Gloss format like num:1-2-3 and fa:A-B-C were introduced with the idea of shortening annotation process and saving space.
Actually, it creates a lot of exceptions and complexities in the code, and inconsistencies in the MMS paradigm.
Finally, it makes it nearly impossible to automatically compute inflections parameters by comparing dictionary and sentence data.
It will be removed in favor of 1-row-per-digit/letter approach.
Gloss format like num:1-2-3 and fa:A-B-C were introduced with the idea of shortening annotation process and saving space.
Actually, it creates a lot of exceptions and complexities in the code, and inconsistencies in the MMS paradigm.
Finally, it makes it nearly impossible to automatically compute inflections parameters by comparing dictionary and sentence data.
It will be removed in favor of 1-row-per-digit/letter approach.