Naming and versioning scheme

1. The Reuters Corpus is divided into a number of volumes.
2. Each volume is identified by a volume number and a description, e.g. Reuters Corpus Volume 1 (English Language, 1996-08-20 to 1997-08-19).
3. If changes to the format take place and a volume is reissued then the format version number will be incremented. The format version is an integer which is set at one for a new volume.
4. If changes to content take place and a volume is reissued then the correction level number will be incremented. The correction level is an integer which is set at zero for a new volume.
5. The correction level and format version are independent and neither is ever reset for any particular volume.
6. Each volume has a release date which relates to the total content of that volume.
   

If any one volume requires multiple CDs (or other media) then each CD will carry the same volume identification, description, format version, correction level and release date and each CD of the set will be distinguished by designation as "Disk x of y".