inconsistencies with output on new CLAN
Brian MacWhinney
macw at mac.com
Tue Aug 17 20:05:44 UTC 2004
Dear Michelle,
There have indeed been big changes to the sample exercises that will
mean that you will need to run through these lessons afresh. When
testing this out now for your class, it is important to use the online
(electronic, PDF) version of the manual. The hard copy was printed in
2000 and it is now out of date in many ways. In late 2002 and through
2003, we went through big changes involved in the transition to XML.
To correct for the idiosyncratic use of main line segmentation, we
focused energy on the construction of an automatic tagger for English
called MOR. That system has now been applied to all of the corpora in
the English segments of the database with the exception of about 4 UK
corpora that are still in progress. All of the corpora in the database
now correspond with the CHAT guidelines in the electronic manual and
corpora.
Having said this, I think I need to double check now each of the
commands and problems you reported.
This will take me an hour or two. I will tell you the results soon.
--Brian
More information about the Chibolts
mailing list