first-stage collation processing in the Frankenstein Variorum Project. For post processing and Variorum development, see our GitHub organization: https://github.com/FrankensteinVariorum
If run over prior collated text, will re-generate other problems (old spurious collation alignments w/ 1831)
Attempt to stop/reduce this: screen some stop words from the algorithm? "the"
Revisit Schematron checking outputs
Can XSLT help with recombining patterns of spurious alignment?
Perhaps still best NOT to re-run whole collation process on the first previously collated sections, but address later in the pipeline, so as not to interfere with previous corrections.