We should look into adding some sort of mapping so that searches for language
will return matching results for either the ISO 639-1 (2 letter) or ISO 639-2
(3-letter) form of the code (and possible the ISO 3166 country codes as well?).


Received via email from Rudi Baccarne <http://www.khk.be/hikmed>
----------------------------------------------------------------------

While limiting a search on the PKP harvester by language I receive diffent
results when performing a search with the code 'en' and the code 'eng'. It
seems
that different institutions are usin different ISO639 codes. This problem might
occur in future with dutch data providers as well, because following the ISO
639
norm there are two possibilities 'dut' and 'nld' and I already noticed that
these are already use inconsisent by different institutions. Do you think there
something to do about this in the sense of mappings between different codes?

It's a pitty that people can use different codes for the dc.language I believe,
because it's lowering the possibilies for what Dublin Core was set up at the
first place.