[sword-devel] The poor man's interlinear
dfhmch at googlemail.com
Mon Sep 10 02:18:01 MST 2012
Further update ....
I referred Jonathan's reply to my friend in MissionAssist, with the
following accompanying remarks.
Somehow, I think he's missed the main point.
i.e. You already have a legacy to Unicode conversion, yet because of the
complexities of the original documents and how it the legacy script works,
the purpose of the required comparison is to discover any "corner cases"
which the conversion algorithm didn't yet address.
Is this a right understanding of the task?
My friend observes:
Your correspondent has not understood what we are trying to do. We are
looking for mistakes in the algorithm.
We have already converted the encoding. It is a given that the source file
is error-free as displayed. What we have to check is that our encoding
conversion mapping (algorithm if you like) is accurate. Machine methods of
doing this have so far turned out to be circular in methodology. The only
sure, but slow, method is sight checking. Until we find other means we would
like to be able to view the text on lines stacked one above the other.
Viewing Word files side-by-side at 75% zoom is hard on the eyes when the
fonts in question are not standard Latin. Given that our source files always
use customised TTF fonts we would need tools that operate on plain text to
have import filters to handle RTF/DOC/DOCX/PDF as a minimum.
Many thanks for your interest in our problem.
View this message in context: http://sword-dev.350566.n4.nabble.com/The-poor-man-s-interlinear-tp4650950p4650961.html
Sent from the SWORD Dev mailing list archive at Nabble.com.
More information about the sword-devel