each from can be derived from the other

["Dillon, Chris" <c.dillon@xxxxxxxxx>]

I am Co-Chair of the Translation & Transliteration of Contact Information PDP 
WG, but making this comment in a private capacity.



It is a solid and practical report and obviously the result of a huge amount of 
work.



I would like to make some points about Table 4 on p.11:

*         I feel it is important to stress that Table 4 is an ideal of clean 
data, that currently may only be produced manually and that it contains aspects 
of both transliteration (e.g. 千代田 -> Chiyoda) and translation (e.g. ビル -> 
Bldg.).

*         Moreover, the relationships between the original and the transformed 
records are complex and it is not possible to move automatically in either 
direction e.g.“first” is usually 第一 /daiichi/ but in this case is ファースト 
/faasuto/ from the English, where // represents a transliteration. In fact 
there are other Japanese possibilities, but I’ll omit those in the interests of 
brevity. Unfortunately ファースト may be either “first” or “fast”. (One wonders 
whether the Japanese fast food chain ファースト・キッチン /faasuto kitchin/, also 
advertised in Japan as the translated form, “First Kitchen” may have originally 
been a mistranslation of “Fast Kitchen”.)

*         This may affect text such as “original data could have been in either 
form” and, especially, “each from can be derived from the other”.

Scripts where letters can be read in more than one way or which do not use 
spaces to define word boundaries (Japanese falls into both of these categories) 
will be the most resistant to automated transliteration/translation.



Chris Dillon.
--
Research Associate in Linguistic Computing, Centre for Digital Humanities, UCL, 
Gower St, London WC1E 6BT Tel +44 20 7679 1599 (int 31599) 
www.ucl.ac.uk/dis/people/chrisdillon<http://www.ucl.ac.uk/dis/people/chrisdillon>