Chicago Style

I notice that in the newest revision of the chicago-a.csl file, the
"journal-article" and “article” styles for CMS have been merged into one.
This will result in incorrect output when working with journal articles,
since the Chicago Manual of Style defines different formats for these two
types (see CMS 15th ed., 17.150). I now notice that both the old style and
this style are also handling issues incorrectly, but to handle both
magazines and journals properly, these two types will need to be split
apart. Is it okay if I commit a corrected version? (This time, I’ll validate
it beforehand.)

Simon

I notice that in the newest revision of the chicago-a.csl file, the
"journal-article" and “article” styles for CMS have been merged into one.
This will result in incorrect output when working with journal articles,
since the Chicago Manual of Style defines different formats for these two
types (see CMS 15th ed., 17.150).

Yeah, practically speaking I think these sort of rules are silly. I
prefer 17.161; if you have the data, print it.

Omitting volume and issue with magazines seems just a convenience.
Note the language in 17.150 is rather weak; they don’t say “you
should/must do x” but rather “usually, x.” Re-read the last sentence
in particular, which seems to support my reading above that separate
templates aren’t really required.

But …

I now notice that both the old style and
this style are also handling issues incorrectly, but to handle both
magazines and journals properly, these two types will need to be split
apart. Is it okay if I commit a corrected version? (This time, I’ll validate
it beforehand.)

… if you like, sure, that’s fine.

The style was invalid for two reasons:

  1. the value of the name attribute was wrong

  2. the position of the cs:type element was wrong (while RNG has an
    unordered “interleave” content model, it does not apply to
    attribute-based content; as a result, the required three templates
    must come first, in a particular order)

Bruce

OK, I see; they have dates rendered a little differently.

Hopefully, over time this sort of weirdness will whittle away!

Bruce

Actually, one thing to consider when you design a style where you need
to include more than the required defaults is what is default, and
what unique? E.g. does it make more sense to treat the journal article
handling as default and add an “article-magazine” template, or vice
versa?

I’m not providing an answer here; just pointing it out. It might
indeed make more sense to go this way, since “article” then covers all
of them except journal articles.

Bruce

I notice that in the newest revision of the chicago-a.csl file, the
"journal-article" and “article” styles for CMS have been merged into one.
This will result in incorrect output when working with journal articles,
since the Chicago Manual of Style defines different formats for these two
types (see CMS 15th ed., 17.150).

Yeah, practically speaking I think these sort of rules are silly. I
prefer 17.161; if you have the data, print it.

Omitting volume and issue with magazines seems just a convenience.
Note the language in 17.150 is rather weak; they don’t say “you
should/must do x” but rather “usually, x.” Re-read the last sentence
in particular, which seems to support my reading above that separate
templates aren’t really required.

It may not be 100% incorrect, but there’s no reason the CSL file has to act
as if it doesn’t know if it is or isn’t a journal article. By following the
rules for these various citation styles, we can produce the best output
possible and (hopefully) speed adoption. This means at some point I’ll also
have to fix newspapers in the APA file.

Chicago also says that generally, you shouldn’t print the article page
numbers for an electronic citation, but there’s not really any way to signal
that in CSL. I think that this issue is minor enough that no one will really
care about it, but it (and the possibility that others may encounter more
significant situations like it) makes me wonder if we need a real "if"
construct of some kind in CSL.

I now notice that both the old style and
this style are also handling issues incorrectly, but to handle both
magazines and journals properly, these two types will need to be split
apart. Is it okay if I commit a corrected version? (This time, I’ll validate
it beforehand.)

… if you like, sure, that’s fine.

The style was invalid for two reasons:

  1. the value of the name attribute was wrong

  2. the position of the cs:type element was wrong (while RNG has an
    unordered “interleave” content model, it does not apply to
    attribute-based content; as a result, the required three templates
    must come first, in a particular order)

Yeah, I figured that out when I found out that it wasn’t correctly
formatting journal articles. It now validates.

And now, for another exceedingly minor issue:

<access>
  <url/>
  <text term-name="accessed" prefix=" ("/>
  <date suffix=")">
    <month suffix=" " text-transform="capitalize"/>
    <day suffix=", "/>
    <year/>
  </date>
</access>

This works as long as you have the access date, which there’s not really any
reason not to have. If you don’t have the access date, it looks completely
awful. I’m wondering if we couldn’t do something like:

<access>
  <url/>
  <date prefix=" (" suffix=")">
    <text term-name="accessed"/>
    <month suffix=" " text-transform="capitalize"/>
    <day suffix=", "/>
    <year/>
  </date>
</access>

isn’t currently allowed in the element, but it’s an easy way
of fixing this problem.

The other thing I could do is assume the access date is today if none is
given, but this seems like something that should be fixed in the style? Not
sure.

Simon

It may not be 100% incorrect, but there’s no reason the CSL file has to act
as if it doesn’t know if it is or isn’t a journal article.

Yeah, I just want to encourage people to get the defaults right,
rather than always just create a new template, because then the
formatting is generally more robust.

Chicago also says that generally, you shouldn’t print the article page
numbers for an electronic citation, but there’s not really any way to signal
that in CSL. I think that this issue is minor enough that no one will really
care about it, but it (and the possibility that others may encounter more
significant situations like it) makes me wonder if we need a real "if"
construct of some kind in CSL.

But I actually think this is an unhelpful rule. Consider a NYT
article. Say I cite it and include the URL, which is inaccessible two
months later. It’s helpful to have the page number there too if it’s
in the data.

I do (or did) have some sort of rule attribute that had to do with
volumes or issues, but I’ll have to look and remind myself.

I agree it makes sense to be pretty true to the styles. OTOH, no one
(journal publishers, professors, etc.) will ever complain if you have
too much information in a citation.

If i had to guess, I’d say we’re already more compliant than
alternatives (Endnote, etc.). I did publish a book with a major
publisher with an earlier version of this system after all, complete
with two copy-editors poring over the citations!

[…]

isn’t currently allowed in the element, but it’s an easy way
of fixing this problem.

And I think more correct. Do you feel comfortable adding it to the
schema yourself?

Bruce

But I actually think this is an unhelpful rule. Consider a NYT
article. Say I cite it and include the URL, which is inaccessible two
months later. It’s helpful to have the page number there too if it’s
in the data.

I completely agree. If only the editors of the Chicago Manual of Style were
so enlightened.

isn’t currently allowed in the element, but it’s an easy way
of fixing this problem.

And I think more correct. Do you feel comfortable adding it to the
schema yourself?

Done.

Simon