Template talk:Sfn/Archive 1

Ref param
Contrary to what the doc says, does not honour the ref=/Ref= parameter. Could someone fix this please? -- 89.15.54.157 (talk) 18:01, 22 February 2010 (UTC) ps: Could emits of etc please also be changed to  ? That way any leading/trailing whitespace in the anonymous parameters would be ignored.
 * It's not terribly clear admittedly, but the ref parameter is not intended to be placed in, but is used with the associated cite template - whether that be , , , etc.
 * Not sure what you mean by your "ps" though. -- Red rose64 (talk) 18:09, 22 February 2010 (UTC)

As would be obvious to anyone familiar with sfn and the harv family, sfn's failure to support ref= forces an editor to use. Ugh.
 * Ref= in etc is not much use unless you have a matching ref= elsewhere. As, for example, in harvnb and all the other  templates, where Ref= works just fine and always did.


 * Never mind my ps. Previously, and  did not emit the same thing. Thus my ps. Evidently, someone fixed {sfn} already. (harvnb etc are still broken though)
 * -- 89.15.54.157 (talk) 19:45, 22 February 2010 (UTC)
 * Re your #1. You can't control the ref element created by Sfn; you adjust the ref parameter of the target citation when you use Sfn with something other than author names. So, for example,  would need a citation such as   or (using SfnRef)  . You can't control both ends, but you can adjust the target, and that's enough. — John Cardinal (talk) 20:01, 22 February 2010 (UTC)
 * What you can do is to use (or its alias ) as in this example:
 * However this approach is not necessary if the last1 (etc.) and year parameters are used in the citation template, together with harv:
 * -- Red rose64 (talk) 21:19, 22 February 2010 (UTC)
 * However this approach is not necessary if the last1 (etc.) and year parameters are used in the citation template, together with harv:
 * -- Red rose64 (talk) 21:19, 22 February 2010 (UTC)
 * -- Red rose64 (talk) 21:19, 22 February 2010 (UTC)
 * -- Red rose64 (talk) 21:19, 22 February 2010 (UTC)


 * The way you guys go on you'd think someone was asking you to solve the world's global warming problems. So, I just fixed the problem myself.

Foo said "Hello". Bar said "Nice weather here". Foo also announced that his chickens had the pox.


 * Nothing especially exotic. This same sort of thing would be true for any collection of snippets where the individual snippets have attribution but not their own titles. -- 89.15.54.157 (talk) 23:28, 22 February 2010 (UTC)
 * Re: "The way you guys go on you'd think someone was asking you to solve the world's global warming problems." See WP:DICK. — John Cardinal (talk) 00:06, 23 February 2010 (UTC)


 * Do you seriously hold your response (last or any other) to be constructive, receptive, and thoughtful? Or are you just having a bad day? -- 89.15.54.157 (talk) 00:58, 23 February 2010 (UTC)


 * My comment directly above probably won't help because you seem committed to wasting time with snide remarks rather than concentrating on communicating clearly. Lastly, my day is going quite well. I never let ill-humored remarks from strangers with a bad attitude get in the way of that. — John Cardinal (talk) 01:24, 23 February 2010 (UTC)


 * I'm not committed to anything but a simple answer to a simple question. To that simple question, I was told that it wasn't possible, but that I was instead to twist myself into contortions (and that too, to accomplish something that I didn't want to accomplish). Then, when I pointed out that you were going over the top, and when I made the fix myself (showing how simple it was in reality) I was told that I was a dick. Then, when I pointed out that that was not a particularly constructive thing to say, and that you must be having a bad day to say such a thing, I'm told that I'm committed to wasting time with snide remarks.
 * In summary: first I was told to piss off, and then told I was a dick to not piss off, and then told I'm snide and ill-humored because I spot a pot calling a kettle black. You're quite right that your calling me (or anyone else) a m:DICK "won't help"; as the page tells you, that's a pretty DICKy thing to do. But whatever. Since its unlikely that you will start reflecting on what you or I have said, and since I have what I want, I'm going to vanish now. I wish you a better day tomorrow than you had today. Ciao. -- 89.15.54.157 (talk) 02:39, 23 February 2010 (UTC)
 * The thing is, your original post suggested to me that you had misunderstood the documentation because it was ambiguous. Accordingly, I clarified the documentation. You didn't actually say why you wanted to use ref inside until after you had modified the template. The thing is, we tend not to modify a high-use template like this without good reason, and we hadn't actually seen the good reason. -- Red rose64 (talk) 09:44, 23 February 2010 (UTC)

<If you fellows are through and can shake hands, I'm afraid I have to point out that the ref parameter broke the template, at least in Safari. I've rolled back the changes. I tried to do it in the sandbox, but its a little beyond me. Feel free to try to finish it, but please use the sandbox. Thanks. CharlesGillingham (talk) 20:55, 24 February 2010 (UTC)

Rolllback and the ref/Ref parameter
Charles Gillingham rolled back the recent addition of  with an edit summary that said the change broke the template in Safari. Before  is re-introduced, I think we should discuss it. IMO, Sfn should be as lightweight as possible, and the  added a bunch of parameter references and a switch statement. I don't think the usage case is very common, and the example given above produces a bad result: the short references mention author names that aren't visible in the full citation. How is the reader supposed to know which full citation is paired with the short citations? Yes, the reader can click the link to have the full citation highlighted, but that only works in browsers that support the ":target" pseudo-class, and that means IE8 users don't see the highlight. (There may be a Javascript solution for that.) Anyway, I am not sure I'd vote for adding the  functionality. There is a workaround: the user can use harvnb with a REF element.

If we do add, I recommend that we not implement both   and. Pick capitalized or not, and given the other parameters are not capitalized,  seems preferable. — John Cardinal (talk) 01:00, 25 February 2010 (UTC)


 * I agree that the ref parameter is unnecessary. It actually defeats the whole purpose of the template, which is to provide standardized link names (one for the footnote and one for the citation).


 * A better idea for your example is use  and  . Of course, if you really need your own special anchor name, you can always just type  . You don't really need sfn to do this and John is right, there's no reason to complicate the code and introduce errors in some browsers.  CharlesGillingham (talk) 04:58, 25 February 2010 (UTC)

Adding a group parameter
Is it possible to add an optional "|group=" parameter so that notes can be separated from other citations, such as journal citations. –  VisionHolder  « talk »  18:58, 4 March 2010 (UTC)
 * Have a look at Reading Southern railway station to see used in conjunction with books and journals, particularly refs 25, 40, 54. -- Red rose64 (talk) 19:15, 4 March 2010 (UTC)
 * My problem is that I have a mix of references types, as discussed here. I'd like to be able to separate out the notes from the more complete references since I'm using LDR.  If you want, after reading my LDR question take a look at one of my sandboxes, and you'll see what I'm dealing with.  I just figured having a way to separate out my short footnotes (which I rarely use) might help me solve 3 of my 5 situations: journal citations (under References), regular books with page citations (under Notes), and regular books with chapter citations (under Notes). –   VisionHolder  « talk »  19:47, 4 March 2010 (UTC)
 * Why not use short references for everything? There's no need to separate sources by type, and you can use a short reference for any source. — John Cardinal (talk) 20:37, 4 March 2010 (UTC)
 * I'm starting to play with that possibility in my sandbox, but so far I would have to say that I would prefer a "|group=" option. –  VisionHolder  « talk »  22:01, 4 March 2010 (UTC)
 * Okay... one big problem with using Sfn for everything is that many of my journal articles have up to 10 authors. When that happens, the sfn ref looks like this:  .  It doesn't look like much, but since you can never abbreviate it (unless you cheat using Template:SfnRef) numerous journal citations get ugly and difficult to read.  I was hoping to eliminate this by using Template:R, but that can't happen if I'm using sfn for everything.  (Too bad you can put all your sfn calls in the LDR by wrapping named ref tags around them.)  Simply put, citing journals is best with LDR and the Template:R, while citing books is best (usually) with Sfn.  This is why I would prefer to split out the short footnotes from the journal and web references.  This is why I want a "|group=" parameter. –   VisionHolder  « talk »  23:03, 4 March 2010 (UTC)
 * If you have a long list of authors, use and then make the ref= value match. — John Cardinal (talk) 00:16, 5 March 2010 (UTC)

This is what I'm looking to do. I managed it by doing my own CITEREF calls. I'm not sure how usable it would be or if the layout is acceptable for a references section, but if people like it, I could easily create a Sfn2 template. It looks like it's would only require stripping out the ref tag so that you can nest in a LDR block. Anyway, tell me what you think. (When you're looking at it, just imagine a huge article with a ton of references to journal articles, web pages, and complex books. Also notice the relative simplicity of the code compared to the complexity of the sources cited.) –   VisionHolder  « talk »  01:07, 5 March 2010 (UTC)
 * I don't care for the presentation of the source information on that page. Using "Yoder et al 1996" as a short footnote, and similar solutions for the other sources, is much better. — John Cardinal (talk) 04:42, 5 March 2010 (UTC)
 * P.S. I have used short footnotes on articles with hundreds of citations and it works fine. Most of the sources had at most a few authors, but some had 9 or 10, and there's no problem. — John Cardinal (talk) 04:44, 5 March 2010 (UTC)
 * P.P.S. (Sorry about the multiple edits.) I am not sure why you think using SfnRef is cheating. Some sources do not have author names, or have more than should be listed in a short footnote, and it's not clear how to make the cite template know what value the sfn has used. it's a simple and straightforward solution, and not cheating at all. — John Cardinal (talk) 04:54, 5 March 2010 (UTC)
 * I didn't mean to imply bad behavior on anyone's part. I'm sorry.  I'll just retract that comment than trying to explain it.  It was poorly worded.  However, I do favor LDR and R (for consolidating and grouping), but Sfn doesn't work with it.  Maybe instead of adding a group parameter, there could be a "noref" option to omit the ref tag so that the citation and CITEREF could be incorporated into LDR.  But that's probably not even worth it, and honestly, I'm too tired at this point to think anything through.  I'll take a look at everything again tomorrow and see if I can find a way to make it all work.  I just have a serious problem with redundancy for large books of collected articles, and my solution is the only one that makes sense to me so far.  Again, I'm sorry for the trouble. –   VisionHolder  « talk »  05:19, 5 March 2010 (UTC)
 * It should be noted that no special construct for "et al" is necessary - it is triggered automatically when is given four authors. You can see this at work in Daniel Kinnear Clark, refs 11 & 14. Admittedly this uses, but the effect for  would be similar. -- Red rose64 (talk) 11:36, 5 March 2010 (UTC)
 * Good point. I was taught to use "et al" after 2 authors and so I do it manually. — John Cardinal (talk) 14:27, 5 March 2010 (UTC)
 * Visionholder, you didn't cause any trouble and there is no need to apologize. I didn't think you were implying bad behavior on anyone's part, nor did I think you were behaving badly. We're just discussing the issue, and you asked for input on the method used in your sandbox, so I gave it! (Be careful what you ask for, I guess! )
 * I understand that you favor LDR, no problem. I prefer Sfn, probably for at least one of the reasons that you like LDR (it reduces clutter in the prose), but I also like that Sfn automatically generates ref names so editors don't have to do it. Some editors complain that Sfn and LDR are too involved, and I think that mixing the two together in one article exacerbates that issue and will also be confusing for readers. I'd prefer to keep Sfn simple and avoid parameter clutter rather than add a group parameter/feature. If you are determined to add citation groups, why not approach it from the other direction? Put your non-Sfn refs into the non-default group. The REF element supports that, and I think that's true even for LDR. — John Cardinal (talk) 14:27, 5 March 2010 (UTC)

I agree with John. As I said in an earlier post, I think sfn is all about standardization, not flexibility. If I understand you right, you want to use both WP:CITE and Shortened footnotes in the same article. Why not let shortened footnotes be the one without a group? And give a group to the LDR?

There is a way to use sfn with LDR. What if you used the standardized footnote name created by sfn? Like so: Article text with a citation to an article.

Article text with a citation to a page in a book.

More thoughts on using sfn with list defined references
I've just had another thought, more radical this time. What if we changed the footnote name in sfn to be "Smith2009", rather than "FOOTNOTESmith2009". Then r and sfn would have exactly the same behavior: would create a footnote named "Smith", just as  does. The only difference would be that (1) r puts a page number up in the superscript, merges with it's neighbors and takes a group. (2) sfn puts the page number into a shortened footnote and links to a full citation. Without the page number, they do the same thing.

If we want to make this change, we should do it soon, before someone uses the "FOOTNOTE" anchor directly. (That would be weird anyway.) CharlesGillingham (talk) 05:12, 6 March 2010 (UTC)

Yet another thought: we could also create a template SfnFootnote to be used in list defined references, so you could do this: {{reflist|refs=


 * This allows the user to chose their own ref name (which I prefer) and create a simple CITEREF-linked short footnote very similar to Sfn, but perfect for use with LDR and R. Basically, the {{tl|RCR}} template is {{tl|Sfn}}, minus the ref tag that is wraps around it... and maybe some other stuff I haven't been able to decipher yet.  Anyway... I admit that I'm probably bringing this up on the wrong page.  I should probably be preaching to the LDR crowd, not the Sfn crowd.  But again, I do appreciate your feedback. –   VisionHolder  « talk »  07:28, 6 March 2010 (UTC)


 * I'll be blunt: I looked at the page, and I don't like it. You are using too many citation methods, including two different superscript styles ([nb N] and [N], two different inline styles (superscripted link and with and without parenthesized text) and three different sections at the end of the article. The sources for the page ought to be simpler, and they can be. Choose one style, and follow it. You're making an arbitrary distinction between books, journal articles, and other sources, and it doesn't help. — John Cardinal (talk) 10:55, 6 March 2010 (UTC)


 * Thanks for the example. It's still not clear to me why you need the special footnote name "2006Lemurs2", as opposed to the standard footnote name that is generated automatically by sfn.


 * I looked at the article, and I appreciate the hard work you are putting into it. It looks like a well researched article. Good work.


 * However, I agree with John that the citation style is too complicated. (1) I think it is a mistake to implement the shortened footnotes as list-defined references the way you are doing it. Now an editor has to add each new shortened footnote in two places, once in the text, and once down in the code. This seems very complicated to me. Also, many of the best editors are not computer professionals, and will find it difficult or impossible to find the shortened footnote inside what is, essentially, a dense block of computer code. (2) It's not necessary to separate the shortened footnotes from citations-in-footnotes. (Note that many thousands of articles in Wikipedia combine shortened footnotes and footnotes into the same section, for example Starship Troopers. This is the de-facto standard. )  CharlesGillingham (talk) 18:30, 6 March 2010 (UTC)
 * If your requirement is for a {{tlx|sfn}}-equivalent that does not have the  around it, try {{tlx|harvnb}}. -- Red rose64 (talk) 13:38, 7 March 2010 (UTC)


 * Thank you for pointing out the existence of {{tlx|harvnb}}! I should have looked around at the related templates more.  As for merging the notes with the general references, I may do that.  I think it looks neater this way, but it's not the most important thing to me.  (Anyway, I like regular numbered footnotes, and not the "nb X".)  I also agree that one of the drawbacks to LDR (and not just my specific approach) is that you have to edit twice, once to add the material and once to add the reference.  I'm not denying that Sfn has a strong appeal.  However, I'm just not sold on the use of short footnotes for every citation, including journal articles.  I like separating out books cited, as I have in Ruffed lemur, which I modeled after many other Wiki articles in the biology section.  Either way, I think I'm just going to run my approach by the people who regularly review my GACs and FACs and see how they feel, and make my tweaks based on that.  Had I noticed {{tlx|harvnb}} prior to all of this, this conversation might not have been so long.  Thank you very much for your time and for sharing your opinions.  Although we differ in opinion, I'm glad I got to hear your side.  You have made excellent points. –   VisionHolder  « talk »  14:37, 7 March 2010 (UTC)

{{outdent|:::::}} How about using Sfn for all the in-text cootnotes, but use the following sections at the end of the document: You can come up with your own scheme for how to divide the full citations; the key part is to simplify the method used to link the footnote in the prose to the full citation. — John Cardinal (talk) 15:39, 7 March 2010 (UTC)
 * 1) Notes (for the footnotes produced by Sfn)
 * 2) References (header for the following two sections)
 * 3) Books (for any books referenced)
 * 4) Other sources (for all other sources including journal articles)

Current status?
Ref param and Rolllback and the ref/Ref parameter said sfn and/or SfnRef and/or the ref=param way be incompatible with some browsers. Unforunately the ends of these discussions were ambiguous. Is it safe for non-technical editors to use these facilities? If not, how to warning an editor that facility X can't be used? --Philcha (talk) 23:48, 21 March 2010 (UTC)
 * What is there in the current documentation to suggest that ref is provided in the and  templates? -- Red rose64 (talk) 00:08, 22 March 2010 (UTC)
 * The current version of Template:Sfn |ref= is a param of all the templates and is documented at Template:Sfn. I included in my last comment because Ref param and Rolllback and the ref/Ref parameter may be part of a browser incompatibility.
 * Back to the most important question - is it safe for non-technical editors to use these facilities? If not, how to warning an editor that facility X can't be used? --Philcha (talk) 05:12, 22 March 2010 (UTC)
 * sfn and SfnRef with |ref= seem to work OK on IE 8. What about Sarafi? --Philcha (talk) 10:19, 22 March 2010 (UTC)

Sorry about the confusion. The discussion above is about a ref parameter for this template. The ref parameter is working fine in the cite family of templates. Part of the point of the discussion above is that we only need ref in the cite templates: there's no reason for non-standard anchor names on both ends.

The browser issue only existed for the few days that ref was added to this template. It probably had something to do with the particular way it was implemented. If you want to get to the bottom of whatever happened, see the old version before the roll back.

There is nothing wrong with this template as far as I know. CharlesGillingham (talk) 20:28, 23 March 2010 (UTC)


 * I've some research and:
 * IE (PC) is a pain:
 * The only reliable ways to test old versions are to have spare PCs or to install virtualisation software and then set up a virtual PC for each a Win with the required IE version. There are hacks to shoehorn old versions into the same PC, but they are unreliable.
 * Netmarketshare for 3 Apr 2010 shows IE 8 and 6 have larger share than 7 - it seems IE users are either quick or very slow to upgrade.
 * Firefox (PC) users seem to upgrade quickly - see Netmarketshare.
 * At about 6% Google Chrome (PC) appears the 3rd browser after IE and Firefox. Netmarketshare shows Google Chrome in order older first. However, Netmarketshare's stats seem to start in Mar 2010 (see the checkbox in the report I've referred to) except that Chrome 2 peaked in Jul-Aug 2009. I have a copy that says it's 2.149.27, and that shows Sfn and SfnRef correctly, although it apparently can't calculate the position of citations in the 2nd and later columns of
 * Safari (Win + Mac) has about 3%.
 * Opera (PC) has about 1%. The latest version is 10, and version 9.24 shows Sfn and SfnRef correctly, but ignore.
 * I know nothing about mobile browers. --Philcha (talk) 22:56, 3 April 2010 (UTC)


 * Can't handle cite doi --Philcha (talk) 06:39, 4 April 2010 (UTC)
 * Puzzled by above two posts - have left message at user's talk page. -- Red rose64 (talk) 13:48, 4 April 2010 (UTC)

. Seems like just confusiion. No known issue here. CharlesGillingham (talk) 06:33, 17 February 2011 (UTC)

A version of Sfn that does not create a ref tag round it
I tried it at Template:SfnNoRef, but it doesn't work, probably because I couldn't drop both. If anyone wants to have a go, I'd be a grateful. IIRC  VisionHolder   mentioned it, but I can't remember where. --Philcha (talk) 12:57, 3 April 2010 (UTC)
 * What you're looking for is Template:Harvnb. You can see an example of its usage at Lemur. –   VisionHolder  « talk »  13:32, 3 April 2010 (UTC)
 * See earlier on this talk page. -- Red rose64 (talk) 14:23, 3 April 2010 (UTC)


 * I dislike Template:Harvnb and its variants as all insist on writing the author and date into the main text - if I think it's justified, I can type it in.
 * But Lemur currently uses Template:R, which generates . Managing the name= of can be a pain, although LDR can help. However, the way it's used in Lemur seems to duplicate citations of chapters of books. Sfn's loc param makes it possible to have just 1 citation for the book but different "refs" for each chapter - and then, like, collates all instances of  , e.g. 1. a b c d. See examples at Brachiopod. --Philcha (talk) 07:38, 4 April 2010 (UTC)


 * I'm sorry I don't have time to thoroughly look at the example Brachiopod, but from what I can tell, the books (or journal articles) have only one set of authors. In this case, Sfn works beautifully.  In the case of Lemur, there are books with individual chapters written by different authors (and editors for the book in general), as well as a book with chapters composed of dozens of "contributions" written by a variety of authors.  In the latter case, you need a full citation for the contribution, but it also helps to link back to the book that it's found in (since there is no DOI or anything else to help the reader find the source).  This is an extreme case, and not something that most Wiki writers have to deal with.  If your sources resemble those of Brachiopod, use Sfn.  If your sources resemble the behemoths used in Lemur, either use my technique or come up with your own... or use Sfn anyway.  There is no hard and fast rule, just opinions and general guidelines at this point. –   VisionHolder  « talk »  12:39, 4 April 2010 (UTC)
 * and related templates don't "insist on writing the author and date into the main text" - they put the author, year, etc. at the place where the template was itself placed. If the template is enclosed by, then the author, year, etc. will appear in the  . The primary difference between  and  is that the latter generates its own   whilst the former doesn't; but  will also handle redundancy without needing to fiddle about with ref names. So, to the end user, the following two examples are equivalent:
 * which displays as:
 * A statement. Another statement. A third.
 * A statement. Another statement. A third.


 * which displays as:
 * A statement. Another statement. A third.
 * A statement. Another statement. A third.


 * So, I'm sorry, but I don't see what the problem is. -- Red rose64 (talk) 14:02, 4 April 2010 (UTC)

How does Sfn combine 2 or more refs into 1 ref
If the markup has 2 or more refs with the same specification, e.g. name(s) and date and page(s), Sfn combines them into the page as displayed. What is the concealed magic that does this? --Philcha (talk) 10:32, 17 May 2010 (UTC)
 * It's done by combining all the info - up to five unnamed parameters (up to four authors/editors and a year), plus the named parameters (p, pp, loc), all preceded by the word "FOOTNOTE", into a single string, which then becomes the value of a  construct. That is,
 * generates the named ref
 * It so happens that if there are two or more ref tags with identical, then they are merged and only the contents of the first are displayed. -- Red rose64 (talk) 11:24, 17 May 2010 (UTC)
 * Oh, and you can test this without using by doing something like this:
 * Note how I have used the same  - this is deliberate, by way of demonstration. This displays as:
 * Statement 1.
 * Statement 2.
 * Note how I have used the same  - this is deliberate, by way of demonstration. This displays as:
 * Statement 1.
 * Statement 2.
 * Statement 2.


 * that is, the content of the second has been ignored, and the two merged because they have the same name. -- Red rose64 (talk) 11:29, 17 May 2010 (UTC)
 * I known about merging refs with the with the same specification. What I to know the action(s) the template does that produce the effect - what under the cover? --Philcha (talk) 13:21, 17 May 2010 (UTC)

Here's a heavily-edited version of the template, which will recognise just one author and neither pp nor loc; it also expects p to be present; and I've boldfaced the relevant bits:

Now let's assume the first example above ; the parameters substitute as:

and then the  magic word constructs the   tag thus:

Does that help? -- Red rose64 (talk) 14:34, 17 May 2010 (UTC)


 * I don't know the template syntax well enough to do what I'm hoping to do. I'd like to upgrade the harvard templates to combine citations - as Lemur shows, not combining citations leads to a ridiculous number of citations in the articles (I know Lemur does not use harv). I have a test template User:Philcha/harvnb, to be use on User:Philcha/Sandbox/Reference methods/Harvard test 2, and at present it's not combining the citations. What do you think is going wrong. --Philcha (talk) 10:28, 17 June 2010 (UTC)


 * I'm don't understand the problem you are trying to solve. You mention the Lemur article, so let's use that as an example.  I see many instances of what I would call "combined citations" there.  For example, for what is now footnote 2 (Garbutt 2007), there are 9 references, indicated by the backlinks a-i after the footnote number.  If this is not what you mean by combining, then what?  — JPMcGrath (talk) 03:30, 18 June 2010 (UTC)

. Just a question, long since settled, I think. CharlesGillingham (talk) 06:45, 17 February 2011 (UTC)

Supporting "bundling"
What is the best way for this template to support "bundling"?

I.e., suppose we have two back-to-back sfn calls, such as Some editors prefer to combine these into a single footnote. Right now, they must do this (as the doc page recommends): Is it possible to implement something like this? Where the sfn code notices the blank argument and combines them appropriately? This looks difficult to me, but does anyone know an easy way to do this? CharlesGillingham (talk) 23:03, 4 November 2010 (UTC)
 * I would prefer to avoid this altogether. Two sources=two references.
 * Technically speaking: the page number bit certainly won't work, due to the way MediaWiki handles named parameters. Regardless of the way that the template is written, when you have named parameters, and the same named parameter is specified more than once (even when the values it passes differ), all except the last instance are ignored. Thus, the 4 overrides the 3. -- Red rose64 (talk) 22:15, 6 November 2010 (UTC)


 * Good point about the p. That pretty much kabashes this idea. CharlesGillingham (talk) 01:53, 7 November 2010 (UTC)



Supporting "et al" in harv and sfn
Just a thought:

It would be nice if there were a simple technique for handling links coming from: or

This would require something in Citation/core, e.g. ref which would create an anchor of the form. CharlesGillingham (talk) 18:33, 5 November 2010 (UTC)
 * Use as in
 * -- Red rose64 (talk) 22:18, 6 November 2010 (UTC)
 * -- Red rose64 (talk) 22:18, 6 November 2010 (UTC)


 * I'm just saying it would be convenient if instead of having to type all that you could just type ref . CharlesGillingham (talk) 02:25, 7 November 2010 (UTC)

Withdrawn. CharlesGillingham (talk) 06:51, 17 February 2011 (UTC)

sfn with List defined references
An (almost) fully automated system could be built that uses sfn and List defined references. It would require a template something like: Which would be functionally equivalent to: And the LDR support in cite.php would have to show footnotes that are not used, rather reporting an error.

See User:CharlesGillingham/Wikipedia/Wish list. CharlesGillingham (talk) 18:42, 5 November 2010 (UTC)

Withdrawn. CharlesGillingham (talk) 06:51, 17 February 2011 (UTC)

Broken example
The examples looked great when I finished them last week. Now attribute. -- Red rose64 (talk) 22:59, 18 May 2011 (UTC)
 * Thank a lot! simple solution and it works ! A ntv (talk) 23:25, 18 May 2011 (UTC)

Duplicate word 'then' in template documentation
There's a duplicate word in the documentation: "If the date parameter is not a full date, then then the extraction will fail." --Siipikarja ♫ 10:52, 8 September 2011 (UTC)


 * The doc page is not protected, so you can fix this yourself. ---— Gadget850 (Ed)  talk 11:46, 8 September 2011 (UTC)


 * . Thanks  Gadget850 (Ed) . --Siipikarja ♫ 13:48, 8 September 2011 (UTC)

Help
Would someone explain why the template which appears in the Coordinated Universal Time does not link to the following template which appears in the reference section of that article: "*"

(By the way, I tried it without quotation marks around the author too.) Jc3s5h (talk) 15:48, 18 September 2011 (UTC)


 * Because you have a comma after ITU Radiocommunication Assembly. ITU Radiocommunication Assembly is the publisher, so can be used to create the ref id. ---—  Gadget850 (Ed)  talk 16:31, 18 September 2011 (UTC)
 * The Science of Timekeeping and "Astronomical Time" are in the reference list but are not used. ---— Gadget850 (Ed)  talk 16:34, 18 September 2011 (UTC)
 * Thanks. By the way, the ITU Radiocommunication Assembly really is the author; the ITU is the publisher. Also, the two "unused" references are correct as is; they are general references. Jc3s5h (talk) 22:14, 18 September 2011 (UTC)

The |ref=harv field
I've added harv to the twelve instances of cite book and similar. Prior to 07:17, 26 September 2009, if there was no ref field in such templates, an anchor would be created automatically as a suitable target for a link from the note dropped into reflist. Since then, with this edit to Template:Citation/core, anchor creation no longer occurs in the absence of ref, and so harv must now be explicitly provided for a suitable anchor to be created. --Redrose64 (talk) 21:37, 27 September 2009 (UTC)

Enabling link from citation to external on-line text
Currently the template links to the source listed in the article with the appropriate ref= tag. In many articles, where there is an on-line version of the source available, it would also be desirable for the template to link to the relevant page of the online source i.e. equivalent to using to yield
 * myrefs citations


 * Sources

So in this case the page number in the short citation links to relevant page in the external document, the name and year in the short citation link to the source list. In the source listing, the url= parameter is made to link to the contents page of the book rather than a specific cited page (because there are many citations to different pages of the same book).

Any thoughts? Anyone like to have a go? Stephen Kirragetalk - contribs 14:32, 25 November 2010 (UTC)
 * You could use the following:


 * -- Red rose64 (talk) 21:10, 25 November 2010 (UTC)


 * Awesome Redrose! Can't believe I didn't try it myself - never occurred to me that it would work with a link in the page parameter. Thanks. Stephen Kirragetalk - contribs 00:20, 26 November 2010 (UTC)

Allowing quotes as an attribute
I think it would be very useful to have an attribute for quotes in this template, basically at the page level. Since at this point the quote is only in the Cite book level, for the entire book. Please take a look at List of ancient cities in Thrace and Dacia where An Inventory of Archaic and Classical Poleis and The Cambridge Ancient History, Volume 3, show up again and again, with full detail, since different pages are cited, with quotes by previous editors. I would love to convert it to Sfn similar to how I wrote Amutria from scratch, and how I already converted the references to Grumeza in the city list article. However, I can't carry over the quotes, and I don't want to lose them or annoy previous editors. As a side note, I am trying to convince a reviewer for the value of this template and citation style here. Maybe you can give hand, with some pertinent examples ;-) Thanks!--Codrin.B (talk) 18:22, 6 January 2011 (UTC)
 * You can do this using the existing loc parameter, something like this:


 * I'm suggesting that both page and quote go into loc, rather than tacking the quote onto the page in p, because the p parameter should be a single page number, nothing else. -- Red rose64 (talk) 19:40, 6 January 2011 (UTC)
 * Very cool! I didn't know that loc is multipurpose, as the examples only show chapters. Thanks!--Codrin.B (talk) 19:56, 6 January 2011 (UTC)

et al. and italics
Is there a reason why et al. is not italicized by this template? –  VisionHolder « talk » 00:16, 21 February 2011 (UTC)
 * Please see the reply to the similar q at Template talk:Cite journal. -- Red rose64 (talk) 16:31, 21 February 2011 (UTC)


 * Per MOS:FOREIGN: "Loanwords and borrowed phrases that have common usage in English—Gestapo, samurai, vice versa—do not require italics. A rule of thumb is not to italicize words that appear unitalicized in major English-language dictionaries." I would consider et al. to be common usage. ---— Gadget850 (Ed)  talk 18:17, 24 February 2011 (UTC)


 * Regardless, the templates are inconsistent. In the cite templates, et al. is italicized for author lists but not for editor lists.  Sfn does not italicize.  Either way, it should be consistent. –  VisionHolder  « talk » 19:35, 24 February 2011 (UTC)
 * I just fixed Citation/core to never italicize "et al.". (Previously, it italicized in some contexts and did not in others.) I'd also be fine with having that template and this one always italicize, but as Visionholder says, they should be consistent. Ucucha 00:10, 9 March 2011 (UTC)
 * So, I take it that the issue is now resolved? VisionHolder, do you agree? CharlesGillingham (talk) 03:03, 9 March 2011 (UTC)

new template : merge or fork? And related new templates...
I just created a new template. The purpose of this is to create shortened footnotes that look like


 * , p. 25.

Instead of



Essentially, this:

is equivalent to this:

Except that it automatically combines identical footnotes, just like.

The motivation was the Old Chinese page, which I've been editing, and which used the style with parens, and repeatedly used the longer idiom with &lt;ref>.

, who created, suggested on Template talk:sfnp that we should possibly just replace the existing definition of with my  definition. Do people think this is reasonable, or should we keep them separate?

Here's what else I was thinking:
 * Create that outputs in the same format as  does, e.g.

outputs
 * , p. 25


 * Create that behaves like  but outputs using using colon page number notation, just like  and friends.
 * If remains split, create  and, as paren-format versions of  and.
 * Whatever ends up created, make sure it's documented in . Benwing (talk) 04:02, 11 May 2011 (UTC)


 * Merge, but make it an option to, something like y, so that existing behaviour of sfn doesn't change. -- Red rose64 (talk) 16:45, 11 May 2011 (UTC)


 * Merge. As Benwing said, I think that ", p. 26" might be a better look for footnotes created by sfn.


 * I don't think we should create all these new templates just yet. Obscure templates tend to be poorly maintained, they complicate the documentation and create needless confusion for new editors. CharlesGillingham (talk) 17:00, 11 May 2011 (UTC)


 * OK, an alternative is to change to use parens, but add a y or y option to eliminate the parens.  I'm not averse to changing {tlx|sfn}}; I know that Wikipedia has this "let a thousand flowers bloom" attitude about different styles, but at a certain point I think it just makes things confusing. Benwing (talk) 01:27, 13 May 2011 (UTC)


 * I favour the parentheses because they match what does, so I'd use the proposed  in shortened footnotes that refer to multiple sources.  The format of  also seems common, so an  variant for that would be useful.  Kanguole 12:01, 12 May 2011 (UTC)


 * I like Redrose64's and Benwing's ideas: merge minor style variations into this template, but add options so that the style can be customized. Ucucha 07:09, 31 May 2011 (UTC)

A couple other ideas
A couple things that came to mind:

1. I hate the name "sfn"; it's very obscure. I suggest we rename it to "snote" for "shortened (foot)note", which is clearer and (for me at least) easier to remember.

2. Incorporating multiple citations in a single note: Old Chinese has a lot of cases where two or sometimes three citations are grouped together. Currently the only way to do this is to manually create the reference and insert using or whatever. I'm thinking something like this:

This would allow a maximum of (say) 3 citations, which need to be separated by a blank argument.

There are two ways to implement this: either include it as part of the template or create a new template (e.g. sfnx) that allows this behavior. The advantage of the former is that it is easier to use; the advantage of the latter is that it avoids adding additional code to, which might slow it down (not sure whether this will actually happen).

Do people think there is any need for this?


 * I already created a template for multi-reference Sfn: Sfnm. Ucucha 02:17, 13 May 2011 (UTC)


 * "sfn" has the advantage of being short, which avoids edit-mode clutter. Note that the full name is shortened footnote, which is redirected to sfn. CharlesGillingham (talk) 04:33, 31 May 2011 (UTC)

Error detection
The links that Sfn generates are often broken, because it is difficult to check them all. created and I updated a script that automatically catches these errors. See the documentation at User:Ucucha/HarvErrors.

To install the script, add:

to Special:MyPage/skin.js. Ucucha 09:23, 3 June 2011 (UTC)


 * Excellent script. ---— Gadget850 (Ed)  talk 21:24, 18 July 2011 (UTC)

How to check anchors?
How can I examine the name of an anchor created by Cite book and related templates when "ref = harv" is specified, so that I can compare it to the anchor that Sfn is trying to link to? Jc3s5h (talk) 14:01, 18 July 2011 (UTC)
 * Look at the HTML source for the page, where it says somewhere near your citation span id="CITEREF... Ucucha 14:16, 18 July 2011 (UTC)


 * User:Gadget850/Citation templates— anchors shows how the anchors are built by the various templates. Time to move that to WP space I think. ---— Gadget850 (Ed)  talk 14:18, 18 July 2011 (UTC)


 * Thanks, both hints were helpful. Jc3s5h (talk) 16:52, 18 July 2011 (UTC)
 * Also see User:Ucucha/HarvErrors where there is a suggestion on modifying your Special:MyPage/skin.js. This will show a red error message if there are incorrect links. -- Red rose64 (talk) 20:14, 18 July 2011 (UTC)

Display of author and year
Would it be possible to tweak the template so that it adds a comma after the author's surname, so instead of "Foo 2011, p99" it displays "Foo, 2011, p99". Mjroots (talk) 20:35, 3 September 2011 (UTC)
 * I believe that is supposed to produce output similar to , which also doesn't use a comma. -- Red rose64 (talk) 20:40, 3 September 2011 (UTC)

Requesting more flexibility to allow wider deployment.
Hi. I have been using this referencing system for some time now, and I am a fan. I would like to request that this system be made more flexible for wider deployability. There are two features which I would like to request. One is that the sfn template should have a “group=” parameter, where we can use codes of the type. A second feature I would like to request is that the template should allow a command which would allow it to link directly to the referenced book or journal or whatever. Something like, which would allow linking to the cited work directly. The uses of the “group=” parameter are obvious. The uses of the second one would be connected with the “group=” parameter. For example, if an ed wants to show some explanatory footnotes in a “Footnotes” section, as in Kunbi article, there appears to be no satisfactory “sfn” way to do it. It can be done with the “#tag:ref” code. But the trouble with using this method is that it leaves a lot of clutter in the article text (which could otherwise have been put into the “Footnotes” section). The reason for using the sfn system that the ed wants to have the article text code as free of referencing clutter as possible (so as to make editing easier). Not being able to do this completely leaves the ed unsatisfied and feeling defeated. Presently, I find that eds do not like to use quotes while using the sfn system. This is because, if quotes are introduced using the “loc=” parameter, the “Notes” section begins to look ugly (because of the presence of one or two quotes in it). Now, if an ed wants to use quotes or some other explanatory footnotes, and take the article to GA or FA level, the ed is in a difficult position because combining the sfn system with some other system becomes a minus point at the GAR or FAR review. Having standardized and a good looking system makes it easier to pass GA or FA. On the whole, if the sfn had these two features, the ed would not have to worry about being able/unable to do this or that, while choosing the sfn system. I believe, having these two feature in sfn would immediately result in wider deployment of the system -- give us more articles which are easier to edit -- and make other referencing systems mostly obsolete. As such, I think addition of these two features should be given a consideration. M W ℳ 15:56, 24 October 2011 (UTC)


 * Another approach that would allow one group of footnotes for discursive notes and a different group for citations is explained at Help:Footnotes. So the &lt;ref group=whatever> tag could be used for discursive notes and sfn used for citations.


 * I don't see how the proposed revisions would make any change to how quotes are handled.


 * Looking at the "Kunbi" article, I would have done it a little differently. I would have reserved one note section, perhaps titled "Notes", for commentary that did not cite any outside works (although I would have included wikilinks). A second note section, perhaps titled "Citations", would include all notes that contain a citation, even if it also contains commentary or quotes.


 * I personally do not see the need to have two separate sections. One section could have a mixture of commentary and citations. Jc3s5h (talk) 16:23, 24 October 2011 (UTC)


 * Kunbi uses the shortened footnotes system; sfn is a template that may be used to create the in-text cite.
 * The practice of separating explanatory and other notes is well-used, but I don't see how it should be done with sfn. You should look at r which does support groups.
 * Shortened footnotes does not use quotes in the Notes section, and I have never seen this as a practice before.
 * I don't understand the link request. Do you want to link to an article or webpage? Either should be linked in the full citation.
 * ---— Gadget850 (Ed)  talk 16:37, 24 October 2011 (UTC)
 * The idea for a group parameter has potential. The idea for a link parameter akin to the ref parameter provided by has been proposed and rejected several times before (see earlier threads on this page): one of the main reasons for  being constructed the way that it is is to keep it simple - there are up to five unnamed parameters (four authors and a year), plus one named parameter selected from the three that are recognised.
 * If the target that links to doesn't exist, it may be added to the citation template by means of the  template; so for your example of, the matching  would contain ref -- Red rose64 (talk) 16:45, 24 October 2011 (UTC)


 * I have already gone through most of this whole page. I think I was not clear enough on what I had meant by direct linking. Presently, the sfn template generates a numbered blue link of this[5] sort in article text. When the blue link is clicked, it links to a "Notes" section, (as in the Kunbi article) where we get the authorlastname+year+page no. This, when clicked, again links to the "References" section, where we see the full citation with full authorname, bookname, year, publisher, ISBN etc. I was proposing that there should be a command which would skip the "Notes" section. I mean, when the numbered bluelink [5] gets clicked, (if the new command has been used in the sfn template), it would make the numbered bluelink link directly to the full citation in the "References" section of the article. The benefit of being able to do this is that being able to do it would allow the numbered blue link to be coded to link directly to the explanatory footnote in the "Footnotes" section. It would obviate the need for using the #tag:ref code, as it is being used in the Kunbi article presently. This would result in a situation where we would have the explanatory footnotes in the article, but there would be no cluttering in the article wikitext (while using sfn throughout), and there would be no need to use harvnb or R or any other referencing template in conjunction with sfn, and the whole article would be written by using the sfn template only. I understand the desire to keep the sfn parameters simple. But I feel that it is resulting in a situation which is making sfn less than what it could be, and eds would be somewhat reluctant to use the sfn template due to lack of flexibility. Creating this flexibility would increase its usage. Having used this template in some other article, I at least have this impression. And I hope I have been able to explain what I had meant by direct linking. M W ℳ 17:37, 24 October 2011 (UTC)
 * When you have a footnote as big as this it's time to consider forgetting about a footnote, and linking to an article covering that topic instead. -- Red rose64 (talk) 17:46, 24 October 2011 (UTC)


 * I am OK with that. What I had in mind was that we could put explanatory footnotes, and also the quotes, in the footnotes section, and then link them to the "Notes" section. Maybe its not time yet. I am in no hurry. Thanks for your responses anyway. M W ℳ 18:03, 24 October 2011 (UTC)

It is trivial to add group to sfn. I'll look into it. We could also add a link parameter. I assume you mean a link outside Wikipedia, am I right? (discussion above notwithstanding). I'm thinking this is the right design:

Where sfn has http ://en.wikipedia.org/wiki/Main_Page:

With a p,pp or loc:

With no p,pp or loc: 

Sound good? CharlesGillingham (talk) 23:39, 25 October 2011 (UTC)
 * I think he wants the link to go from the in-text cite to the full citation, not the shortened footnote. Jusr like r. ---— Gadget850 (Ed)  talk 23:46, 25 October 2011 (UTC)
 * Ah yes, you're right. Sorry I was hasty.


 * sfn is for shortened footnotes. If the link went straight to the book, then it's not a shortened footnote. It's a "normal" footnote.


 * (The purpose of shortened footnotes is to make multiple references to the same source, where each citation is to a different page number. Kunbi has quite a few of these.)


 * Gadget suggests conversion to r. There are two issues to consider:


 * (1) You have to handle the page numbers somehow. r handles them like this: . (Following the superscripted number with a colon.) Some people like this, some don't.


 * (2) MangoWong notes that they would prefer to edit the article without a lot of edit-mode clutter. The only way to do this with r is to also use list-defined references. This is certainly possible, but, similarly, some people like this and some people don't. It's a big change, so proceed cautiously with the other editors on the page.


 * At any rate, I think the "group" parameter makes sense, and is relatively easy. CharlesGillingham (talk) 01:43, 26 October 2011 (UTC)


 * I agree that it is simple to implement, but I would still like to see an example where it would be useful. Is the intent to have multiple groups of shortened footnotes with one reference list? ---— Gadget850 (Ed)  talk 02:59, 26 October 2011 (UTC)


 * I don't follow this either. sfn is designed for footnotes that are citations, for which the anchor is generated from the author(s) and year.  I don't see how it could be used for explanatory footnotes.  The above suggestion of using r with list-defined references for these footnotes seems to be the only way to achieve the de-cluttering MW wants.  Kanguole 14:07, 26 October 2011 (UTC)
 * r can also be used with in-text defined references. ---— Gadget850 (Ed)  talk 14:38, 26 October 2011 (UTC)
 * Yes, though that seems to be what MW was trying to avoid. But my point was that sfn seems inappropriate for footnotes that aren't citations (so adding group to it wouldn't help).  Kanguole 15:31, 26 October 2011 (UTC)


 * The main strength of sfn (as I see it) is that it leaves the article neat and clean, and also the wikitext uncluttered in the edit mode. This increases the readability and editability of the article. The "r" template leaves page numbers in the article text and this may be even more undesirable if we are using the "r" template for bundling refs. The sfn template does not have this limitation. Using Sfnm, one can even bundle ten refs, and still have only one numbered blue link in the article text. But bundling five refs with r, and having five page numbers showing in article text is somewhat undesirable. Other templates too do not seem to be as neat as the sfn is. So, if we want to have an article which would have higher readability (i.e. having the minimal amount of numbered bluelinks in article space) and higher ease of editing factor (minimal amount of clutter in the editing mode) sfn would be the template of choice IMO. But if I want to change all refs in the (say) Hinduism or Jesus article (which have 200-500 refs), I would inevitably be confronted by some naysayers. "No, we have always used quotes and explanatory footnotes." I say "We can do quotes and explanatory footnotes through the #tag:ref code." Reply "No, it is even more complicated. Leave it as it is." End of discussion. But if the sfn had some way of making the numbered blue link jump directly to the full citation, I could show an easy way to doing quotes and footnotes, and still having neater looking and more editable article. Another objection would be that sfn does not allow grouping, and I would get asked what to do if we want to group up refs. So, here too, my argument falters and the sfn does not get applied. But if the sfn did have the direct-linking-to-full-citation option and the group= parameter, I might have and easier time achieving consensus, and making the Hinduism and Jesus article more readable, and more editable like the G. S. Ghurye article. (Its a very short article, but you get the idea.) Another use of the group= parameter could be in an article like Kurmi. There, we are using lots of dated sources and it is explained in an explanatory footnote that the dated refs are unreliable. But the average reader has no easy way of knowing which is a modern source, and which is a dated source. If want that article to have sfn template, and then want to separate out the dated refs as "Primary sources" in the "Notes" section, the group= parameter would come in handy. We would put the group=(say)lower-alpha, code on all dated refs and separate them out in the Notes section as "Primary sources". The reader would have an easier time knowing which is which. We could use the group= parameter in the Hinduism or Jesus articles to separate out the primary sources, or any other group of sources as necessary. But the difficulty in doing quotes and explanatory footnotes does not get solved by the addition of group= parameter only. The sfn template would also need a direct-linking-to-full-citation option to be able to do quotes and explanatory footnotes neatly. Basically, I have requested the sfn being made more flexible so as to allow wider deployability, and to increase the readability and ease of editing of complex articles. As long as sfn does not have these features, it may be difficult to achieve consensus to change to sfn. That is why I request that the limitations of sfn be removed. Regards. M W ℳ 19:26, 26 October 2011 (UTC)
 * Er... well, NBR 224 and 420 Classes got to GA-class, and that uses mostly plus one  quite happily. Admittedly the latter isn't visible in the wikicode, being generated by the, but it's there just the same. -- Red rose64 (talk) 19:39, 26 October 2011 (UTC)


 * @MangoWong: I'm sorry but "direct linking" would make no sense for this template. It's primary purpose is to create only shortened footnotes (where shortened footnotes are for citations to many different pages of the same source). Direct linking would defeat the purpose of the template. Where should Wikipedia show you the page number? The page number is what this template is all about.


 * If you want to (a) avoid edit-mode clutter, and (b), use "direct linking", the only option is   (which can't show the page number) or   which shows the page number in this:  (page number is 332, btw), and in either case you must also use WP:LDR (list defined references), which means you to move all the references into the reflist template and add  . The full citations would appear in the footnotes, rather than in a separate, alphabetized bullet list. (My own view is that this is not better or worse, just a lateral move to a less popular format, but that's just me. You make up your own mind. But be careful: lateral moves can sometimes generate terrific opposition from other editors.)


 * As for groups, you can always create a group for the explanatory footnotes (using  inline. I think explanatory footnotes make sense inline; they're source text isn't horrifically ugly or unreadable (as citations often are). To generate the shortened footnotes, you use reflist, to generate the explanatory notes, you use


 * @Gadget: A group would make sense for an article that had separate sections for explanatory notes and short citations (and you really really wanted both groups to have a name). That's a reason to add the parameter (maybe not a great one, but it's a reason). CharlesGillingham (talk) 22:57, 26 October 2011 (UTC)

The difficulty in using LDR for doing explanatory footnotes is that explanatory footnotes have some further refs, i.e. numbered blue links. It would mean having refs within tags. This may not work properly and may be too confusing. While doing quotes, the intention was to show the authors last name with the quote (to avoid copyvio problems), and then put in another numbered blue link at the end of the quote which would lead to the page no. in the notes section and then to the full citation. However, if the group= parameter can be added, that too would add to the power and flexibility of the sfn template and is heartily welcome. M W ℳ 03:16, 27 October 2011 (UTC)


 * Looking at the use of citation for explanatory footnotes in G. S. Ghurye, my feeling is that you're stretching both citation and sfn a long way past their intended use, and this isn't going to work out. It might be better to design what you want for explanatory footnotes from scratch rather than trying to bend these citation templates to fit.  Kanguole 11:38, 27 October 2011 (UTC)


 * I agree, more below. CharlesGillingham (talk) 20:14, 27 October 2011 (UTC)


 * As a stop gap measure, to avoid using the citation in a non standard way, I have replaced it with Wikicite, which is probably more suited for non standard usage. If there be some better way of doing the same thing, I would certainly like to use it. M W  ℳ 12:16, 10 November 2011 (UTC)

Sandbox
I added group to the sandbox. Here is an example:

The brontosaurus is thin at one end. Then it becomes much thicker in the middle.

The Norwegian Blue Parrot will not move if its feet are nailed to the perch. Its metabolic processes are a matter of interest only to historians.

Notes

Citations

References



---— Gadget850 (Ed)  talk 10:58, 27 October 2011 (UTC)


 * If this is ever put in production, it will be necessary to change the documentation. I don't see any reason why this function should be incorporated in harv, but at present, essentially the same documentation serves for sfn and harv. I find trying to keep the sfn and harv syntax straight in my mind as I read the documentation confusing already. With this new option, perhaps it will be time to split the documentation. Jc3s5h (talk) 11:15, 27 October 2011 (UTC)
 * I still don't see the need for this, but it illustrates the issues. You have to include the entire note content twice: once in the sfn to create the link and again in the citation to create the anchor. I would certainly like to see it applied to Kunbi. ---— Gadget850 (Ed)  talk 11:28, 27 October 2011 (UTC)


 * On the substance of the example, I don't think it achieves what the original poster wanted, because the text of the note is mixed in the running text of the article. To achieve what the original poster wants, it would be necessary to give the note a name, put the note inside the reflist with the same name, and refer to the name in the sfn template. Jc3s5h (talk) 11:31, 27 October 2011 (UTC)


 * The problem with list-defined notes that include references is you can only include one before you hit . ---— Gadget850 (Ed)  talk 12:36, 27 October 2011 (UTC)

I think it is a mistake to use sfn for explanatory notes. Even if we had a group parameter (which I think is harmless), the example above is not a good way to use it. I would use something like. See below. CharlesGillingham (talk) 20:25, 27 October 2011 (UTC)

Another example
Here are several options, which use existing techniques without stretching them to the breaking point:
 * which I think is okay for a quote, but dicey for much more.
 * This screws up the punctuation and misuses sfn. Don't use this. (I just wanted to see if it worked.)
 * This works, but #tag:ref is too ugly for me.
 * I like this.
 * I like this too.

Terence Gumby's head hurt, as did the heads of Waldo Gumby,, Iris Gumby, Patricia Gumby, and Oliver Bolliver Gumby.

to  to aid in distinguishing, otherwise there are two numeric lists.
 * n.b. when doing this, it's important to use because  by itself doesn't convert the numbers to letters -  does this by wrapping the  in a . -- Red rose64 (talk) 21:32, 27 October 2011 (UTC)
 * Did not know that. Thanks. CharlesGillingham (talk) 22:11, 27 October 2011 (UTC)
 * I updated the example with your suggestion. CharlesGillingham (talk) 22:20, 27 October 2011 (UTC)
 * This works, but is now mixing Shortened footnotes and Parenthetical referencing. ---— Gadget850 (Ed)  talk 22:40, 27 October 2011 (UTC)


 * I don't think that's an issue. Parenthetical references in footnotes look fine to me in an article that already has shortened footnotes. After all, shortened footnotes are basically a parenthetical reference in a footnote. This is matter of personal taste, I suppose. CharlesGillingham (talk) 03:21, 28 October 2011 (UTC)
 * It should be the same style then: sfn and harvnb with no parenthesis or sfnp and harvcoltxt with the year in parenthesis. ---— Gadget850 (Ed)  talk 03:41, 28 October 2011 (UTC)
 * And you can use refn in place of #tag:ref. You don't have to worry about the order of name and group. ---— Gadget850 (Ed)  talk 12:07, 28 October 2011 (UTC)

Idea
Hey, Gadget, what do you think about having a template specifically for explanatory notes, and nothing else. Call it fn, the "footnote template". It would (1) use #tag:ref, so you could have ordinary citations inside explanatory notes, (2) set the group to lower-alpha, (3) ignore name, because this is rarely needed for explanatory notes. I've mocked it up over at User:CharlesGillingham/Wikipedia/fn (along with a few other ideas that are not quite as good).

(My main motivation here is that I hate when Wikitext begins to look like computer code, and when editing begins to feel like computer programming. Thus, for example, I hate when people (who are accustomed to computer programming) use named refs that have unreadable semantics like "JRG12". One of the goals of this template was handle named refs automatically, so I wouldn't have to look at them. group is another one of these things that is more like computer programming than it is like writing. So hiding group from editors would be a plus to me. A "footnote" is something a writer uses. "Lower-alpha" is something a computer program uses. That's my rant, for what it's worth.)

Anyway, tell me if you like the idea. CharlesGillingham (talk) 20:36, 28 October 2011 (UTC)


 * I was considering the same. Fn is an old Footnote2 template that I killed a while back and it is still used in archives (see what links here), so lets not use that. The only difference between refn and this one is the group, so how about refna. ---— Gadget850 (Ed)  talk 22:14, 28 October 2011 (UTC)


 * I don't like the name "refna" (it's not a reference, so "ref" means nothing. The "a" refers to the lower-alpha typesetting, which should not be a writer's concern). How about "efn" for "explanatory footnote". This would line up with WP:CITE very well, which calls these "explanatory notes". CharlesGillingham (talk) 18:31, 30 October 2011 (UTC)
 * Then 'explanatory footnote' with alias of 'efn'. ---— Gadget850 (Ed)  talk 20:53, 30 October 2011 (UTC)


 * Where are we with this? ---— Gadget850 (Ed)  talk 21:51, 10 November 2011 (UTC)
 * Haven't done it. Only takes a minute ... CharlesGillingham (talk) 06:05, 11 November 2011 (UTC)


 * Now it's done. See efn and the example Logic theorist. CharlesGillingham (talk) 21:06, 30 December 2011 (UTC)


 * This rocks. I would like to see name= supported as it would allow long notes to be moved out of the prose. I get that this would then preclude an {sfn} et al. in there, at least until MediaWiki wakes up with a new parser. Rock on, Alarbus (talk) 04:40, 5 January 2012 (UTC)
 * I added name support to efn. ---— Gadget850 (Ed)  talk 10:14, 5 January 2012 (UTC)
 * I'd also like to request adding support for other group types, like upper-alpha. The default can be lower-alpha, but I'd like the option to use the other possibilities. Ed [talk] [majestic titan] 11:05, 5 January 2012 (UTC)
 * Use refn. ---— Gadget850 (Ed)  talk 11:38, 5 January 2012 (UTC)
 * We should be focusing on a single look for site consistency; lowercase looks better and is less SHOUT-y. Anyway, what you're after is refn. Alarbus (talk) 11:24, 5 January 2012 (UTC)
 * I reverted the addition of . It broke the pages using this; all the notes were linked to 'a' and the others were not shown. A pity. Alarbus (talk) 11:24, 5 January 2012 (UTC)
 * Fixed. ---— Gadget850 (Ed)  talk 11:38, 5 January 2012 (UTC)
 * Thanks. Can we have blue text-shadows when you hover over the links? And a text-transform to the other case? Alarbus (talk) 11:43, 5 January 2012 (UTC)

Add this to your JS for tooltips on references with some nice box effects: importScript('User:Yair rand/ReferenceTooltips.js'); importStylesheet('User:Yair rand/ReferenceTooltips.css'); If you mean NAME, Name or the like, I would rather not. The defacto standard is lower case parameters and it really should not be an issue. ---— Gadget850 (Ed)  talk 12:06, 5 January 2012 (UTC)
 * I was kidding; proliferation of looks. People will eschew this over their pref for UPPER_ALPHA; Ed reverted me and went back to #tag over this. One dispairs. Thanks for the links; name= worked in John Diefenbaker. Alarbus (talk) 12:16, 5 January 2012 (UTC)

sfn and sections?
I seem to recall, perhaps incorrectly, that sfn supports the ability to place the links in references sections, like the REF tag's GROUP. Is this correct? If not, could we get it added? Maury Markowitz (talk) 21:23, 28 October 2011 (UTC)
 * ...as in the example above, which uses a tag to do the work instead of SFN.
 * And what do we call the sections? Normally I used "Notes" for the sfn's, and "Bibliography" for the cites. But the Notes aren't really Notes, and they're no more footnotes either. Is there some sort of standard naming? Maury Markowitz (talk) 21:27, 28 October 2011 (UTC)


 * Not sure what you mean. We discussed adding group to sfn, but it doesn't do what the OP thought it would.
 * Yes, adding group to sfn. Who's/What's the OP? Maury Markowitz (talk) 16:50, 29 October 2011 (UTC)
 * The discussion is Given that OP=Original Poster, that would be . -- Red rose64 (talk) 18:36, 29 October 2011 (UTC)
 * This is definitely what I'm looking for, but I don't seem to see any conclusions... the OP requested group=, is there any place where this is rejected? It's hard to follow! Maury Markowitz (talk) 01:34, 30 October 2011 (UTC)
 * See and the comments. ---—  Gadget850 (Ed)  talk 11:01, 30 October 2011 (UTC)
 * See a couple of ongoing discussions on section names at Wikipedia talk:Manual of Style/Layout. ---— Gadget850 (Ed)  talk 22:19, 28 October 2011 (UTC)


 * I think I'll use Notes, Citations and Bibliography then. Maury Markowitz (talk) 16:50, 29 October 2011 (UTC)
 * Until discussion is resolved, the possible section names are shown at MOS:APPENDIX. -- Red rose64 (talk) 18:39, 29 October 2011 (UTC)

loc=
Is it appropriate to use {sfn |Smith |1960} and the loc field for all print material that is part of the book Smith 1960? The full citation, also in the article, may specify that Jones is the author of the foreword, but it seems reasonable to specify Jones in the short note. By default, I suppose, the publisher is the author of "jacket", "cover", "front endpapers", etc. --P64 (talk) 17:34, 19 November 2011 (UTC)
 * Smith |1960 |loc=Acknowledgments
 * Smith |1960 |loc=Back inside jacket
 * Smith |1960 |loc=Foreword by Jones —"Foreword" by Jones? —Foreword (1985) by Jones?
 * The trouble with referencing the covers or endpapers is that they are not necessarily reliable, since they are typically written by the publisher's marketing department, based on "what looks good" or "what will sell the book", not "what facts are contained inside". Any genuine fact that does appear on the cover should also appear inside, and will thus have a proper page number. The foreword may also be written from the perspective of "outside observer". I don't immediately see what can usefully be referenced to the Acknowledgements section. -- Red rose64 (talk) 18:05, 19 November 2011 (UTC)

Sfn and SfnRef - red error message
I like Sfn and SfnRef. But (? a recent changes to the template) the template now shows red error message which hide the article's content. The SfnRef template also fails to understand the refbegin + refend combination. Please give this quickly. You see an example at User:Philcha/Sandbox/MOO 2 - 3. --Philcha (talk) 22:56, 11 December 2011 (UTC)
 * I think you're trying to use templates for other than their intended purposes. The main problem is that the article lacks a - either that, or a  is essential when  is used. The next problem is that the various  etc. within the References section should really be in the form of a bulleted list. -- Red rose64 (talk) 23:53, 11 December 2011 (UTC)


 * The red error message is not being generated by this template or sfnRef. It is being produced by cite.php, which is asking you to add either reflist or . Also, take a look at Template:Harvard citation documentation and Template:Harvard citation documentation (note that harvid == sfnRef).  CharlesGillingham (talk) 00:18, 12 December 2011 (UTC)


 * There are a bunch of problems in the markup:
 * Two citation templates improperly closed- fixed.
 * No reflist- fixed.
 * Misuse of sfn- this is for use in-text, not in the reference list
 * Misuse of sfnref- this should be used in the ref field
 * No * to define the reference list entries
 * Extra closing double braces
 * Use of cite web without title
 * Placing cite web inside sfnref causing the strip markers to be exposed
 * Use of partial dates in date, which mangles the anchor

---— Gadget850 (Ed)  talk 02:40, 12 December 2011 (UTC)


 * User:Philcha/Sandbox/MOO 2 - 5 shows "Cite error: There are tags on this page, but the references will not show without a Reflist template or a and the names are long and hard to spell, they won't like sfn and may prefer ref name= using a short name. So I use sfnRef and one or two names and the year. I'd be interested in other views on this.''

Could I get some input on the trade off between using sfnRef vs using up to four names and a year inline in uses of sfn? I'm a bit wary of using long templates inline as it seems at odds with the idea of being short. I like sfn, but don't want it to be seen as more trouble than it is worth. By selectively using sfnRef, I've been able to keep the inline invocations reasonable. FYI, the most recent page I've been doing this on is Tethys (moon). Tycho Magnetic Anomaly-1 (talk) 22:49, 17 December 2011 (UTC)

Edit request to add support for section= and inset= (for cite map)
This template currently supports books and journals very well, but not maps. Often, a map book will have several grid squares or different insets referenced in an article, and it would be especially useful to use a template for consistency rather than typing them by hand. So, this is to request the addition of the |section= and |inset= parameters, which are locators in map references like a page number is. I believe this is an uncontroversial request, as it merely expands the functionality of this template to cite map.

-->..τ ¢  04:55, 8 January 2012 (UTC)


 * Is there a reason that loc would not suffice? CharlesGillingham (talk) 09:39, 8 January 2012 (UTC)
 * Lack of consistency. Would be brought up at FAC for sure. -  ʄɭoʏɗiaɲ  τ ¢  15:05, 8 January 2012 (UTC)
 * I doubt it. I'd probably prefer to keep the template simpler with only loc, but a few more parameters won't do much harm. Ucucha (talk) 15:10, 8 January 2012 (UTC)


 * Consistency with what? CharlesGillingham (talk) 15:30, 8 January 2012 (UTC)
 * Template:Cite map, I think; it has section and inset parameters. Ucucha (talk) 15:40, 8 January 2012 (UTC)
 * And other templates have similar fields; for example cite video has time. ---— Gadget850 (Ed)  talk 15:42, 8 January 2012 (UTC)
 * Anything that allows you to locate the position of something on a source should be usable with this template... It'd still be just as simple if you didn't use those parameters, but I think that's a poor reasoning to reject improvements. Consistency would be with the grammatical formatting of long references using {cite foo}. On map references, ", section X34–Z38" is used, as well as ", Toronto inset". I'd like to use this template because of its linking ability, but right now this would certainly be brought up at the FAC for Highway 401 (the article I intend to introduce this on). Alternatively a new template could be created for non-book sources, but I think in the end they'd just be merged for the overlapping functionality. -  ʄɭoʏɗiaɲ  τ ¢  17:23, 8 January 2012 (UTC)
 * I used on Reading Southern railway station. The only objection concerned my use of a normal space instead of a   in "p. 4". I don't think there's any need to complicate what is deliberately a relatively simple template.  -- Red rose64 (talk) 18:36, 8 January 2012 (UTC)
 * How is it complicating it? Is making a simple template just as simple, but with a few more lines, too much for some to fathom? -  ʄɭoʏɗiaɲ  τ ¢  19:18, 8 January 2012 (UTC)
 * It is unnecessary to do so. The point about the family - of which  is a member - is that they keep the inline citations simple, since they push all the complicated stuff (the other parameters used by, etc.) off to the bottom where it doesn't clutter up the wikitext of the actual article. The  family also have a pretty consistent set of parameters - if I find  being used in an article which otherwise uses , it's a simple matter to amend   to   and wrap it in a pair of </ref> tags - I don't need to worry about whether the parameters have different meanings. So, if extra parameters are to be added to , they must also be added to the others in the  family.
 * These templates as they stand can handle any way of describing a point within a work, by using loc for all of them. Two alternative parameters, p and pp, are provided for the most common situations (i.e. one or more page numbers with no further qualification) - the visual effect of using 12–14 is identical to using pp. 12–14. But map sections and insets are much less commonly used than book or journal pages, so I don't see why we need to create extra parameters for them. Without amending the templates, you can already put section X34–Z38 or Toronto inset - has anybody at FAC actually complained about being used in this manner? -- Red rose64 (talk) 22:07, 8 January 2012 (UTC)
 * No. It does lend itself to errors though, but if that's the way it's to be done then that's the way. -  ʄɭoʏɗiaɲ  <sup style="color:#3AAA3A;">τ <sub style="color:#3AAA3A;">¢  22:16, 8 January 2012 (UTC)

This issue is design. I guess it is a subtle issue (to most users) but it is important to me personally. I prefer a simple and clean design. A simpler design helps in several respects: (1) It shortens the learning curve because there is less documentation to read and less choices to consider. (2) It helps with maintenance, because there is less that can go wrong. (3) It helps prevent edit-warring, because the less choices there are, the less there is to disagree about.

(Typically what "goes wrong" is mild confusion. For example, suppose you find an article that uses loc for some map sections but uses section for others. Someone might feel they have to "fix" the ones that use loc because they are "wrong". Or consider an article that cites a legal document that contains "sections" (normally written as "&sect;") and they use the parameter section you've designed to work for maps. Is this wrong? I'm not sure. What if one citation uses  and other  ? Does one of these need to be "fixed"? Not really, of course, but I can imagine that someone might see this and wonder if one of these is a "mistake". They may wind up searching the archives until they find this discussion before they are certain there is nothing magical about section. Another editor might disagree about any of these choices and we have yet another edit-war over nothing.  Do you begin to see what I mean? More choices causes more complexity, new consistency problems and unintended consequences.)

The difficult question is exactly how much simplicity we want or how much complexity we should permit. This is a subtle, subjective and arguable question. We (as a group) have to make a judgement call on these issues. In this case, my individual vote is for simplicity: use loc for everything other than page numbers. CharlesGillingham (talk) 20:31, 9 January 2012 (UTC)

why are there errors in the template?
Big red e.g. "Harv error: link to #CITEREFSmithJonesJohnson2005 doesn't point to any citation. Harv error: link to #CITEREFJonesJohnsonSmith2004 doesn't point to any citation."

What's the deal? MathewTownsend (talk) 00:52, 19 March 2012 (UTC)


 * You're looking at template:sfn/doc, right? Those are examples of syntax and they don't actually point at things. You're only seeing them because you have the script installed. Alarbus (talk) 00:58, 19 March 2012 (UTC)


 * oh, thanks. I did just install a bunch of scripts! I was wondering why that was suddenly showing up and no one else (but me) seemed to see them. humm, maybe I should ditch that script. MathewTownsend (talk) 01:05, 19 March 2012 (UTC)
 * The script is very useful at spotting problems with article citations. You do a lot of reviewing and should keep the script to help with articles you're commenting on. If you want to see the article without the red, use a non-logged-in browser. I do think the red is a bit over the top and was thinking of taking a local copy to tone it down. Ucucha might be game for a tweak. Alarbus (talk) 01:34, 19 March 2012 (UTC)
 * yes, I should learn how to fix those citations. Agree its over the top. BIG RED. A little frightening. But I was wondering why I was the only one to see it. Now I know! Thanks. MathewTownsend (talk) 01:37, 19 March 2012 (UTC)
 * The script has doc: User:Ucucha/HarvErrors. If you encounter an article with one of these citation issues it should be fixed before passing any review. Note that there are two messages the script generates; the second, “Harv error: There is no link pointing to this citation”, may not be a problem (and it can be turned off; see doc). citation always uses ref=harv and not all articles use the sfn/harv system. In this case, the red is ignorable.
 * If you need help fixing any of these, let me know; I do this all the time. Alarbus (talk) 01:53, 19 March 2012 (UTC)

If you see this error, you should take one second and tag it with citation not found. This points to specific instructions for local editors, and categorizes the article in a maintenance category. (I wish the script could add this automatically, but I'm told that's impossible with current technology.) CharlesGillingham (talk) 18:35, 19 March 2012 (UTC)
 * (I fixed the two errors in harvard citation documentation. CharlesGillingham (talk) 19:01, 19 March 2012 (UTC)
 * ok, I'll use the citation not found. Thanks, MathewTownsend (talk) 19:08, 19 March 2012 (UTC)

Option to remove terminal full stop in short form
I've just discovered this template – really nice and much easier to use when one main textbook is cited repeatedly at different pages! However, I always use the Citation template if I possible can rather than the "Cite" family (I prefer the output and it's easier than deciding which one to use). By default this template does not place a full stop/period at the end of a citation. So when combined with the Citation template, the result is a mixture of styles (see Cactus). It would be good to have an option to suppress the full stop. Peter coxhead (talk) 18:35, 7 March 2012 (UTC)
 * I use sfn and citation but manually add a full stop at the end of the citation template. It looks good to me. Aa77zz (talk) 18:42, 7 March 2012 (UTC)
 * In, the preferred method for adding a full stop is to set . -- Red rose64 (talk) 18:50, 7 March 2012 (UTC)
 * The documentation states it ends with a period, but looking at the markup, there is no default terminator. ---—  Gadget850 (Ed)  talk 19:00, 7 March 2012 (UTC)
 * Er, it states "Omit or leave blank to remove the trailing full stop.", therefore the full stop is only present if explicitly specified by . -- Red rose64 (talk) 19:33, 7 March 2012 (UTC)
 * I usually put the full citations in a separate subsection, because a mixture of full and short citations is visually jarring, and because full citations are usually given in alphabetical order. Kanguole 19:15, 7 March 2012 (UTC)
 * As for example with NBR 224 and 420 Classes. -- Red rose64 (talk) 19:33, 7 March 2012 (UTC)

In response to the comments above, I don't like a full stop at the end of citations: they aren't sentences and as per the spirit of MOS:CAPTION I prefer to reserve full stops for sentences and abbreviations. So I don't use. unless this is already the article style. I don't agree that a mixture of full and short citations is visually jarring. Further, there's no point in a short citation which links to a single full citation; it just creates redundancy. Anyway, all this is irrelevant; WP:CITEVAR allows me to choose the citation style in an article I create or first add citations too. I asked here because I can't edit the template and didn't want to create yet another one. I hope that this won't be my only option. Peter coxhead (talk) 19:49, 7 March 2012 (UTC)

I have now created Sfn/ps which by default creates a shortened footnote without a terminal full stop/period, and added cross-references to the documentation. This allows more consistent use by the minority, who like me, prefer the Citation template. Peter coxhead (talk) 09:48, 11 March 2012 (UTC)


 * Urgh. CITEVAR is not license to endlessly fork the existing citation templates for the sake of utterly trivial aesthetic quibbles. The last thing anyone should be doing in 2012 is adding even more inconsistency to our citation system. Chris Cunningham (user:thumperward) (talk) 10:54, 11 March 2012 (UTC)


 * I'm not adding inconsistency; quite the reverse. It's inconsistent to mix Citation formats with Sfn formats in the same list. Thus a bot (forgotten the name just now) goes around adding "postscript=." to citations which don't have a terminal full stop in articles where most do. Greater consistency would have been achieved if by default Sfn did not add a full stop, only doing so with the use of . This would make  the same as , which it should be, but isn't. However, it's too late for this better solution now. Peter coxhead (talk) 11:27, 11 March 2012 (UTC)

As no-one has objected to the last compromise suggestion made above, and it's been around for 6 days now, can some admin please make the change? I will then be happy to alter the documentation and fix sfn/ps. To be absolutely clear, the change requested is that the line which is currently -->}}.}}
 * Why is it "too late for this better solution now"? Was that actually established before the fork was created? Chris Cunningham (user:thumperward) (talk) 13:32, 11 March 2012 (UTC)
 * If this is a requested feature then why not add another parameter to the original template instead of making a new template? And while we are all here - could I request more feedback on centralizing citation discussions? See Help_talk:Citation_Style_1. Thanks.  Blue Rasberry    (talk)   15:35, 11 March 2012 (UTC)
 * We have three discussions going on here: A request to add a feature (which is only four days old), a discussion on a the creation of a new template that added that feature and a discussion on adding a similar feature to other templates. Could we please stick to one topic at a time? Or; at least split these up. ---— Gadget850 (Ed)  talk 16:03, 11 March 2012 (UTC)
 * Actually I can only see one discussion. Originally I thought the solution was to add a parameter which caused the trailing full stop not to appear, e.g. something like true. Then I realized that this was inconsistent with the way that "postscripts" work in other templates, e.g. Citation. The ideal solution would be if the original sfn template produced no "postscript", then . would do what is needed in a "cite family" context, and e.g., some note would allow additional notes to be added to the short footnote as you can via . This would be ideal because   and   would display exactly the same short note. But changing sfn in this way would render most existing uses incorrect. So unless someone wants to go back and change all the articles which use the template, it's too late. Hence I created a new template to do what I think sfn should have done in the first place.
 * To reduce maintenance, the content of sfn can now be changed so that all it does is call sfn/ps with the addition of . Peter coxhead (talk) 00:01, 12 March 2012 (UTC)
 * There is no need for a template fork, nor even a subtemplate. It can be done with a single additional parameter in . Consider - by default, this has a terminal period, but by specifying postscript with no value, the terminal period is suppressed. The necessary change to  to give similar behaviour is trivial, and concerns just one line - at present there is this:
 * -->}}.}}}}
 * Yes check.svg Done -- Red rose64 (talk) 15:40, 19 March 2012 (UTC)


 * Thanks. I've updated the documentation, as best I can, but it's in a rather complex form, so it needs checking. The template I created temporarily, "sfn/ps" has been deleted. Peter coxhead (talk) 12:40, 20 March 2012 (UTC)

Core
I created Harvard citation/core and sandboxed each template to use core. This ensures consistency and eases changes across the series. See sfn/sandbox for a sample of the simplified markup. ---— Gadget850 (Ed)  talk 15:53, 21 March 2012 (UTC)


 * “I love it when a plan comes together”
 * See this related talk re sfn/sfnm: User talk:Ucucha
 * I've not looked at the code yet, but *yes* — this is goodness. Alarbus (talk) 16:01, 21 March 2012 (UTC)

what about this article?
DragonFly BSD - here there do seem to be named references - but it shows harv errors when the article doesn't use them. MathewTownsend (talk) 20:54, 19 March 2012 (UTC)
 * That's another set of false positives. I'll deploy a fixed version of the script soon that should get rid of most false positives. Ucucha (talk) 21:03, 19 March 2012 (UTC)
 * I've disabled the script meanwhile. MathewTownsend (talk) 21:13, 19 March 2012 (UTC)
 * What about this article: Gray wolf? It comes up with many harv errors. MathewTownsend (talk) 21:19, 20 March 2012 (UTC)
 * 24: There is no Mech1974 in the Bibliography
 * 201: There is no Coppinger2001, it is CoppingerCoppinger2001 in the Bibliography because of the two authors; see #258 which is correct
 * Bekoff, Marc (1977): nothing in Notes links to this
 * Further reading: These are all false positives; using harv here enables anchors that are unused.
 * ---— Gadget850 (Ed)  talk 21:39, 20 March 2012 (UTC)
 * I have fixed for you, since it was clear that the omission of the second Coppinger was an oversight. I can't fix the Mech one, because not only are there several candidates: there's a distinct possibility that 1974 is correct and the book simply hasn't been listed at the bottom. -- Red rose64 (talk) 21:57, 20 March 2012 (UTC)
 * ok, thanks! I'm trying to learn this. So you're saying the error messages are correct in this case? Just wanted to check before I go muddling in. MathewTownsend (talk) 22:10, 20 March 2012 (UTC)
 * The ones in the Notes section are/were correct: I fixed one, and the other should still be fixed. The one under Bibliography may indicate an inadvertent omission of an inline reference (i.e. a ), or it may indicate a source which was not in fact used (if this is the case it may be moved to the Further reading section). Those under Further reading may be ignored. -- Red rose64 (talk) 22:25, 20 March 2012 (UTC)
 * ok, thanks! (I'm thinking that this is a little over my head.) MathewTownsend (talk) 22:43, 20 March 2012 (UTC)

I discovered that if I add isbn, all the following references get the error messages. MathewTownsend (talk) 19:28, 21 March 2012 (UTC)
 * I presume you mean isbn. This should have nothing to do with the anchors. Example please? ---— Gadget850 (Ed)  talk 19:39, 21 March 2012 (UTC)
 * I can't find it because I didn't save the article, once I saw what was happening.
 * On another article, what about Nullah which doesn't use harv? MathewTownsend (talk) 18:22, 22 March 2012 (UTC)
 * It uses 1911, which does. Kanguole 18:28, 22 March 2012 (UTC)
 * To be exact, it uses which passes harv into  which in turn uses . -- Red rose64 (talk) 20:01, 22 March 2012 (UTC)
 * hummm, ok. MathewTownsend (talk) 20:12, 22 March 2012 (UTC)

ok, if you add isbn to Vocabulary development under the "References" section, errors come up for every entry. (I didn't save it because I don't want to screw up the article.) MathewTownsend (talk) 23:04, 22 March 2012 (UTC)
 * You keep typing isbn, but have you actually clicked on it? I fixed three entries in the bibliography that did not have harv. I fixed two that used coauthors: if you use that parameter, the anchor will not match the link. If you edit the Bibliography as a section, every entry show an error because the References section is not being rendered. If you have a live article where isbn causes a Harv error, please let me know. ---—  Gadget850 (Ed)  talk 10:20, 23 March 2012 (UTC)
 * So if an article uses, it's going to throw a Harv error? As in Cliff dwelling?  MathewTownsend (talk) 20:27, 25 March 2012 (UTC)
 * Yes, but the ordinary user won't see that message. There is no harm in a web page containing link anchors with no incoming links. -- Red rose64 (talk) 20:50, 25 March 2012 (UTC)

Title-Date with wiki markup in link
Hi. I'm fine with the tweaks for wikicite/doc. But in a, I recall nikkimaria calling for consistency between sfn'd footnotes and the full cite, so .27.27 went in. You wouldn't know a way to do this without having to encode the apostrophes in the sfnRef would you? (fyi, the example is off Woodes Rogers.) Alarbus (talk) 01:04, 25 March 2012 (UTC)
 * The above was posted on my talk page. It's follow-up to.
 * It doesn't link when normal double-apostrophe italicisation is used, this is true:


 * but when you put non-intuitive encoded markup into, like this:
 * but when you put non-intuitive encoded markup into, like this:


 * it does work, but at the expense of puzzling non-techie editors. I think that it would surely be better to amend either or  so that their behaviour is mutually consistent. That is, if it is felt that the italicisation of titles in Title-Date  is a Good Thing. What do others think? -- Red rose64 (talk) 13:29, 25 March 2012 (UTC) clarified  Red rose64 (talk) 08:34, 26 March 2012 (UTC)
 * it does work, but at the expense of puzzling non-techie editors. I think that it would surely be better to amend either or  so that their behaviour is mutually consistent. That is, if it is felt that the italicisation of titles in Title-Date  is a Good Thing. What do others think? -- Red rose64 (talk) 13:29, 25 March 2012 (UTC) clarified  Red rose64 (talk) 08:34, 26 March 2012 (UTC)


 * The simplest technical fix would be to add title to both templates. ---— Gadget850 (Ed)  talk 13:48, 25 March 2012 (UTC)


 * How about geting  fixed to encode apostrophes in {sfnRef}? title in both rather gets away from the idea of short footnotes. I do agree that the .27.27 is awkward. Alarbus (talk) 16:50, 25 March 2012 (UTC)


 * I think it would be better to fix this in anchorendcode, if that's possible. It's better if the user doesn't have to think about this issue. CharlesGillingham (talk) 20:58, 25 March 2012 (UTC)


 * Let me rephrase that. A few ideas:
 * Add ititle to core so it uses title-date instead of author-editor-date and includes the italic markup in the anchor.
 * Add ref to sfn like the other Harv templates to create a custom link.
 * Add ititle to sfn for an title that is formatted in italics but does not include the italics markup in the link.
 * Add ititle to sfnref
 * I suppose there is the potential for the title in quotes. ---— Gadget850 (Ed)  talk 21:47, 25 March 2012 (UTC)


 * Is this actually fixed, now? V1.19? Redrose64's examples seem to work, albeit with the script still highlighting one. Will have to have a test somewhere. See Template talk:Citation for examples of this in a fistful of articles. As I've used this, it is frequently the case that it is only a portion of the footnote that is in italics. I believe I've seen a few uses of quotes in {sfn}, too. Alarbus (talk) 23:44, 25 March 2012 (UTC)

It worked because the title is enclosed in .27.27:

---— Gadget850 (Ed)  talk 00:16, 26 March 2012 (UTC)
 * Thanks. I tried it in an article preview and it still needed the .27.27. I think this needs mentioning somewhere on mw: (but I don't really know where). Alarbus (talk) 01:54, 26 March 2012 (UTC)


 * Looking a bit deeper: sfn creates a link without using any encoding, but sfnref encodes the anchor using anchorencode:. Using works, but the italic formatting is lost and spaces are converted to underscores. We could add anchorencode: to sfn, but I would not be surprised if it breaks some existing uses. ---—  Gadget850 (Ed)  talk 10:40, 26 March 2012 (UTC)
 * Are you thinking that adding anchorencode to {sfn} would allow normal use of apostrophes in both? If so, that would be ideal. We certainly don't want to have to use .27.27 in both. FWIW, I've seen encoded commas, quotes, and explicit underscores as well as a fair number of forced ref=CITEREF… but not much else. Alarbus (talk) 21:20, 26 March 2012 (UTC)


 * Again, I think  is the right way to go. What is most important is that sfn and sfnRef do exactly the same thing. Note also that citation/core also uses  . This is what is most simple and this is what users will expect. I feel strongly that this is an issue users should not have to think about.


 * Gadget: I don't think breakage will be an issue; there might be one or two articles where someone has built an ugly workaround for this problem, using CITEREF or god knows what. We just have to find them and fix them.  CharlesGillingham (talk) 08:17, 27 March 2012 (UTC)

Inconsistency with Template:sfnp
Sfnp had become inconsistent with this template because recent edits had not been applied. I think I've fixed it now, but there needs to be a better solution, e.g. creating sfn/core which is then called by other templates in the "sfn family" as is done for taxobox templates. Peter coxhead (talk) 11:09, 31 March 2012 (UTC)


 * Already in sandbox; see above. ---—  Gadget850 (Ed)  talk 11:36, 31 March 2012 (UTC)


 * Ah, right, I'd seen the earlier discussion but missed the fact that you'd actually created it. Ok, so what's needed is for the templates you have in the sandboxes to be made live... Peter coxhead (talk) 11:39, 1 April 2012 (UTC)