[mkgmap-dev] Transliteration

Tue Jan 26 10:59:00 GMT 2010

Hi

Back in October I created the translit branch to add all of the
transliterations from the perl Text::Unidecode module, but
I have heard no comments on it.

I shall merge it, since it is probably better than showing question
marks even for countries where it doesn't work very well.

But there is one controversial point, in that I have included
separate tables for latin1 in some cases, that include accented
characters.  These were derived from icu4j.  However the romanisation
schemes use many accented characters outside of latin1, so I had to
drop those and the resulting mixture may not make any sense at all.
In fact I will probably drop them, unless someone says that they
are useful / or provides better tables.

There is a better transliteration program being developed by
'greencaps' on the forum and is used by Lambertus's maps.  This would
be preferred but the code is not released, although the author
says they intend to do so at some point.

So the current transliteration on trunk and in the branch
is neither language nor context dependant.  That means
that each letter is considered on its own and not in relation
to the letter(s) that preceed or follow it.

You can download a ready built jar at http://www.mkgmap.org.uk/snapshots/
(mkgmap-translit-r1314.jar).

..Steve