TextCritical.net

There are a few ways to handle this:

• Use SGML/XML: this provides formatting and the XML structure would allow morphological parsing data to be included. Searching doesn't really work unless using an XML database.
• Use database model of chapter » verse » word: no formatting allowed but allows easy morphological parsing and seraching.

May want to redo the model such that the verses and chapters use a foreign key relation. This is necessary to make deletions of the work cascade to the verses and chapters.

See http://stackoverflow.com/questions/3937194/django-cascade-deletion-in-manytomanyrelation

I could use something like markdown in order to process the text to handle formatting, footnotes and even lemmatization. See https://github.com/trentm/python-markdown2#readme

I switched the models to use foreign keys so that the deletes cascade.

I am refactoring the models such that Section and Chapter models are replaced a single model that does both: Division.

The problem with the model that has verses foreign key to divisions is that it is difficult to find the lowest level of divisions. The lowest level of divisions are the ones that need to be used for chunking.

It may be necessary to have verses be a many-to-many field within Division objects. However, this means that verses will not be deleted when divisions are deleted.