Serialization
We need to serialize the four levels of annotations we record for a syntactically analyzed text (text, sentence, verbal unit, token). That is trivially accomplished with four delimited-text data sets, which can be coordinated through a CEX data model, and serialized to a single plain-text representation.
See examples, and compare the example tables with these visualizations of the same data.