Table 4.

Differences between the Auto-CORPus BioC and PMC BioC JSON outputs.

Difference	Auto-CORPus	PMC
Section titles	Section titles, subtitles, subsubtitles (and so on) are linked to the passage text they apply to	Section titles, subtitles, subsubtitles (and so on) precede the passage text they apply to
Section types	Section types are annotated using IAO terms	Section types are described using custom labels
Offset counts	Offset increased by 1 for every character (including whitespace) in a passage	Offset increased by the number of bytes in the text of a passage plus one space
Table and figure sections	Structured table data are stored in table JSON. Figure captions are included in the BioC JSON in the sequential order in which they occur within paragraphs.	Table data and figure captions occur at the end of the JSON document. Table content is given as XML.
Abbreviations section	Abbreviations section stored in abbreviations JSON. Abbreviation and definition components are related. Incomplete/one-sided definitions are not stored.	Abbreviations and definitions from the abbreviations section are stored separately as text with no relations between the two components. Incomplete/one-sided definitions are stored.
Link anchor text	Link anchor text retained (HTML element tags removed).	Link anchor text removed.
Character encoding	UTF-8 used for outputs	Available in Unicode and ASCII