An example file, data.gml, is included in the Samples directory.

Source language is expressed using GlossML's srclang attribute. Target
language is, implicitly, the non-source language. If there is more than
one non-source language, the file is not convertible.

Subject field is expressed in a subjectField attribute of the glossary
element. This attribute is imported from another namespace, as seen in
the sample file. Subject field is mandatory for convertibility.

A glossary-wide note is expressed as in all GlossML.

Source and target terms are expressed as in all GlossML. Each glossentry
must contain one langentry each for the source and target languages,
and no other langentries; otherwise the glossentry is not convertible.

Source and target part of speech are expressed in a partOfSpeech attribute
of the term element. This attribute is imported from another namespace,
as seen in the sample file. Part-of-speech attributes are mandatory
for convertibility, but if the attribute is present on only one term
of a source/target pair, it will be presumed identical for the other
one. The convertible part-of-speech values are adjective, adverb, noun,
properNoun, and verb -- properNoun is an addition to the list of values
in the ISOcat data category.

Note on an entry, definition, and source of definition are expressed as
in all GlossML. Per the GlossML specification, the language of the note
should be the srclang of the document.

Contextual example is expressed in an example attribute of the term
element. This attribute is imported from another namespace, as seen in
the sample file.

Source of contextual example is expressed in a source attribute of the
term element. This attribute is imported from another namespace, as
seen in the sample file. In its original namespace, this data category
is multipurpose, but in convertible GlossML it is only used to cite
sources of examples.

No data category may take a value containing a tab character; such values
are not convertible.

