<div dir="ltr"><div><div>G'day,<br><br></div>Mike and I noticed (after a hint from Matic) that one of the rules in jaen did not fire in ACE in the same way it did with LOGON.<br><br></div>in jaen/erg.tdl there is a rule to get rid of any left over marks:<br><br>all_mark_ditch_ef := elision_mtr &<br>[ INPUT.RELS < [ PRED "~mark$" ] > ].<br clear="all"><div><div><div><br></div><div>This should match any predicate ending with "mark" --- but does not. When we rewrote it to: <br><br>all_mark_ditch_ef := elision_mtr &<br>[ INPUT.RELS < [ PRED "~_mark" ] > ].<br clear="all"><div><div><div><br></div><div>it matched most of the things we want, but will over match (including, e.g., "_mark_n_1").<br><br></div><div>Is this a principled difference in the regular expression handling? Should we be writing it in another way?<br></div></div></div></div><div><br></div><div>Yours, <br></div><div><br>-- <br><div class="m_6281574704490005525gmail_signature">Francis Bond <<a href="http://www3.ntu.edu.sg/home/fcbond/" target="_blank">http://www3.ntu.edu.sg/home/<wbr>fcbond/</a>><br>Division of Linguistics and Multilingual Studies<br>Nanyang Technological University<br></div> </div></div></div></div>