Structurally-Based Citation Formatting
Existing citation formatting specifications — from Endnote and Reference Manager style files to BibTeX — are based on awkward and limited data models that depend upfront on type classifications that can only ever apply to narrow communities. I propose instead formatting engines be based on lessons learned from more structurally-based and comprehensive metadata models as expressed in standards such as Dublin Core and MODS.
Introduction
Existing citation formatting specifications are based on awkward and limited metadata models. Consequently, the formatting specifications themselves are likewise limited. The most signifcant problem is their reliance on concrete classification types, which by definition means they can only ever accurately format those references which are specifically defined. All else falls through the cracks of generic and miscellaneous types.
What I propose instead is basing formatting classification on structural concepts drawn from newer metadata models such as Dublin Core and MODS. Here what matter is not whether a formatting engine is dealing with books or articles, but the structural relationships that define those very objects.
My concrete proposal is that classification distuinguish between class and type. Class is broad and structural — including those objects that stand alone, and those that are part of other objects — while types are concrete and flexible.
Examples
Rather than argue this in the realm of theory, let’s look at some concrete examples from two comprehensive style specification manuals that serve quite different communities: the Chicago Manual of Style, and the ALWD Citation Manual. I divide the examples by the classes proposed above.
Monograph
A monograph is a standalone object, which may or may not contain parts. Examples include books, legal treaties, stamps, statues, web sites, musical albums, lectures, plays and data sources.
In contrast to parts, monographic citations only contain one title and one group of names/creators. Otherwise, most monographic objects are formatted essentially the same. Let’s take a few examples from the AWLD. Color key in the sidebar (if your browser renders CSS correctly at least).
The first example is a book:
Let’s break the citation down structurally:
This is straightforward enough.
Now let's look at a musical album. It shares the exact same structure as a book, and consequently the exact same formatting ... save one detail.
The only difference with the musical album is that it includes the medium information. Many citation styles specify the inclusion of such information for non-text media.
For a final example, here is a speech:
This adds location information generally needed for unpublished and online sources.
Parts ...
... in Monographs
The most common part contained in a monograph is a book chapter.
Let’s compare the above with a song on an album:
... in Serials
Legal case citations are also generally published as parts of periodicals, called court reporters. As there is no author per se, the case title functions in its place. Legal citations also often include related citations; in this case the citation history.