You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We have had some pushback in the past for using pos as a property on Token which was not so controversial), but what people did not like was that the value is a string rather than a PosTag annotation type. With PosTag we would have had a nice way to have more than one tag or confidence measures and whatnot as well as a very specific spot to put meta data like tag sets. However, changing to PosTag would be a pain in the extreme lower back and that ship had sailed and we basically decided to live with a string value and allow other optional properties to deal with extra information (pos-list, pos-confidence, etcetera).
Now that we are about to introduce a semtags property on token we may want to open this discussion again. And perhaps this is relevant for the discussion on Morpology (issue #65).
The text was updated successfully, but these errors were encountered:
We have had some pushback in the past for using
pos
as a property on Token which was not so controversial), but what people did not like was that the value is a string rather than a PosTag annotation type. With PosTag we would have had a nice way to have more than one tag or confidence measures and whatnot as well as a very specific spot to put meta data like tag sets. However, changing to PosTag would be a pain in the extreme lower back and that ship had sailed and we basically decided to live with a string value and allow other optional properties to deal with extra information (pos-list
,pos-confidence
, etcetera).Now that we are about to introduce a
semtags
property on token we may want to open this discussion again. And perhaps this is relevant for the discussion on Morpology (issue #65).The text was updated successfully, but these errors were encountered: