* Known problems and future improvements

- There are numerous errors in the annotation resulting in
  inconsistencies.

- A number of tokens have an extra pos-tag (mor_pos) or an extra
  msd-tag (syn_msd) which need to be manually adjudicated. In many of
  these cases there may be a difference in category between the
  ordinary pos-label and msd-label.

- Inconsistent treatment of end-of-line hyphenation in tokens.

- Sources are only provided in the form "scrubbed" texts, not any
  html/pdf files.

- Annotation guidelines missing.
