Obvious corrections done in version 1.1:

  - most of the obvious corrections concern the inexact identification
    of STIME and ETIME boundaries; usually, a missing final dot "." at
    the right boundary.  Example:

      RISE version (cmu.andrew.official.cmu-news-2477:0):
        Extruders," will be given at <stime>10:45 a.m</stime>., Tuesday, Feb. 21, in <location>Doherty

      Correction:
        Extruders," will be given at <stime>10:45 a.m.</stime>, Tuesday, Feb. 21, in <location>Doherty

  - other obvious corrections concern the inexact identification of
    LOCATION boundaries (usually the fact that the closing
    parenthesis is not included):

      RISE version (cmu.andrew.org.epp-58:0):
        <location>Scaife Auditorium (Room 125</location>)

      Correction:
        <location>Scaife Auditorium (Room 125)</location>

  - no substantial modification to SENTENCE and PARAGRAPH tagging; we
    only changed examples such as the following where a SENTENCE tag
    crosses an STIME one.

      RISE version (cmu.cs.robotics-1245:0):
        at <stime>7PM</stime> in <location>Wean 4623</sentence></location>.

      Correction:
        at <stime>7PM</stime> in <location>Wean 4623</location></sentence>.

    BTW, it seems that SENTENCE and PARAGRAPH tags have been usually
    skipped before running the IE algorithms so we are thinking of
    removing them in future versions of the corpus.