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 cannot use numbers, like in anon, because referencing an edge is not enough. Reference needs to be to a specific label on an edge. (If att in the example were mispelled, there would also be an O tag, and we wouldn't know if C relates to S-CON or O.)
Related problem: "secondary edges", i.e. edges between edges.
The text was updated successfully, but these errors were encountered:
(Using "labeling" to refer to a specific label on a specific edge.)
Referencing a specific labeling directly is difficult. Edges ids are "volatile" (they often change) so a reference like e-s23-t23/S-CON would have to be updated a lot, and it would be easy to fall behind.
Could we add indices to labels? A labeling would consist of a label and possibly an index, something like labels: ['S-CON:1', 'O'] and labels: ['C:1']. We would need a UI widget for adding and removing indices from labelings.
Should it be possible to add more than one index to a single labeling? labels: ['C:1:2']
Indices should probably be essay-wide, not sentence-wide. Otherwise, if one later decides to re-segment two sentences to one, indices might clash. Also better for searching in the future.
Tror att dessa referenser bör kunna göras även till element utanför meningen med C-markeringen. Det är till exempel tänkbart att man ändrar ett ord från singular till plural i en mening, och att det leder till att man behöver göra en konsekvensändring av en anafor i efterföljande mening.
In this example, the C relates to the S-CON.
We cannot use numbers, like in anon, because referencing an edge is not enough. Reference needs to be to a specific label on an edge. (If att in the example were mispelled, there would also be an O tag, and we wouldn't know if C relates to S-CON or O.)
Related problem: "secondary edges", i.e. edges between edges.
The text was updated successfully, but these errors were encountered: