Parallelizing Translation

Nathan L. Cutler livingston at
Wed Jul 5 11:30:57 PDT 2000

Apologies in advance for O-T but if you're reading this message at all you
obviously don't mind...

I would say that the contributions to this thread show that there is a
broad consensus that, yes, translation software COULD benefit, in terms of
speed, from parallelization.

But speed is not the most important factor in translations.  To be useful,
a translation must satisfy a minimum quality standard.  A poor quality
translation can do real damage in the business world, which obviously is
where these translation programs would be used.  

Having evaluated the work of many translators (Czech - English) I can say
that the worst mistakes happen when the translator fails to understand the
real meaning of the original.

But computers by their very nature do not "understand".  They merely carry
out instructions.

I don't think translation software can be useful as a replacement for human
translators in most real-world cases in which human translators are
commonly used.  Once the frequency of mistakes goes too high, the
"translation" has to be thrown out and re-done completely from scratch
because it's simply too human-labor-intensive to root out and correct all
the mistakes.

I can, however, imagine "translation assistance" software -
i.e. mainly a combination bidirectional dictionary, thesaurus and voice
recognition package that could be useful to a human translator.  But even
here, the biggest problem is going to be putting together the
dictionary/thesaurus part, not the actual computing part.  (Don't know
anything about voice recognition, but that's a different field entirely,

Nathan L. Cutler < livingston @ >         telephone: +420-2-51611648
Livingston Professional Translations		 fax:       +420-2-6514377
** When "pretty good" is not enough **           mobile:    +420-602-259964

More information about the Beowulf mailing list