# delimiters

**URL:** <https://discourse.citationstyles.org/t/delimiters/643>\
**Category:** CSL Development\
**Created:** [May 25, 2009, 9:36pm UTC](https://discourse.citationstyles.org/t/delimiters/643 "2009-05-25T21:36:18Z")\
**Posts on this page:** 12\
**Page:** 1

<div class="post-metadata">

**Author:** ![Bruce\_D\_Arcus1](https://avatars.discourse-cdn.com/v4/letter/b/4bbf92/32.png) [@Bruce\_D\_Arcus1](https://discourse.citationstyles.org/u/Bruce_D_Arcus1)\
**Post date:** [May 25, 2009, 9:36pm UTC](https://discourse.citationstyles.org/t/delimiters/643/1 "2009-05-25T21:36:18Z")

</div>

I’m blanking on this; do we support #1 in this user’s list?

[http://forums.zotero.org/discussion/7152/citation-style-for-biotropica-little-help-needed/](http://forums.zotero.org/discussion/7152/citation-style-for-biotropica-little-help-needed/)

Do we have a test for it?

If not, we need to add this to the TODO.

Bruce

---

<div class="post-metadata">

**Author:** ![Frank\_Bennett](https://yyz1.discourse-cdn.com/flex029/user_avatar/discourse.citationstyles.org/frank_bennett/32/198_2.png) [@Frank\_Bennett](https://discourse.citationstyles.org/u/Frank_Bennett)\
**Post date:** [May 25, 2009, 10:08pm UTC](https://discourse.citationstyles.org/t/delimiters/643/2 "2009-05-25T22:08:26Z")

</div>

> I’m blanking on this; do we support #1 in this user’s list?
> 
> [http://forums.zotero.org/discussion/7152/citation-style-for-biotropica-little-help-needed/](http://forums.zotero.org/discussion/7152/citation-style-for-biotropica-little-help-needed/)
> 
> Do we have a test for it?

Currently, we can do this:

(Condit et al. 2000, 1995, 1998; Connell 1971; Cortes & Hunziker 1997)

But not this:

(Condit et al. 2000, 1995, 1998; Connell 1971, Cortes & Hunziker 1997)

He’s needing to change the trailing delimiter on collapsed blocks.  
Currently we can customize the delimiters within collapsed sets, but  
use a single default delimiter between collapsed blocks. This would  
need another option to supply the special delimiter value. Not sure  
what to call it. Something like “after-collapse-delimiter”?

---

<div class="post-metadata">

**Author:** ![Frank\_Bennett](https://yyz1.discourse-cdn.com/flex029/user_avatar/discourse.citationstyles.org/frank_bennett/32/198_2.png) [@Frank\_Bennett](https://discourse.citationstyles.org/u/Frank_Bennett)\
**Post date:** [May 25, 2009, 11:08pm UTC](https://discourse.citationstyles.org/t/delimiters/643/3 "2009-05-25T23:08:59Z")

</div>

> > I’m blanking on this; do we support #1 in this user’s list?
> > 
> > [http://forums.zotero.org/discussion/7152/citation-style-for-biotropica-little-help-needed/](http://forums.zotero.org/discussion/7152/citation-style-for-biotropica-little-help-needed/)
> > 
> > Do we have a test for it?
> 
> Currently, we can do this:
> 
> (Condit et al. 2000, 1995, 1998; Connell 1971; Cortes & Hunziker 1997)
> 
> But not this:
> 
> (Condit et al. 2000, 1995, 1998; Connell 1971, Cortes & Hunziker 1997)
> 
> He’s needing to change the trailing delimiter on collapsed blocks.  
> Currently we can customize the delimiters within collapsed sets, but  
> use a single default delimiter between collapsed blocks. This would  
> need another option to supply the special delimiter value. Not sure  
> what to call it. Something like “after-collapse-delimiter”?

Have added a test for this, code to come. The test incidentally  
uncovered a couple of silly bugs elsewhere that were very good to  
discover.

---

<div class="post-metadata">

**Author:** ![Bruce\_D\_Arcus1](https://avatars.discourse-cdn.com/v4/letter/b/4bbf92/32.png) [@Bruce\_D\_Arcus1](https://discourse.citationstyles.org/u/Bruce_D_Arcus1)\
**Post date:** [May 26, 2009, 1:48am UTC](https://discourse.citationstyles.org/t/delimiters/643/4 "2009-05-26T01:48:02Z")

</div>

> > I’m blanking on this; do we support #1 in this user’s list?
> > 
> > [http://forums.zotero.org/discussion/7152/citation-style-for-biotropica-little-help-needed/](http://forums.zotero.org/discussion/7152/citation-style-for-biotropica-little-help-needed/)
> > 
> > Do we have a test for it?
> 
> Currently, we can do this:
> 
> (Condit et al. 2000, 1995, 1998; Connell 1971; Cortes & Hunziker 1997)
> 
> But not this:
> 
> (Condit et al. 2000, 1995, 1998; Connell 1971, Cortes & Hunziker 1997)

Oh, I missed that wrinkle. That is a dumb rule that may well be a “won’t fix.”

Thoughts?

> He’s needing to change the trailing delimiter on collapsed blocks.  
> Currently we can customize the delimiters within collapsed sets, but  
> use a single default delimiter between collapsed blocks. This would  
> need another option to supply the special delimiter value. Not sure  
> what to call it. Something like “after-collapse-delimiter”?

Bruce

---

<div class="post-metadata">

**Author:** ![Frank\_Bennett](https://yyz1.discourse-cdn.com/flex029/user_avatar/discourse.citationstyles.org/frank_bennett/32/198_2.png) [@Frank\_Bennett](https://discourse.citationstyles.org/u/Frank_Bennett)\
**Post date:** [May 26, 2009, 3:33am UTC](https://discourse.citationstyles.org/t/delimiters/643/5 "2009-05-26T03:33:42Z")

</div>

> > > I’m blanking on this; do we support #1 in this user’s list?
> > > 
> > > [http://forums.zotero.org/discussion/7152/citation-style-for-biotropica-little-help-needed/](http://forums.zotero.org/discussion/7152/citation-style-for-biotropica-little-help-needed/)
> > > 
> > > Do we have a test for it?
> > 
> > Currently, we can do this:
> > 
> > (Condit et al. 2000, 1995, 1998; Connell 1971; Cortes & Hunziker 1997)
> > 
> > But not this:
> > 
> > (Condit et al. 2000, 1995, 1998; Connell 1971, Cortes & Hunziker 1997)
> 
> Oh, I missed that wrinkle. That is a dumb rule that may well be a “won’t fix.”
> 
> Thoughts?

If you’re willing to add an option for it, it’s nearly implemented.  
Up to you. I’d kind of like to go forward with it as a way of putting  
something back – the test I wrote for it turned to bugs in the  
citeproc-js code, which was very helpful. 🙂

---

<div class="post-metadata">

**Author:** ![Frank\_Bennett](https://yyz1.discourse-cdn.com/flex029/user_avatar/discourse.citationstyles.org/frank_bennett/32/198_2.png) [@Frank\_Bennett](https://discourse.citationstyles.org/u/Frank_Bennett)\
**Post date:** [May 26, 2009, 12:42pm UTC](https://discourse.citationstyles.org/t/delimiters/643/6 "2009-05-26T12:42:31Z")

</div>

> > > > I’m blanking on this; do we support #1 in this user’s list?
> > > > 
> > > > [http://forums.zotero.org/discussion/7152/citation-style-for-biotropica-little-help-needed/](http://forums.zotero.org/discussion/7152/citation-style-for-biotropica-little-help-needed/)
> > > > 
> > > > Do we have a test for it?
> > > 
> > > Currently, we can do this:
> > > 
> > > (Condit et al. 2000, 1995, 1998; Connell 1971; Cortes & Hunziker 1997)
> > > 
> > > But not this:
> > > 
> > > (Condit et al. 2000, 1995, 1998; Connell 1971, Cortes & Hunziker 1997)
> > 
> > Oh, I missed that wrinkle. That is a dumb rule that may well be a “won’t fix.”
> > 
> > Thoughts?
> 
> If you’re willing to add an option for it, it’s nearly implemented.  
> Up to you. I’d kind of like to go forward with it as a way of putting  
> something back – the test I wrote for it turned to bugs in the  
> citeproc-js code, which was very helpful. 🙂

It’s been implemented.

[http://bitbucket.org/fbennett/citeproc-js/src/tip/std/humans/collapse\_TrailingDelimiter.txt](http://bitbucket.org/fbennett/citeproc-js/src/tip/std/humans/collapse_TrailingDelimiter.txt)

Since the requester did post a link to a proper style guide with this  
requirement, it probably makes sense to go ahead and support it.

---

<div class="post-metadata">

**Author:** ![Bruce\_D\_Arcus1](https://avatars.discourse-cdn.com/v4/letter/b/4bbf92/32.png) [@Bruce\_D\_Arcus1](https://discourse.citationstyles.org/u/Bruce_D_Arcus1)\
**Post date:** [May 26, 2009, 1:22pm UTC](https://discourse.citationstyles.org/t/delimiters/643/7 "2009-05-26T13:22:58Z")

</div>

…

> It’s been implemented.
> 
> [http://bitbucket.org/fbennett/citeproc-js/src/tip/std/humans/collapse\_TrailingDelimiter.txt](http://bitbucket.org/fbennett/citeproc-js/src/tip/std/humans/collapse_TrailingDelimiter.txt)
> 
> Since the requester did post a link to a proper style guide with this  
> requirement, it probably makes sense to go ahead and support it.

That’s not the criteria we’re using here. There are a lot of little  
quirks in real world styles that we could implement, but where doing  
so adds unwarranted complexity.

I’ll look at your proposal later, but my question in evaluating is:

Can we explain its function in clear English, such that it can be  
included in a GUI in a way that a user would have a clear  
understanding of what it means.

---

<div class="post-metadata">

**Author:** ![Bruce\_D\_Arcus1](https://avatars.discourse-cdn.com/v4/letter/b/4bbf92/32.png) [@Bruce\_D\_Arcus1](https://discourse.citationstyles.org/u/Bruce_D_Arcus1)\
**Post date:** [May 26, 2009, 1:52pm UTC](https://discourse.citationstyles.org/t/delimiters/643/8 "2009-05-26T13:52:55Z")

</div>

OK …

…

> He’s needing to change the trailing delimiter on collapsed blocks.  
> Currently we can customize the delimiters within collapsed sets, but  
> use a single default delimiter between collapsed blocks. This would  
> need another option to supply the special delimiter value. Not sure  
> what to call it. Something like “after-collapse-delimiter”?

I think this discussion exposes other problems in how delimiter  
applies to a citation layout. I’ve hinted at this earlier.

Right now, “delimiter” is vague: does it refer to citation references  
within a citation group, or author groups within it?

So unless I just haven’t had enough caffeine yet, I think we first  
need to clarify this, and document it better.

---

<div class="post-metadata">

**Author:** ![Bruce\_D\_Arcus1](https://avatars.discourse-cdn.com/v4/letter/b/4bbf92/32.png) [@Bruce\_D\_Arcus1](https://discourse.citationstyles.org/u/Bruce_D_Arcus1)\
**Post date:** [May 27, 2009, 1:02pm UTC](https://discourse.citationstyles.org/t/delimiters/643/9 "2009-05-27T13:02:07Z")

</div>

…

> I think this discussion exposes other problems in how delimiter  
> applies to a citation layout. I’ve hinted at this earlier.
> 
> Right now, “delimiter” is vague: does it refer to citation references  
> within a citation group, or author groups within it?

So, for example, what does this mean?

```
    <names variable="author" delimiter="; ">
        <name delimiter=", "/>
    </names>

```

Would the result be “Jane, Doe; John, Smith”?

Related: why is the “delimiter-precedes-last” attribute on cs:name and  
note cs:names?

Am I wrong that this is all too fuzzy?

Bruce

---

<div class="post-metadata">

**Author:** ![Frank\_Bennett](https://yyz1.discourse-cdn.com/flex029/user_avatar/discourse.citationstyles.org/frank_bennett/32/198_2.png) [@Frank\_Bennett](https://discourse.citationstyles.org/u/Frank_Bennett)\
**Post date:** [May 27, 2009, 1:29pm UTC](https://discourse.citationstyles.org/t/delimiters/643/10 "2009-05-27T13:29:59Z")

</div>

> …
> 
> > I think this discussion exposes other problems in how delimiter  
> > applies to a citation layout. I’ve hinted at this earlier.
> > 
> > Right now, “delimiter” is vague: does it refer to citation references  
> > within a citation group, or author groups within it?
> 
> So, for example, what does this mean?
> 
> ```
> <names variable="author" delimiter="; ">
> <name delimiter=", "/>
> </names>
> 
> ```
> 
> Would the result be “Jane, Doe; John, Smith”?

No, the spacing in the straight-up name form is implicit. The  
delimiters are for joins between names, and between namesets. The  
semicolon in the example above would have no effect, because there is  
only one nameset. The comma could fall between the names: Jane Doe,  
John Smith. The delimiter on names comes into play when there are  
multiple namesets. So:

Jane Doe and John Roe (eds.); John Smith and Jane Brown (trans.)

> Related: why is the “delimiter-precedes-last” attribute on cs:name and  
> note cs:names?

Given that behaviour, delimiter-precedes-last is in the right place:

Jane Doe, and John Roe (eds.); John Smith, and Jane Brown (trans.)

> Am I wrong that this is all too fuzzy?

A lot of the confusion arises from the names (no pun intended) given  
to the elements. Names would be more clear as namesets. Name would  
be more clear as names. And there is the off-by-one naming of the  
collapse options.

The item I’m arguing for in this thread is pretty straightforward, if  
you set aside the problem of what to call it. As semicolon is used  
after any series of collapsed citations, instead of a comma. This is  
an aide to readability, highlighting that multiple references by a  
particular author are being cited – a fact that might otherwise be  
missed where a citation is truncated to the year suffix only.

There is a sound editorial reason for this rule (readability), and if  
we weren’t encountering it at the end of a long trek through a number  
of more demanding rules that also result in the elimination of a  
single space or the replacement of a single character, we probably  
wouldn’t think twice about it.

---

<div class="post-metadata">

**Author:** ![Bruce\_D\_Arcus1](https://avatars.discourse-cdn.com/v4/letter/b/4bbf92/32.png) [@Bruce\_D\_Arcus1](https://discourse.citationstyles.org/u/Bruce_D_Arcus1)\
**Post date:** [May 27, 2009, 3:23pm UTC](https://discourse.citationstyles.org/t/delimiters/643/11 "2009-05-27T15:23:44Z")

</div>

> > …
> > 
> > > I think this discussion exposes other problems in how delimiter  
> > > applies to a citation layout. I’ve hinted at this earlier.
> > > 
> > > Right now, “delimiter” is vague: does it refer to citation references  
> > > within a citation group, or author groups within it?
> > 
> > So, for example, what does this mean?
> > 
> > ```
> > <names variable="author" delimiter="; ">
> > <name delimiter=", "/>
> > </names>
> > 
> > ```
> > 
> > Would the result be “Jane, Doe; John, Smith”?
> 
> No, the spacing in the straight-up name form is implicit. The  
> delimiters are for joins between names, and between namesets. The  
> semicolon in the example above would have no effect, because there is  
> only one nameset. The comma could fall between the names: Jane Doe,  
> John Smith. The delimiter on names comes into play when there are  
> multiple namesets. So:
> 
> Jane Doe and John Roe (eds.); John Smith and Jane Brown (trans.)

Ah, right. So this clearly needs to be better documented.

> > Related: why is the “delimiter-precedes-last” attribute on cs:name and  
> > note cs:names?
> 
> Given that behaviour, delimiter-precedes-last is in the right place:
> 
> Jane Doe, and John Roe (eds.); John Smith, and Jane Brown (trans.)
> 
> > Am I wrong that this is all too fuzzy?
> 
> A lot of the confusion arises from the names (no pun intended) given  
> to the elements. Names would be more clear as namesets. Name would  
> be more clear as names.

Well, there’s actually three levels here:

1. what you call name sets are names grouped by role (cs:names)

2. the list of names specific to a role (cs:name)

3. the individual contributor (??)

> And there is the off-by-one naming of the  
> collapse options.
> 
> The item I’m arguing for in this thread is pretty straightforward, if  
> you set aside the problem of what to call it. As semicolon is used  
> after any series of collapsed citations, instead of a comma.

That’s one way to explain it; the way you understand it as an implementer.

But there’s another way to understand this: that it’s the delimiter  
for #2 above, but just that this style says “if number of contributors  
is 1, use ', ', else '; '”

---

<div class="post-metadata">

**Author:** ![Frank\_Bennett](https://yyz1.discourse-cdn.com/flex029/user_avatar/discourse.citationstyles.org/frank_bennett/32/198_2.png) [@Frank\_Bennett](https://discourse.citationstyles.org/u/Frank_Bennett)\
**Post date:** [May 27, 2009, 8:53pm UTC](https://discourse.citationstyles.org/t/delimiters/643/12 "2009-05-27T20:53:04Z")

</div>

> > > …
> > > 
> > > > I think this discussion exposes other problems in how delimiter  
> > > > applies to a citation layout. I’ve hinted at this earlier.
> > > > 
> > > > Right now, “delimiter” is vague: does it refer to citation references  
> > > > within a citation group, or author groups within it?
> > > 
> > > So, for example, what does this mean?
> > > 
> > > ```
> > > <names variable="author" delimiter="; ">
> > > <name delimiter=", "/>
> > > </names>
> > > 
> > > ```
> > > 
> > > Would the result be “Jane, Doe; John, Smith”?
> > 
> > No, the spacing in the straight-up name form is implicit. The  
> > delimiters are for joins between names, and between namesets. The  
> > semicolon in the example above would have no effect, because there is  
> > only one nameset. The comma could fall between the names: Jane Doe,  
> > John Smith. The delimiter on names comes into play when there are  
> > multiple namesets. So:
> > 
> > Jane Doe and John Roe (eds.); John Smith and Jane Brown (trans.)
> 
> Ah, right. So this clearly needs to be better documented.
> 
> > > Related: why is the “delimiter-precedes-last” attribute on cs:name and  
> > > note cs:names?
> > 
> > Given that behaviour, delimiter-precedes-last is in the right place:
> > 
> > Jane Doe, and John Roe (eds.); John Smith, and Jane Brown (trans.)
> > 
> > > Am I wrong that this is all too fuzzy?
> > 
> > A lot of the confusion arises from the names (no pun intended) given  
> > to the elements. Names would be more clear as namesets. Name would  
> > be more clear as names.
> 
> Well, there’s actually three levels here:
> 
> 1. what you call name sets are names grouped by role (cs:names)
> 
> 2. the list of names specific to a role (cs:name)
> 
> 3. the individual contributor (??)
> 
> > And there is the off-by-one naming of the  
> > collapse options.
> > 
> > The item I’m arguing for in this thread is pretty straightforward, if  
> > you set aside the problem of what to call it. As semicolon is used  
> > after any series of collapsed citations, instead of a comma.
> 
> That’s one way to explain it; the way you understand it as an implementer.
> 
> But there’s another way to understand this: that it’s the delimiter  
> for #2 above, but just that this style says “if number of contributors  
> is 1, use ', ', else '; '”

Ah! I got it. The example was ambiguous, isn’t it, and could be  
understood to turn on the number of names/namesets. If that were the  
case, I would agree that it’s a silly rule. But it’s not about the  
names, it’s about collapsing (I checked in the style guide linked by  
the forum poster). Here’s the example from the poster:

(Condit et al. 2000, 1995, 1998; Connell 1971, Cortes & Hunziker 1997)

Here’s the rule from the linked style guide:

Use commas (Yaz & Taz 1981, Ramirez 1983) to separate citations, BUT  
use semicolon for different types of citations (Fig. 4; Table 2) or  
with multiple dates per author (Yaz et al. 1982a, b; Taz 1990, 1991).  
Order references by year, then alphabetical (Azy 1980, Yaz 1980, Azy  
1985).

So it’s the collapsing of multiple cites (2000, 1995, 1998) that  
triggers the rule, not the number of creators (Condit et al.).  
There’s a glitch in the poster’s example – the sort is all messed up  
– but the use of the semicolon is consistent with the journal’s rule.

The explanatory text would read something like:

## Delimiter to use between any series of collapsed citations and a

## citation that is not collapsed.