MRodz authored Jul 05, 2018 and mrodzis committed Feb 05, 2019

For the removal of tei:handShift duplicates we first enhance all tei:handShift
with all missing attributes by adding an empty attribute node - this makes the
comparison easier. Afterwards these empty nodes are removed again.

MRodz authored Jul 05, 2018 and mrodzis committed Feb 05, 2019

The code and tests for handshift-omissions are deprecated and therefore
deleted since it's easier to handle this phenomenon in the second stage of
processing.

MRodz authored Jul 04, 2018 and mrodzis committed Feb 05, 2019

This should make sorting out handShifts in the next processing step easier.

MRodz authored Jul 04, 2018 and mrodzis committed Feb 05, 2019

We noticed that the elimination of double tei:handShift is buggy and decided
that the solution of this should better be implemented in the second
serialisation step that tidies up the whole document because the hierarchy of
the XML is flatter.

MRodz authored Jul 04, 2018 and mrodzis committed Feb 05, 2019

For easier serialisation the creation of tei:div[@type = "section"] takes
places in a separate transformation.

MRodz authored Jun 29, 2018 and mrodzis committed Feb 05, 2019

After a brief discussion we decided it would be better to keep the tei:choice
construct as a whole. While we only need the text content of tei:abbr in the
serialization of the edited text, keeping tei:choice might yield better
results while searching the website.