The CNTR collation provides several improvements not normally found in other collations:
- Displays orthographical differences, scribal corrections, and the condition of characters.
- Highlights variant patterns and boundaries.
- Includes interlinear information so that the average user can get an idea of what the words mean.
- Created directly from the raw manuscript data to help eliminate errors.
The collation was computer-generated by using three different algorithms:
- The maximum text was created as a templet containing all known variants for each verse by using a recursive longest common sequence first algorithm without reference to any base text.
- Each individual witness was then aligned to this templet using a non-recursive longest common sequence algorithm considering multiple sequences. Each distinct lexical/morphological/phonological word form was aligned according to a phonetical set of rules.
- The resulting matrix was compressed using a gap-reducing algorithm to minimize the number of columns.
The collation currently only contains extant manuscript data from up to 400AD along with some modern critical texts. This beta version is still a work in progress and therefore remains restricted by copyright, but when it is ready for production it will be released under an open license. For more details, consult the CNTR Technical Reference.