embedded copies of Unicode data?

[Paul Wise <pabs@debian.org>]

There are several copies of the Unicode data files in the archive, examples:

charmap.app: /usr/lib/GNUstep/Applications/Charmap.app/Resources/UnicodeData.txt
charmap.app: /usr/share/doc/charmap.app/UnicodeData.txt.gz
clisp: /usr/lib/clisp-2.44.1/data/UnicodeDataFull.txt
gnulib: /usr/share/gnulib/tests/uniname/UnicodeDataNames.txt
libcamomile-ocaml-data: /usr/share/doc/libcamomile-ocaml-data/UnicodeData.html
libopensrs-perl: /usr/share/opensrs/RACE/RACE/UnicodeData-Latest.txt
libsaxonb-java-doc:
/usr/share/doc/libsaxon8-java-doc/api/net/sf/saxon/codenorm/UnicodeData.html
moodle: /usr/share/moodle/lib/typo3/unidata/UnicodeData.txt
perl-modules: /usr/share/perl/5.10.0/unicore/UnicodeData.txt
pypy-dev: /usr/share/pypy-1.0/pypy/module/unicodedata/UnicodeData-3.2.0.txt
pypy-dev: /usr/share/pypy-1.0/pypy/module/unicodedata/UnicodeData-4.1.0.txt
pypy-dev: /usr/share/pypy-1.0/pypy/module/unicodedata/UnicodeData-5.0.0.txt
rails: /usr/share/doc/rails/html/classes/ActiveSupport/Multibyte/UnicodeDatabase.html
typo3-src-4.2: /usr/share/typo3/typo3_src-4.2/t3lib/unidata/UnicodeData.txt
typo3-src-4.3: /usr/share/typo3/typo3_src-4.3/t3lib/unidata/UnicodeData.txt

In addition, some of the following packages appear to embed copies of
the Unicode data in their source code:

http://source.debian.net/source/search?q=IMIFTHORON&defs=&refs=&path=&hist=
http://source.debian.net/source/search?q=LHAVIYANI&defs=&refs=&path=&hist=

In addition, there is a copy of parts of the Unicode data in
fonttools, gucharmap, libuninameslist and probably others.

I'm wondering what, if anything, Debian should do about this
situation. We generally don't like embedded code copies, but I'm not
sure if the same arguments apply to embedded data copies.

Any thoughts?

In addition, the CMap and AGLFN data are in similar situations, but I
think they are embedded in much fewer packages.

-- 
bye,
pabs

http://wiki.debian.org/PaulWise