We’re now running into a situation where there are many journals
with identical styles, and we need a good way to handle this problem
from the perspective of developers and of end users. The CSL schema,
for example, could support multiple tags the way it currently does
for “category” so that several journal titles could be added to a
single CSL. End users might want to have an interface in preferences
that allows them show/hide styles in the export/citation dialog.
Users might, for example, want to have “Earth and Planetary Science
Letters” and “Metabolic Engineering” both appear, and not care to
know that having one or the other or both all point at the same CSL.
It seems to like we need to allow multiple titles, each with its own
set of categories, preferably without replicating all other metadata.
Of course, this would mean that we need to break (slightly) with Atom.
I’d propose something like:
http://purl.org/net/xbiblio/csl/styles/apa.csl
2007-09-06T06:36:07+00:00
Simon Kornblith
@Simon_Kornblith
American Psychological Association
Journal of Primitive Zoology
Here, is used for similarity with Atom. A more intuitive title
might be appealing, given that this is not actually an Atom feed.
Outstanding questions:
Should each style have one main title outside of individual
tags?
Does each need its own ID? (This is, IIRC, required in
Atom, but again, this isn’t a feed, and I’m not sure this would serve
any purpose.)
I think there are two cases here, one where a journal might say, “References
for this journal should be formatted as in journal XXX”. (or as in, say,
Chicago Manual of Style). In this case I think a link is required, as we
don’t necessarily want to pollute a base style with lots of journal names.
However probably the more common case is that a journal uses a style that is
identical to another journal. This might be convergent evolution, copying or
evidence of a house style by the publishers. In some cases the
specifications between two journals use the same format and the same exact
examples. In other cases the examples are different, but the outcome the
same.
For the time being, having multiple titles for the same style seems
convenient, if they diverge, then a copy will have to be taken and the split
made. I’ve done one style just by renaming and retitling in the zotero
archive, but this seemed counterproductive in the long run - especially in
the FEMS journals where a style is shared between 6 journals. As I fully
expect there to be bugs in these styles it also means I will have to fix N
different files.
I think there are two cases here, one where a journal might say,
“References for this journal should be formatted as in journal XXX”. (or
as in, say, Chicago Manual of Style). In this case I think a link is
required, as we don’t necessarily want to pollute a base style with lots
of journal names.
However probably the more common case is that a journal uses a style
that is identical to another journal.
I’m not so sure these are two different cases. At first glance, I’d
argue both involve different named styles using the same formatting rules.
Identification in CSL (as in Atom) is not based on title, but on ID
(URI). The title is just a convenient label.
This might be convergent
evolution, copying or evidence of a house style by the publishers. In
some cases the specifications between two journals use the same format
and the same exact examples. In other cases the examples are different,
but the outcome the same.
For the time being, having multiple titles for the same style seems
convenient, if they diverge, then a copy will have to be taken and the
split made…
But won’t that break things for the hypothetical example I gave?
This seems reasonable to me, although shouldn’t be in
because it’s not metadata. Maybe:
Journal of Primitive Zoology
http://www.zotero.org/styles/mla
2007-09-06T06:36:07+00:00
This seems like the most reasonable way of dealing with a journal that
takes plain APA style. It’s probably not the most reasonable way of
dealing with a set of journals with the same style (e.g., PLoS Biology
vs. PLoS Medicine, or the various Springer styles), which would do
better with a single CSL that simply contained all of their names.
But, in the interest of simplicity, I think that this would work fine.
Not sure I’m quite using Atom conventions correctly, but I’m sure you
get the general idea. The rest of the file would probably be empty.
I’m not sure the best solution, but we need to be careful, since the
details have far-ranging implications.
This seems reasonable to me, although shouldn’t be in
because it’s not metadata. Maybe:
Journal of Primitive Zoology
http://www.zotero.org/styles/mla
2007-09-06T06:36:07+00:00
This seems like the most reasonable way of dealing with a journal that
takes plain APA style. It’s probably not the most reasonable way of
dealing with a set of journals with the same style (e.g., PLoS Biology
vs. PLoS Medicine, or the various Springer styles), which would do
better with a single CSL that simply contained all of their names.
But, in the interest of simplicity, I think that this would work fine.
So can I get some votes on this?
The change would involve adding an optional “source” attribute on the
root. Where present, a processor would instead load the content
(everything but the info element) from the style identified with that URI.
As I mentioned previously, this is a trivial change with big
implications, so I want to make sure everyone’s comfortable with it.
I also need some feedback on the following changes to the info element:
Looks good to me. Is there any thought to a source part of the info that
could be used to refer to documentation about the style, such as the current
source that I’ve been (ab)using.
Looks good to me. Is there any thought to a source part of the info that
could be used to refer to documentation about the style, such as the
current source that I’ve been (ab)using.
a) some way to specify from what style a variant style descends. There
are really only a handful of core styles, with almost infinite variation
beyond that. I’d to be able to capture that in the metadata.
b) a pointer to some canonical location for a journal or a book publisher
c) a pointer to a book or webpage that documents the style, but which
may be at a separate domain from the homepage
We did have much of this earlier, but it dropped out in the transition
ot the new schema.
Any thoughts on any of this (use cases, the markup details, etc.)?
BTW, I just checked in the start of the bluebook style to the Zotero
repo. I’m experimenting with seeing how little I can rely on types.
It looks like in most cases CSL automatically places commas and
periods within the quotation mark if it follows a title enclosed in
quotes. This doesn’t seem to happen, however, in the “layout” element
when the citation includes no page number and the title is the last
item in the citation (this happens from time to time). To simplify:
If no locator is provided, the result is:
White, “Battle of the Century”.
But it should be:
White, “Battle of the Century.”
Note that the semicolon should fall outside of the quotation mark:
White, “Battle of the Century”; Jacob, “Eventful Transformations.”
I may be missing something, but if I am right, can you fix this?
Well, the obvious way is to include quote entities in the prefix and
suffix instead of using the “quotes” attribute. Alternatively, since
positioning of quotes should be a locale-specific option (although I’m
not sure if that’s set up properly in Zotero ATM), the user could
switch the locale to EN-GB. Perhaps we need to specify "default"
locale for each style in , and Zotero should use this unless
overridden?
I don’t think the locales will be a good solution. The journal in question
says that that is the house style. So if you wish to submit an article to
the journal in its specified format, that’s how it should look. So typically
people will want to select that style and submit the paper, not want to go
changing locales and stuff.
I tried putting ` and ’ as prefix and suffix, but it doesn’t look very good.
I;m not sure what is used in word but its not the regular ascii single
quotes. Maybe there is an HTML or Unicode character that is better?
Well, the obvious way is to include quote entities in the prefix and
suffix instead of using the “quotes” attribute. Alternatively, since
positioning of quotes should be a locale-specific option (although I’m
not sure if that’s set up properly in Zotero ATM), the user could
switch the locale to EN-GB. Perhaps we need to specify “default”
locale for each style in
this seems like the best solution, especially since zotero javascript
already includes british style quotes option:
this.useBritishStyleQuotes = false;
could this be switched on and off depending on style?
still doesn’t solve my issue though–let me know if it’s
unsolvable, in which case i’ll stop worrying about it.
Well, the obvious way is to include quote entities in the prefix and
suffix instead of using the “quotes” attribute.
No, I think this is the wrong solution; or at least, it’s one we
rejected earlier (it’s how we did it previously). Prefix and suffix
should not contain ANY locale-specific content.