Karen Coyle writes:

>Doh! You're absolutely right, Rick. Thanks. - kc

Well, no, I don't think so.  The language is still that of the original
script.  Hebrew transliterated into Arabic is still Hebrew.  And since XML
uses Unicode you can deduce the script from the character encoding (or else
there is an explicit declaration of the character set).  What you might
want is an indication of whose transliteration rules are being used, and
you already have that in your suggested approach.

The KISS approach would probably be to just prohibit transliteration.  And
be silent about what you do when you convert an existing MARC record.