From: SourceForge.net <no...@so...> - 2004-10-29 13:24:26
|
Bugs item #1056784, was opened at 2004-10-29 22:24 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=520347&aid=1056784&group_id=68187 Category: None Group: 1.4 beta Status: Open Resolution: None Priority: 9 Submitted By: Jean-Christophe Helary (brandelune) Assigned to: Sandra Jean Chua (sachachua) Summary: Match bug in Japanese source file Initial Comment: Segments that are not at all similar produce 100% matches. I created a simple 11 lines basic Japanese text, with strings close enough that they ought to produce some close matches. Source and target files are uploaded for reference. The non translated lines in the target file are the ones that produce false 100% matches: ------------------- ???????? is considered as: 1) ???????? Je suis japonais. < 100% > ------------------- ????????? is considered as: 1) ????????? Est-il japonais ? < 100% > ------------------- ????????? is considered as: 1) ????????? Est-il japonais ? < 100% > ------------------- ??????????? is considered as: 1) ??????????? J'aime les pommes. < 100% > ------------------- I used the totally translated french target file as a source file to see what the matches were and I find reasonably sane results. With word strings clearly identified as possible matches. (i played a bit with the ""assign to", "group" and "priority" settings just to see what the report would look like, don't pay attention and please modify them if I made a mistake :) ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=520347&aid=1056784&group_id=68187 |