[jats-list] Tagging supplementary material in JATS

Subject: [jats-list] Tagging supplementary material in JATS
From: Alexander Schwarzman <aschwarzman@xxxxxxxxx>
Date: Sat, 26 Apr 2014 19:24:52 -0400
I wonder what JATS list members think would be an optimal way of
tagging supplementary materials in JATS 1.1d1.

Suppose that an article contains a supplementary section. The section
is useful, but not critical, for understanding the article; that is,
in the NISO/NFAIS terminology, it is "additional supplemental
material." The section itself contains a narrative and two videos. The
narrative part of the section will be presented in a PDF format, with
links to the videos (for archival reasons, videos are not to embedded
into the PDF). The supplementary section, i.e., the PDF with the
videos linked to it, will only appear in the HTML representation of
the article.

Suppose further that the processing system, i.e., the publishing
platform and the databases that feed it, need to "know" about not only
the narrative part of the section (the PDF) but also about the videos
linked to it, e.g., in order to be able to render the videos, for
statistical/reporting purposes, etc.

The first impulse might be to mark up the structure as:

            <p>See <inline-supplementary-material id="S1"
                    >Supplement 1</inline-supplementary-material> for
supporting content.</p>
            <media id="S2" xlink:href="article-id-s002.avi">
                <label>Media S1</label>
            <media id="S3" xlink:href="article-id-s003.avi">
                <label>Media S2</label>

This "flattens" the hierarchical structure of the supplementary
section (i.e., it cannot be inferred that the videos are "children" of
the narrative) but, at least, it gives enough information to the
processing system about all supplementary components. However, as the
Tag Library documentation reminds us, the <media> elements within the
<supplementary-material> element are not intended to be used this way;
rather, they are intended for providing a representation of a
supplementary material: for example, if the supplementary object being
tagged is itself a 5-hr long video, then the <media> element within
could be used to point to its little snippet or a promo. Thus the
above tagging, while technically valid to the Tag Set, goes in fact,
against JATS's spirit and constitutes "tag abuse."

What, then, about the following tagging:

    <sec sec-type="supplementary-material">
        <p>See <inline-supplementary-material id="S1"
                Supplement 1</inline-supplementary-material> for
supporting content.</p>
        <supplementary-material id="S2" xlink:href="article-id-s002.avi">
            <label>Media S1</label>
        <supplementary-material id="S3" xlink:href="article-id-s003.avi">
            <label>Media S2</label>

Here, we "wrap" each supplementary component in its own
<supplementary-material> element, flattening hierarchy even further,
but we don't seem to violate the spirit of JATS. Or do we? What markup
would you suggest?

Note that we have more than one PDF files with linked videos or more
than one ZIP files with the contents of each being described JATS
makes it rather difficult to infer which components belong to which
supplementary object. But should it even care?

In addition to the technical question of which markup is more
appropriate, this makes me think of two issues:
- is JATS an appropriate tool to express the (hierarchical) structure
and contents of supplementary materials? It's a *Journal Article* Tag
Set after all, and not necessarily the right expressive instrument for
dealing with the sprawling hierarchies of (potentially infinitely
nested) components that may have a certain relationship to the said
- If it is though, what would be the best way to achieve that goal?
Fully integrate the NISO/NFAIS RP, which is designed to deal with such
hierarchies? Introduce an explicit "Degree of Importance" attribute on
the <supplementary-material> element, e.g., Integral, Additional, etc.
(this does not have to be binary; it could be a "sliding scale")? Or
perhaps something else, simply to clarify the intended semantics of
the relevant JATS elements?

Thank you,

Alexander ('Sasha') Schwarzman, Content Technology Architect
phone: +1.202.416.1979 | e-mail: aschwarzman@xxxxxxx

The Optical Society (OSA)
2010 Massachusetts Ave., NW
Washington, DC 20036 USA

Current Thread