Question about joining/splitting segments

[Yves Savourel <ysavourel@enlaso.com>]

I'm still working on testing how our specification can be implemented and here are another couple of questions:

=== Joining:

- A unit has two segments s1 and s2
- The segment s1 has a translation candidate, not s2
- A user agent joined the two segments together

--> what is the expected behavior for the candidate?

a) We remove because it is not a match to the new s1 anymore.

b) We keep it attached to the new s1 but we change its similarity/score attributes to some arbitrary value to indicate the match is not the same.

c) We attach the candidate to the span of text that corresponded to the content of s1 before the join, using some <mrk> annotation for this (that we don't have defined yet).

d) we do something else?


=== Splitting:

- A unit has a unique segment
- That segment has a candidate attached to it
- A user agent splits the segment into two new segments

---> what is the expected behavior for the candidate?

a) We remove it because it is not a match to either new segments.

b) We keep it in the first new segment and change its similarity/score attributes to some arbitrary value to indicate the match is not the same. Note that tool will be able to reset it if the two segments are joined.

c) We attach the candidate to the span of text that corresponded to the content of s1 before the join, using some <mrk> annotation (across two segments). Note that tools can reset it to its previous segment if the two segments are re-joined.

d) We do something else?


Cheers,
-ys