[ omegat-Feature Requests-1666867 ] Short string match mode

[SourceForge.net <no...@so...>]

Feature Requests item #1666867, was opened at 2007-02-23 10:48
Message generated for change (Comment added) made by leuce
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=520350&aid=1666867&group_id=68187

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: Short string match mode

Initial Comment:

This feature request was prompted by a post "[OmT] new user - translation memory & fuzzy matching question" by hud...@ya... on the Yahoogroup on 23 February 2007, about poor matches on short segments.

Different text types do better with different match policies.  For general texts, the length of the match should be roughly the length of the current segment.  For l10n purposes, however, a translator can often experience higher translation speed if he can get matches which would ordinarily fall below the fuzzy threshold, because the matches are too short.

For this reason, I propose a setting in OmegaT "Enable short matches" which will cause OmegaT to drop the match threshold and give preference to shorter matches.  This can be useful for translating GUIs and tooltips, for example.

When the translator reaches the point where he starts translating longer sentences, he can disable short matches, reload the project, and continue using the "normal" match mode.

I'm not 100% sure how the short match mode would work, but it should give preference to shorter matches even if the current segment is so much longer than the match that normally it would render the match percentage below the threshold.

(posted by Samuel/leuce, not logged in)

----------------------------------------------------------------------

Comment By: Samuel Murray (leuce)
Date: 2007-02-23 14:06

Message:
Logged In: YES 
user_id=168045
Originator: NO

Regarding "1295367", yes I have now seen it, and it in fact it gave me a
different idea.  Wordfast has a feature called "phrase search" (or
something similar) in which is searches for matching phrases.  It tries to
match a sequence of words from the current segment to TUs in the TM, and
displays the TU's source and target above the current segment.

A similar thing for OmegaT would be nice... perhaps as a separate window
(there would then be three -- match, phrase and glossary).  Then TUs with
matched phrases can be displayed.  A method to "insert" a phrase match's
target text into the current segment would also be nice (but being able to
copy and paste it might be sufficient).



----------------------------------------------------------------------

Comment By: Jean-Christophe Helary (brandelune)
Date: 2007-02-23 11:33

Message:
Logged In: YES 
user_id=915082
Originator: NO

I've already proposed a similar idea with substring matches: 1295367. Did
you take a look at it ?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=520350&aid=1666867&group_id=68187