Suggestions on Requesting New Styles

bdarcus

I'd like to suggest some changes to style request. At this point, there are now over 1000 styles, so any new style is likely to be either exactly like an existing one, or to differ only trivially. As such, I'd like to suggest the following template.

1. what existing style is closest to the style you need? If it's exactly the same, tell us. If not ...


2. itemize, in plain English, the precise differences that need to be implemented


3. give us a link to the style guide

The first two steps are the most important, as the most time-consuming part of style creation is not the technical part, but rather just understanding how a style works. The less a volunteer has to do to figure that out, the more likely it is they can help.

I'd also suggest it might be more helpful to have separate forum topics for each style request? Does that make sense?

PS - can someone (Dan?) fix the CSS on this forum to properly display the lists? The above includes an ordered list, but it displays as unordered (and has funky alignment too!).

Rintze

At this point, there are now over 1000 styles, so any new style is likely to be either exactly like an existing one

Please note that there are currently (only) 96 independent styles in the repository, and a little over a 1000 dependent styles, so the diversity of styles is not that great.

what existing style is closest to the style you need?

Although I think the support for dependent styles is a useful addition, the large number of dependent styles does make it a lot more difficult to browse the repository in search of the style that best matches what you're looking for.

give us a link to the style guide

An additional consideration is that, especially for the natural sciences, style guides generally plainly suck. If you're lucky you get a quite diverse list of example references (formatting rules are quite rare), but sometimes the examples are limited to just one or two journal articles and a book chapter. In these cases another thing that has to be done is: either a) request that the journal's publisher improves the "guidelines for authors"-section or b) look through some publications in search of how undocumented item types should be formatted.

noksagt

Given the number of styles that are requested, I don't know if separate threads are beneficial or not. I agree that having access to th three pieces of information that you list would be very useful. A fourth (maybe optional) item would be a link to a freely available article from the publication--style guides are often (and unfortunately) too brief to apply to all situations & having more examples is helpful.

We should also strive to make requirement 1 as easy as possible. Some ways to do this:

1. list the ca. 104 unique zotero styles separately (going through 100 styles is easier than 1000)
   
2. organize styles by category (particularly info-class (author date vs. numeric vs...)
   
3. in CSL, add a field for the publisher that uses the style.
   
4. provide automated tools to find the closest matching style

A work-around to the third point is to manually include the publisher in the 'summary:' then they can still be 'grep'ed for information. Eventually having a better search tool for CSL styles would help.

Re. the last point: I often manually markup a citation in HTML & then calculate the edit distance between that file and the examples from the 104 or so core styles. This benefits me quite a bit, but isn't robust enough for others. To make a tool like this easier for others: strip COinS and other extraneous information that people won't manually enter from the compared data; make a brief description of markup available; run HTML tidy on both existing styles & the user-inputted styles. This is weighted towards differences in the bibliography over citations, so having the info-class category would definitely help a tool like this.

bdarcus

All good points (but, rintze, the fact that most of those 1000+ styles are referenced supports my larger point; there's not really as much variation as we tend to think). Part of the reason to have a rich metadata section of the styles is precisely to make it easier to do this sort of thing. The current one-big-list approach isn't too useful, obviously.

Another issue that may need attention is somewhat richer previews.

Tjowens

With this conversation in mind I have drafted up a page with instructions for requesting styles. When you have a few moments give it a look and post feedback here.

Rintze

I have drafted up a page with instructions for requesting styles. When you have a few moments give it a look and post feedback here.

- Maybe it would clarify things a bit if you included some info on the (lack of) support for other style formats, i.e. that Endnote style read-only support will be introduced in Zotero 1.5 (although it is meant mostly as a stopgap measure), and that bib styles aren't supported.
- The style preview box only includes a small selection of items/item types. Perhaps you could include a short discussion on how cslpreview.xul can be used to generate previews from the items in one's library from all installed styles. It's quite a fast way to see how different styles handle different kinds of items and their bibliographic fields.
- The style repository doesn't really explain the difference between dependent and independent styles, and neither does this page.

migugg

Why not add some categories and maybe even sub-categories to the style-repository? The main categories could be: numbered, author-date, author title, date at the end, full quotation in footnote. Then the subcategories could be: year in bracket/ most common separator: comma, period/no publisher/publisher.
These categories are fairly easy to identify and would narrow down the list of styles to comparable styles quickly without makeing a long search and comparison necessary.

bdarcus

@migugg: yes, there are a number of different ways to usefully slice the categorization:

• class (in-text vs. note)


• type (author-date, author-title, number, etc.)


• field (history, chemistry, etc.)


• base style (the style on which dependent styles are based)

An interface that exploited those would be a long way towards streamlining access.

While my ideal is a full web application to deal with accessing, creating, and commenting on styles, I wonder if there might be a more incremental bang-for-buck solution; say a simple Javascript thing where each style gets a div with appropriate class attributes (for apa, say, "psychology author-date base") and the result list could just be filtered based on user-chosen options?

bdarcus

Trevor: looks good, but I agree with Rinte's point about adding some upfront clarification of what CSL is and how it fits in the Zotero universe. E.g. something like "Zotero's primary citation styling format is CSL, which is powerful, flexible, and open." with a link to the xbib page.

dstillman

A new web application for managing styles is in development, though due to other priorities it's probably a couple months off. I'll add a quick filter to the current page when I have a moment. Most of the styles are tagged already, though the schema could use a few more fields.

bdarcus

Dan: great.

Feel free to add those fields to the schema as you wish. The current list is more-or-less "what I thought of off the top of my head," and there's no reason to be conservative about adding new fields.

scot

Trevor: I like the page. Consider expanding the heading 'Differences from similar style:', since that particular point is one which we want to encourage the requesters to focus on. Perhaps:

The next step is to itemize the precise differences that need to be implemented to make that existing style into the style you need. Doing this will give the creation of the style a big jump-start. The most time-consuming part of style creation is not the technical part, but understanding how a style works and how it differs from existing styles . The less a volunteer has to do to figure that out, the more likely it is they can help.

Tjowens

Thanks for the feedback. I have incorporated some suggestions.
- added brief mention of cslpreview
- added a bit of text explaining what CSL is
- added a paragraph about Independent vs. dependent styles
- added some of scot's text to the 'Differences from similar style:' section

Feel free to respond with additional comments here, or to refine wiki page by appending ?do=login to the URL and logging in with your wiki account to edit it. (If you don't have a wiki account append ?do=register to the URL to register for one.

scot

The new page is excellent.

komrade

for Dan Stillman!

It would be really great if you could add a quick filter to the Style repository when you have a moment.
Searching through hundreds of styles is a real pain.

(re: your comment of Dec 3 2008 above)

migugg

bump
just tried to find a style, it is a real pain. There are now roughly 200 independent styles.
The normal use case is this: I am given a style of a journal. The journal name does not show up in the style repository. I do have to find a similar style to request one. To do this, I have to memorize the style and then hover more or less randomly over hundreds of styles and compare, whether the styles uses a comma or no comma after the title etc.
uff. You could save the people out there hundreds of hours of work (and prevent also lots of misguided style requests) with a navigation help.