PMC Tagging Guidelines: General Tagging Practice

PMC prefers to receive data associated with DTDs but will accept content associated with W3C XML Schema, RelaxNG, or RelaxNG compact syntax schema types.

A schema may be associated with an XML document in several ways. A DTD can be referenced using a DOCTYPE declaration, a W3C XML Schema can be referenced using the @schemaLocation or @noNamespaceSchemaLocation attributes on the root element, or any schema type may be associated with the <?xml-model?> processing instruction defined by the W3C (http://www.w3.org/XML/2010/01/xml-model/).

In PMC, we strongly prefer the DOCTYPE declaration for associating DTDs and the @schemaLocation or @noNamespaceSchemaLocation attributes for W3C XML Schema association, but we will accept the <?xml-model?> processing instruction for DTD, W3C Schema, RELAX NG, and RELAX NG compact syntax when the following rules are observed:

1. The processing instruction(s) must be placed after the XML declaration and before the root element.
2. Each processing instruction MUST have an @href pseudo attribute. The content of this attribute must be either the filename of the schema or the complete URL of the schema (e.g. "JATS-journalpublishing1.rng" or "https://jats.nlm.nih.gov/publishing/1.0/rng/JATS-journalpublishing1.rng"). Local or relative paths should not be used (../schemas/JATS-journalpublishing1.rnc or C:\work\schemas\JATS-journalpublishing1.rnc).
3. Each processing instruction MUST identify the schema type that is being referenced with either an @type or @schematypens pseudo attribute according to the list below:
   1. for DTD, set "type='application/xml-dtd'"
   2. for XSD, set "@schematypens='http://www.w3.org/2001/XMLSchema'"
   3. for RNG, set "@schematypens='http://relaxng.org/ns/structure/1.0'"
   4. for RNC, set "type='application/relax-ng-compact-syntax'"

<?xml-model href="JATS-journalpublishing1.rng" schematypens="http://relaxng.org/ns/structure/1.0"?>

<?xml-model href="https://dtd.nlm.nih.gov/publishing/2.3/xsd/journalpublishing.xsd"
     schematypens="http://www.w3.org/2001/XMLSchema"?>

<?xml-model href="journalpublishing3.dtd" type="application/xml-dtd"?>

If more than one <?xml-model?> processing instruction is provided in an article, or if they are provided along with a DOCTYPE declaration or article attributes that reference an XSD file, PMC will process the articles in the following way:

1. DTD (from DOCTYPE declaration or <?xml-model?>)
2. XSD (from @noNamespaceSchemaLocation or <?xml-model?>)
3. RNG (from the <?xml-model?>)
4. RNC (from the <?xml-model?>)

PMC supports XML tagging practices as defined by the JATS4Reuse Recommendation on Accessibility that allow accessible renditions of the articles to be made from the source XML. This recommendation bases its set of tagging best practices on the following definition:

• "Accessibility is the practice of ensuring there are minimal barriers to access and interact with websites and other electronic information for users with physical or situational disabilities."

Specific tagging for XML best practices that support accessibility can be found in the following locations in the Tagging Guidelines. Practices that relate specifically to accessibility are marked with the text "Accessibility:".

General Tagging Practice: Character Encoding and Special Characters.

Elements: <alt-text>, <graphic>, <inline-graphic>, <media>, <table>, <th>, <caption>, <long-desc>, <ext-link>, <sec>, <title>, <list>, <alternatives>, <aff-alternatives>.

Document Objects: Figures, Inline images, Tables, Audio and Video, External Links, Lists, Multi-Language Articles, Math, Accessibility Metadata, Alternate Versions of a Single Object, Headings.

Use title case capitalization for PMC, particularly in <article-title> and <subject>.

When more than one article starts on the same page in a printed journal, treat each as an individual article. Tag each in its own file with a unique filename. Articles that start on the same page will have the same <fpage>. Use the @seq to assign sequence letters so that each article will have a unique fpage/sequence.

Generally speaking, all articles for a journal in PMC have the same delay time between the publication date and the time they are available in PMC. In some cases, however, these delays need to be set at the article level rather than at the journal level. See "Release Delay" under Processing Instructions for details on how to tag embargo delays for individual articles.

Do not use empty elements for formatting or any other purpose.

All required elements (that are not defined as EMPTY in the DTD) should have content.

As a rule, use formatted text (<bold>, <italic>, <sc>, etc.) only to set off a piece of information. Do not set entire elements in formatted text. For example, if a <title> is set completely in boldface, do not tag the <bold>. However, if a title has a word or some words set in boldface for emphasis, tag those words using <bold>. Mainly this will apply to <title>, <p> in <abstract> and <label>. It might also show up in <aff> and special sections, like <ack>.

Based on the agreement between the publisher and NLM, PubMed Central may accept non-English articles and/or English articles with non-English parts (titles, abstracts, etc.).

Non-English content must be identified with @xml:lang. Unlike nearly all other attributes in XML, the value of language is inherited. This means that all elements inside the one with the language attribute (its descendants) are assumed to be in the same language unless they explicitly set their own @xml:lang attribute.

Specify the language in @xml:lang using two-letter, lowercase language codes as described in IETF RFC 5646. Values may be obtained from the IANA Language Subtag Registry: http://www.iana.org/assignments/language-subtag-registry.

The rule for tagging language is that the primary language of the article should be set in the @xml:lang on <article>. Any item within the article that is in a different language must be tagged with an @xml:lang to identify the language of that piece.

Note: English is the default value for @xml:lang on <article>, <response>, and <sub-article> and does not need to be set explicitly at these levels.

See Multi-Language Articles for more information.

Tag all display formula with MathML mixed or presentation markup. Tag inline formula with MathML mixed or presentation markup when it cannot be represented by regular article elements and Unicode™ characters. MathML 2.0 and 3.0 are included in the DTDs. Each <mml:math> should have an ID.

Do not set math as <tex-math> except within <alternatives>.

Mathematical elements (<disp-formula> and <inline-formula>) should contain a single mathematical expression. That is, they should contain a single representation of the expression in plain markup, <mml:math>, or <graphic>. If the expression is being supplied in more than one format, the representations should be tagged to indicate that they are alternative representations of the same expression (see Alternate Versions of a Single Object).

The MathML 2.0 and 3.0 Specifications describe a number of requirements that are not enforced with the MathML DTD or W3C Schema. See <mml:math> for details on these requirements.

Online First (OLF) is the version of the published article that is made available in final, edited form prior to assignment of complete citation data (e.g., volume and/or issue). Other common labels for this version of an article include "early view" and "ahead of print."

See PMC's Online First Article policy and see Processing Instructions on how to add the proper indicator (<?OLF?>) for OLF articles.

OLF articles should follow PMC's minimum requirements for metadata. Specifically:

• Include all known publication dates in <pub-date>.
• Include the following information, but only if specifically applicable to the article:
  • Volume
  • Issue number
  • Pagination and/or elocation-id
  • DOI

Requirements for publication dates are based on the publishing model of the journal in which the articles are published.

There are two basic classes of publication: issue-based and article-based.

• Issue-based publication is when an entire issue is published at one time―in print, online, or both. In this class of publication, issue and article publication dates coincide, so the publication date of each article is the same as the publication date for the issue.
• Article-based publication is when articles are published individually or in small groups. They may be published in issues or collected in some other way―by volume or by year. In this class of publication, each article has two publication dates: the date (including day, month, and year) on which the article was published and the publication’s broader collection date.

For specific examples of various publishing models and their corresponding date types, see <pub-date>.

Sometimes articles are part of a series. This series may either be a group of articles all in one issue or a series of articles (like a recurring column) spanning issues. Use the <series-title> to identify the title of the series to which that article belongs.

If the XML is encoded in other than UTF-8 (or ASCII, which can pass as a subset of UTF-8), then the character encoding should be declared in the XML declaration in the prolog of each file.

Tag special characters with the Unicode hex number in character entity style (&#x2203;, ∃) or directly in the declared encoding. For accented characters that cannot be represented by a single Unicode value, use the base character and Combining Diacritical Marks (U+0300 through U+036F) or Combining Diacritical Marks for Symbols (U+20D0 through U+20E3). The base character and the accent character must be directly adjacent.

Accessibility: this guidance supports the JATS4R Recommendations for accessible capture of characters.

• To ensure that special characters can correctly be interpreted by screen readers or other text-to-speech or text-to-braille systems, take care to use the correct Unicode value for each character. For example, when indicating temperature, take care to use the degree symbol (°, &#x00B0;), and not a modifier small O (ᵒ, &#x1D52;) or the masculine ordinal symbol (º, &#x00BA;).
• Do not use <private-char>; instead, use the correct Unicode symbol for each character.

Do not use values from the Private Use Areas: U+E000 through U+F8FF, U+F0000 through U+FFFFD, and U+100000 through U+10FFFD.

Discouraged XML Characters: U+7F to U+84, U+86 to U+9F, U+FDD0 to U+FDEF, U+1FFFE to U+1FFFF, U+2FFFE to U+2FFFF, U+3FFFE to U+3FFFF, U+4FFFE to U+4FFFF, U+5FFFE to U+5FFFF, U+6FFFE to U+6FFFF, U+7FFFE to U+7FFFF, U+8FFFE to U+8FFFF, U+9FFFE to U+9FFFF, U+AFFFE to U+AFFFF, U+BFFFE to U+BFFFF, U+CFFFE to U+CFFFF, U+DFFFE to U+DFFFF, U+EFFFE to U+EFFFF, U+FFFFE to U+FFFFF, U+10FFFE to U+10FFFF.

PMC uses subjects (under <article-categories>) to sort the issue contents and build the Table of Contents. Subjects can be hierarchical. They may describe the content of the article: Physical Sciences. Or they may give an indication of the type of article: Erratum.

Each named supplement is considered an issue and may have one or many articles. Tag this information in <issue> within the <article-meta>.

See Funding Information for specifics on capturing funding sources, grants, and conflicts of interest.