Major changes in the Version 2.0 of the RDB edition of the CSJ.
Addition of discourse segment table
- A segment table named "segDiscourse" is newly provided. This table covers from the beginning to the end of a discourse.
- Parent-child relationship tables that define the relation between segDiscourse table and all other segment tables are also provided.
Change in the name of column
- The column formerly named "OrthographicTranscription" is renamed as "Text". This column is involved in many segment tables.
Changes Unaligned-segment tables
- Tables formerly called subsegment tables are renamed as "unaligned-segment table". Former "subsegSUW" and "subsegLUW" are now renamed as "usegSUWMorph" and "usegLUWMorph" respectively.
- Parent-child relationship tables that define the relationships between the two unaligned-segment tables and their parent segments are provided. Also a parent-child relationship table is provided that defines the parent-child relationship between the two unaligned-segment tables.
- In the subsegment tables of version 1.0, the same unit IDs ("LUWID" and "SUWID") as in the corresponding segment tables. In version 2.0, however, new unit IDs called "LUWMorphID" and "SUWMorphID" are introduced. In addition to this, the "nth" and "len" columns that were involved in the former subsegment tables were omitted, because the information provided by these columns can be obtained by use of the newly provided parent-child relationship tables.
- A column named "Dep_ObligateComment" is newly provided to the link table "linkDepBunsetsu". This column contains obligatory comments of the bunsetsu-dependency annotation.