Confusing automated translators.

Francois Pinard pinard at iro.umontreal.ca
Mon Jul 14 15:20:36 CEST 2003


[Alan Kennedy]

> My favourite "phrase designed to mess up machine translation" [...] is

> "Time flies like an arrow, but fruit flies like a banana".

The "Time flies" is a classic.  I heard it a few dozens of years ago. :-)
Another similar classic, for French, is "Le pilote ferme la porte.".

> If people think it's too much newsgroup abuse to discuss such a
> non-python subject, [...]

It is Pythonic enough to me :-), as I long intended to rewrite a previous
word categoriser of mine, in Python.  The categorical ambiguities are
always a challenge, and for French at least, there are plenty of those,
especially for texts which are loose on diacritical marks.

My previous categoriser was written in MoSTex, an (unknown) language and
system implemented in Scheme, made in such a way that it was compiling
to optimised machine code.  I'm quite curious about how speedy, legible
and maintainable a Python implementation would be, by comparison.

-- 
François Pinard   http://www.iro.umontreal.ca/~pinard





More information about the Python-list mailing list