates Cite AV media (1) cite book (65) Cite book (8) Cite journal (4) cite journal (6) cite magazine (2) Cite news (1) cite news (12) Cite web (15) cite web (32)
List of cs2 templates citation (1) Citation (18)
List of sfn templates sfn (81)
List of harv templates Harvnb (41) harvnb (8)
explanations

How about an optional new parameter name for specifying an arbitrary name for the <Ref name="name goes here"> tag>? Implementing that looks like it needs a knowledge of WP:Lua and/or mw:Extension:Scribunto, neither of which are in my ready skillset. This would be useful in wikitext as e.g. {{sfn|Deady|2005|p=55|name=Deady2005p55}} instead of <ref name=Deady2005p55>{{harvnb|Deady|2005|p=55}}</ref>. Wtmitchell (talk) (earlier Boracay Bill) 03:08, 12 December 2016 (UTC)

Why? {{sfn|Deady|2005|p=55}} uses the name= attribute so that matching instances of the template will not be duplicated. Here is the output for Deady 2005 p. 55:

<ref name="FOOTNOTEDeady200555">[[#CITEREFDeady2005|Deady 2005]], p. 55.</ref>

—Trappist the monk (talk) 13:29, 12 December 2016 (UTC)

I apparently did not explain this well. I just made this edit to add pagenos to a bunch of previous cites of (Deady 2005). Those cites did not use {{sfn}}, but that article contains a bunch of cites like {{sfn|Constantino|1975}} which likewise need pagenos. Where {{sfn}} is used, I would like to preserve that style. My initial explanation above along with a look at the changes I made re (Deady 2005) in that recent edit ought to illustrate what I envision. I often run into similar needs to add pageno info to existing cites, resulting in a need for one shortened footnote pointing to an entire work to be replaced by several shortened footnotes with unique ref names to allow collecting duplicates, all pointing to the full cite of the work. Wtmitchell (talk) (earlier Boracay Bill) 21:41, 12 December 2016 (UTC)

Wherever you used <ref name=Deady2005p58 /> you could have used {{sfn|Deady|2005|p=58}} (the extra pdf page numbers to my mind are unnecessary and should be left out; but that's neither here nor there). So, no I guess I don't understand what it is that you are trying to achieve.

References

• Deady, Timothy K. (2005). "Lessons from a Successful Counterinsurgency: The Philippines, 1899–1902" (PDF). Parameters. 35 (1). Carlisle, Pennsylvania: United States Army War College: 53–68. {{cite journal}}: Invalid |ref=harv (help)

In the above comment are three instances of the {{sfn|Deady|2005|p=58}} template, all lumped together as they should be in the reflist. For the Constantino references, just add |p= or |pp= as appropriate, and Bob's yer Uncle, right? But, I don't think that you should be mixing styles. Pick a style and use that style throughout.

—Trappist the monk (talk) 22:38, 12 December 2016 (UTC)

I was just about to expand on Trappist's first post, but he beat me to it. I suspect the most important thing is to relax and let the computer do the work. {{sfn}} does the hard slog for you. Have a look at Oaks explosion which I've just been working on and you can see {{sfn}} being left to its own devices. Martin of Sheffield (talk) 22:49, 12 December 2016 (UTC)

OK. I haven't looked at this template since some pre-Lua version. Apparently, it's gotten a lot smarter while I wasn't paying attention to it. Wtmitchell (talk) (earlier Boracay Bill) 09:01, 16 December 2016 (UTC)

Just for the record, the 'smarts' that you are attributing to the Lua version are not new to the Lua version; those 'smarts' have always been there. Both the old wiki markup version and the new Lua version use variants of {{#tag:ref|<content>|name="<name>"}} so that MediaWiki knows to group identical cites together.

—Trappist the monk (talk) 11:32, 16 December 2016 (UTC)

Protected edit request on 11 December 2016

Michael Goodyear (talk) 04:16, 11 December 2016 (UTC)

You didn't include any explanation of what you want changed. Also, I think any change would need to be handled by a change to the Lua code, which can't be done as a protected edit — it needs someone with actual access to that code. —David Eppstein (talk) 04:26, 11 December 2016 (UTC)

Something clearly went wrong there. It would be very helpful if this template would include group= and name=, similar to the companion template {{efn}}.

You raise a related point, protecting edits is one thing, but the code should at least be visible in the interests of transparency, so that one can understand how it functions. --Michael Goodyear (talk) 14:25, 11 December 2016 (UTC)

The code is visible: Module:Footnotes.

If there is consensus to make this change, I will do so.

—Trappist the monk (talk) 14:59, 11 December 2016 (UTC)

• Thanks --Michael Goodyear (talk) 15:34, 11 December 2016 (UTC)

I oppose the change, for now. I find the documentation for {{efn}} mind-numbing, and I am completely unable to imagine what these parameters would mean with {{sfn}}. I suggest Michael Goodyear create some documentation explaining how all these templates would work harmoniously together, and how it would solve some citation problems that are hard to deal with absent the new parameters, and then we can decide if we like it. (Oh, and don't forget the documentation for {{sfn}} also describes {{harv}}, {{harvnb}}, {{sfnm}}, {{harvcol}}, {{harvcolnb}}, and {{harvcoltxt}} so the new documentation would have to explain how those fit in too.) Jc3s5h (talk) 15:14, 11 December 2016 (UTC)

That is certainly a valid point, and a major project to integrate all these related templates. As with most requests this arose from a real life scenario. The need arose when faced with the requirement for a second notelist and reflist arising out of a bibliography. A second set was needed because the citations won't go up the page but only down from where they were called. As soon as I did so it was deleted (several times) under the aegis of "duplicate reflist without use of group=". Attempts to use group= were succesful for efn and notelist but failed for sfn and reflist which is when I realised, these features were present in one and absent in the other. Since both temlates are frequently used to gether to separate notes from citations, this seemed illogical.--Michael Goodyear (talk) 15:34, 11 December 2016 (UTC)

As I understand the proposition, it only changes {{sfn}} and {{sfnp}} and only because these templates are not (must not be) used within <ref>...</ref> tags. For the family of {{harv}} templates that you mentioned, when it is necessary to segregate them in different groups, the group= attribute must be added to the <ref>...</ref> tags that wrap the {{harv}} template. {{sfnm}} is not supported by Module:Footnotes so is not associated with this proposed change.

—Trappist the monk (talk) 16:22, 11 December 2016 (UTC)

Any change must not disturb existing uses of {{sfn}} and {{sfnp}}. A lot of time can be taken from content development (ie wasted) by fixing changes to templates in numberous articles. There is always a tendancy to incrementally change templates to cover a special case. These are meant to be simple, short and clean templates for normal referencing. @Michael Goodyear: - can you show where this problem occurred; expecting references to go up the page might be rather confusing for readers. I might have a look at it for you later in the week, but for the moment I'm trying to knock the Oaks explosion into shape for tomorrow's front page (it's the sesquicentenary). Martin of Sheffield (talk) 17:07, 11 December 2016 (UTC)

Yes, naturally that's the point of edit protection and why we have Sandbox and testcase (I have developed templates previously). The particular context was William T. Stearn, which had a complex bibliography, with notes attached to the citations (and embedded sfn references within those). That needed a separate place for these to be anchored, since the Bibliography is below Notes and References. I created a subsection for Bibliography notes and placed a notelist and reflist there. In the end for the sfn references I utilised the workaround suggested by User:Trappist the Monk in the previous section, of embedding the sfn within ref tags. The actual short footnotes created will go up the page!--Michael Goodyear (talk) 18:32, 11 December 2016 (UTC)

William T. Stearn citation system

Oh, my goodness! Willam T. Stearn is a marvel indeed. There are footnotes to the footnotes to the bibliography! I've never seen anything like it. In one case you can click five times before getting to the bottom. Start at the and click from there. EEng 01:10, 12 December 2016 (UTC)

I have three observations on that article's "Bibliography" section.

When {{notelist}} or {{reflist}} is enclosed in {{refbegin}}/{{refend}}, this causes an accessibility issue because the font size is reduced to 81% (90% of 90%) which is below acceptable limits (85% of the page font size). I've fixed that.

If you need references for the references, there's something seriously wrong - if you need to justify the use of a source, is it really a reliable source?

The section is being overworked - at least three different groups of books are listed here, which per MOS:APPENDIX could easily be separated into three distinct sections: "Works" or "Publications" for the works actually written by Stearn; "Sources" for the works linked from the "References" section; and "Further reading" for the books about Stearn that were not linked from the "References" section. --Redrose64 (talk) 10:27, 12 December 2016 (UTC)

That just makes my head hurt. Especially note [b]. What is that? A list of the contents? Whatever it is, I'm not seeing that page as a good use of the proposed template change; mostly, I guess, because the example isn't clear cut.

—Trappist the monk (talk) 13:45, 12 December 2016 (UTC)

I would just like to point out that if the citations in an article begin to become an intractable mess, one alternative is to just redo it with WP:Parenthetical referencing. References within references, or references within explanatory footnotes, work just fine in that system. I suggest that it might be better to rewrite a handful of problematic articles than to make our footnote citation system even more impenetrable than it already is. Jc3s5h (talk) 14:32, 12 December 2016 (UTC)

Yes! ~ J. Johnson (JJ) (talk) 23:05, 16 December 2016 (UTC)

To answer the original q: if you find that you need a construct like {{sfn|Smith|2016|p=123|group=Foo}} then use {{harvnb}} instead, as in <ref group=Foo>{{harvnb|Smith|2016|p=123}}</ref> --Redrose64 (talk) 23:45, 12 December 2016 (UTC)