{"id":578,"date":"2018-12-22T01:48:02","date_gmt":"2018-12-22T01:48:02","guid":{"rendered":"http:\/\/blueberryhaiku.com\/?post_type=project&p=578"},"modified":"2018-12-22T01:54:09","modified_gmt":"2018-12-22T01:54:09","slug":"front-matter-xml-extension","status":"publish","type":"project","link":"https:\/\/blueberryhaiku.com\/?project=front-matter-xml-extension","title":{"rendered":"Front Matter XML Extension"},"content":{"rendered":"

[et_pb_section fb_built=”1″ _builder_version=”3.18.9″][et_pb_row _builder_version=”3.18.9″][et_pb_column type=”4_4″ _builder_version=”3.18.9″ parallax=”off” parallax_method=”on”][et_pb_text _builder_version=”3.18.9″]<\/p>\n

\n
\n
\n
\n
\n
\n

\"<JATS-Con><\/a><\/p>\n

\n

The Front Matters: Capturing Journal Front Matter with JATS<\/h1>\n<\/div>\n<\/div>\n<\/div>\n
\n
\n

Authors<\/h4>\n

Rachael Carter<\/span>, MLS,<\/span>1<\/sup>\u00a0<\/span>Kathryn Funk<\/span>, MLIS,<\/span>2<\/sup>\u00a0and\u00a0<\/span>Rebecca Mooney<\/span>, MLS<\/span>\"corresponding<\/sup>3<\/sup>.<\/span><\/span><\/p>\n

Contact<\/h4>\n
1<\/sup>\u00a0<\/span>rcarter@dscs.com<\/a><\/div>\n
2<\/sup>\u00a0<\/span>funkkm@ncbi.gov\"corresponding<\/sup><\/a><\/div>\n
3<\/sup>\u00a0<\/span>rmooney@aaas.org<\/a><\/div>\n
NCBI\/NLM\/NIH<\/div>\n
\"corresponding<\/sup>Corresponding author.<\/div>\n
<\/div>\n
<\/div>\n
<\/div>\n<\/div>\n
\n
\n

PMC strives to be a comprehensive archive of biomedical journals. However, JATS currently provides no way to capture journal and issue specific front matter content, such as editorial boards, journal philosophy, submission instructions, etc. Our team developed an extension to the current Tag Suite, the pmc<\/g>–journalmatter<\/g>.dtd<\/g>, that captures these journal artifacts. In mapping multiple publishing models, we found that front matter exists in two basic forms: issue and standing. Our DTD aims to be flexible enough to accommodate a variety of user needs. It allows the users to update different elements of front matter without the burden of updating all elements by offering separate document types. The document types include: author information, cover, editorial board, publisher information, and general information. In order to encapsulate front matter content, we introduced new elements that work in conjunction with JATS.<\/p>\n

This paper will explore the need for a front matter specific JATS extension, the limitations of the article model to represent this type of information, our process for the data mapping and the rationale for subsequent development of new elements, and future implementation.<\/p>\n<\/div>\n

\n
<\/div>\n

<\/h2>\n

Introduction<\/h2>\n

For over ten years, PMC has sought to archive journal articles in a way that supports research in the biomedical community and the preservation of this content for future use. As noted in\u00a0<\/span>Preservation Strategies for Digital Libraries<\/i>, \u201cthe purpose of preserving things is to enable access to them at some unspecified date in the future, very probably for purposes not anticipated by the creators\u201d (Holdsworth, 2007<\/a>). Since the growth of digital publishing and archiving, there has been a conversation within the academic and library communities regarding what is being digitally preserved as well as what should be (Marcum, 2001<\/a>;\u00a0<\/span>Wheeler, 2010<\/a>). Dr. Marcum writes \u201cdecisions must be made about what will actually be saved for future use\u2026Will the content consist only of articles in a journal, or will it also include front matter (such as the names of the members of the journal\u2019s editorial board)?\u201d. Acknowledging this question, the National Library of Medicine requested that PMC investigate options for supporting the archiving and preservation of non-article content. Within the scope of PMC, we focused on front matter as the most significant part of non-article content as it helps define the historical context of the journal (Andrews, 2006<\/a>) and provides a more complete overall picture of a journal\u2019s publication.<\/p>\n

When the NLM DTD was first developed, its initial purpose was exclusively to capture article content. In 2001, PMC staff created the issue-admin.dtd, made available to publishers who expressed an interest in submitting additional material such as journal covers to the PMC archive. While still used today, the issue-admin.dtd is based on the pmc<\/g>-1.dtd<\/g>, a model that is no longer used. In addition, it only provides broad categories that do not give the user the means to tag front matter in a meaningful way. The pmc<\/g>-journalmatter.dtd is an extension to the more current version of the PMC tagging suite and seeks to offer semantic specificity in tagging. Although used interchangeably with \u201cfront matter\u201d, we used \u201cjournal matter\u201d in the naming of the DTD to reflect the current publishing environment where more and more journals are published online without a traditional \u201cfront\u201d.<\/p>\n

As recently as 2011, the publisher Atypon developed its own extension to JATs that specifically addresses the need to capture the table of contents (Markantonatos, 2011<\/a>). Atypon\u2019s Issue XML is designed to build functional tables of contents from a store of articles. Approaching the creation of a new DTD specific to PMC\u2019s needs, we envisioned a broader and different scope for the project. Our goal was to capture front matter as content, in the environment in which it was published, even if it exists outside of an issue.<\/p>\n

Aside from Atypon\u2019s Issue XML, there was no existing issue level<\/g> extension of JATS to capture front matter material, such as editorial boards and journal philosophies. We did not want to capture these elements using the existing JATs article model because,<\/g> simply put, this content is not an article. While the existing JATS model is flexible, there are limitations that would restrict the user\u2019s ability to capture non-article content. For example, there is no meaningful way to capture editorial boards or to tag journal metadata at a level higher than <article-meta>. Consequently<\/g> we needed to create elements and attributes to reflect front matter content accurately.<\/p>\n<\/div>\n

\n
<\/div>\n

<\/h2>\n

Process<\/h2>\n

After receiving the initiative from the NLM, we determined that looking at physical and electronic examples of journal front matter was the best first step (see\u00a0<\/span>Appendix I<\/a>). Making note of the common components present in our sampling, we created a list of several different content types that compose front matter. Appreciating that there are a variety of pieces that could comprise front matter, and that many journals have different approaches in their publishing models, we defined the following content types: cover, editorial board, publisher information, information for authors and general journal information. During our development of these content types, it became apparent that we needed to allow for the distinction between content that is associated with a specific issue (e.g.,\u00a0<\/span>PLoS ONE<\/i>, Volume 7, Issue 9) and content that is not (i.e., a continuous publication model, e.g.,\u00a0<\/span>Frontiers in Cancer Genetics<\/i>); this challenge will be discussed in the following section.<\/p>\n

The next step was to begin to define the different elements that would be present for each particular content type. We generated sample documents of how we anticipated each content type would be tagged (see\u00a0<\/span>Appendix II<\/a>). Working with the existing article model, we used the JATs DTD as a framework for our model while determining occasions where new elements would need to be created, eliminated or restricted. Based on those examples, we were able to see where non-article content diverged from article content and what new elements would need to be created as an extension of JATS. We then defined these new elements,<\/g> and decided where they would exist in our model. Combining these new elements with the applicable elements from JATS, we began developing the first iteration of the PMC Journal Matter DTD.<\/p>\n

Once the syntax of the DTD was established, our next step was to address our goal of creating a flexible DTD while defining what would be of importance to the PMC archive. For this reason, our content types are merely suggestions and not required. Consequently, valid XML could be created without a content type specified. This allows for adaptability for multiple publishing models and usage outside of PMC.<\/p>\n

In order to test the DTD, we then tagged actual content from our original journal samples. After making some minor adjustments, we generated tagging guidelines and documentation pertaining to the project. We then asked PMC journal managers to assess the DTD by tagging front matter samples. Along with the documentation, they were provided with both issue and non-issue samples, from multiple publishers, that covered a variety of content types. Final adjustments to the DTD and documentation were made based on the user feedback.<\/p>\n<\/div>\n

\n
<\/div>\n

<\/h2>\n

Challenges<\/h2>\n

We faced two primary challenges while working on the creation of this DTD. The first was how to generate a foundation for organizing and labeling the front matter content. The second was answering the question of\u00a0<\/span>can we tag all of this content in one document?<\/i>\u00a0<\/span>In the end, these challenges were resolved by creating attributes for the root element.<\/p>\n

We first determined that issue and non-issue material must be captured and labeled separately because there are instances of front matter content that are not published in an issue. For example,\u00a0<\/span>BMC Cancer<\/i>\u00a0<\/span>publishes articles individually, on a rolling basis, and without any issue information specified; yet, they still maintain information that would go in the front of an issue, such as an editorial board. Consequently, of the two attributes on the root element, <journalmatter>, the first is @journalmatter-type, which can either be \u201cissue\u201d or \u201cstanding\u201d. This attribute also prevents a hybrid of issue and non-issue content in the same XML document so content that changes in each issue can be more easily updated.<\/p>\n

The second attribute on our root element, @content-type, allows the user to categorize the nature of the material being captured. The primary goal behind separate documents for each content type is for flexibility in use and rendering. Some pieces of a journal\u2019s front matter change more frequently than others. For example, a journal\u2019s editorial board may change annually, whereas a journal\u2019s philosophy is more likely to remain the same over<\/g> several years. Our model allows publishers to update specific documents as they change, rather than resubmitting all front matter content. It also permits a user to more easily tag or render the content in a way that best suits their needs, whether it is a single section of front matter,<\/g> or all of it. As an example, one could quickly render all the past editorial boards on a single page; or choose to display a combination of content types, such as cover and publisher information, together. Alternately, the user could choose to capture all of the front matter in one document if that is what best suited his or her needs.<\/p>\n<\/div>\n

\n
<\/div>\n

<\/h2>\n

Discussion of DTD<\/h2>\n

The PMC Front Matter DTD calls upon two files from JATS, journalpubcustom<\/g>-modules.ent and modules.ent, and also consists of two custom files created specifically for the project: pmc<\/g>–journalmatter<\/g>.dtd<\/g> and pmc<\/g>–journalmattercustom<\/g>.ent.<\/p>\n

\n

<\/h3>\n

The Root Element<\/h3>\n

To begin tagging a document you must first define the root element attributes. The root element on the DTD is <journalmatter> with two attributes required: @journalmatter-type and @content-type. Below is an example of the root element on an issue-based, editorial board document:<\/p>\n

\n
<journalmatter journalmatter-type=\"issue\" content-type=\"edboard\">\n<\/pre>\n<\/div>\n
 <\/p>\n
\n
Journal Matter Type<\/h4>\n
The first attribute you must select is the @journalmatter-type: issue or standing. The issue attribute should be used for materials that are published in an issue. The standing attribute should be used for non-issue based materials that are not published in an issue, but are static, such as information published on a website. You cannot have both \u201cissue\u201d and \u201cstanding\u201d in a single XML document. Depending on the content being tagged, it is possible for a single journal to tag some material as standing and some as issue. For example, a journal\u2019s submission instructions may be submitted as \u201cstanding\u201d since it is unlikely to change frequently over time; whereas a cover will presumably change with each issue.<\/p>\n<\/div>\n
\n
<\/h4>\n
Content Type<\/h4>\n
The second attribute on our root element, @content-type, allows for six different possible values:<\/p>\n