ENCLOSURE 5: 

RULES OF PARTICIPATION IN SECOND CENSUS OCR SYSTEMS CONFERENCE 93/06/16

1) Participants shall not human classify or human correct any results. If
Participants human check any results as a sanity or blunder check, they
shall report this fact and any resulting changes when returning their
results. 

2) Participants shall return their classification results within two
weeks after receiving the test materials from NIST in the format and
media requested by NIST.

3) Participants shall return one filled out summary sheet per entry
with all obligatory responses completed.

4) Attendance at the Conference may be limited to Participants chosen
by the Committee based on the space available to accommodate Participants.

5) Attendance at Conference session(s) where Participants openly discuss
their procedures and results may be restricted to those Participants who
have agreed to participate openly as demonstrated by the answers that they
provide on the form illustrated in Enclosure 6 that is to be returned at
the same time as the test results. 

6) The Committee will make the final decision if any unforeseen
questions arise, and the only recourse (other than continued
participation) that is open to Participants not satisfied with any of
these decisions is withdrawal from further participation.

7) All Participant's summary scores (error vs. rejection data, etc.)
will be given to the Participants at the Conference and will be
published in the Conference results. Participant's own scores or all
scores might be sent to the Participants in advance of the Conference
at the discretion of the Committee, but if they are sent to one
Participant, they will be sent to all Participants.

8) Anyone to whom NIST sends the training materials will be listed as
participating in the Conference, and if they fail to return their test
results, this fact will be stated in the Conference report.

9) The Conference report will contain a disclaimer that is similar, if
not identical to the following:

The U.S. Bureau of the Census (Census) and the National Institute of
Standards and Technology (NIST) sponsored this Conference as part of
ongoing research into machine recognition of hand-print.  Participants
may have submitted the results of experimental or developmental
systems for scoring even if they have commercial products in the
market place, and the efforts of the participants in conducting the
tests were not proctored in any way.  While some test results from
this Conference may appear in marketing literature, potential buyers
must beware!  Census and NIST can make only one recommendation to
potential buyers: use your own application- specific data to
thoroughly test the performance of any system or component in a
realistic setting.



