fftb on an ACE-processed profile using an older version of the treebank with the
--gold flag, we observe messages in the terminal which seem informative, such as:
gold tree node 'aq0fs00' conflicts with 'aq0fs0'
Indeed, this tag name differs between the version of the grammar which was used to created the old version and the one that was used to create the new version, so this seems expected and reassuring (in that I can probably trust that discriminant).
But other times the messages seem unexpected:
gold tree node 'hd-cmp_v-cp_c' conflicts with 'hd-cmp_v-cp_c'
gold tree node 'vmif3p0@v_np-ppa*native_le' conflicts with 'ncms000@n-_c-tmp_native_le'
In the first example, it would seem the rule is the exact same one?
As for the second, it could be some random result due to somewhat ad-hoc nature of our treebanking procedure or it could be something informative which I cannot yet interpret.