Corpora: US company claims patent on common machine translation

Jacques Guy (j.guy@trl.telstra.com.au)
Sun, 12 Oct 1997 08:25:45 -0700

A post appeared the other day on sci.lang and soc.culture.esperanto,
the import of which ought to be of utmost concern to recipients of this
list.
In short, some tiny US company is claiming patent on a very common
approach to machine translation. Here is the text of the patent, as I
found
it on soc.culture.esperanto:

A universal machine method implemented on a data processing device
utilizing a data base for performing interpretations between any one
of a plurality of first national source languages and into any one of
a plurality of second national target languages comprising:

1. a first step of storing a source text in said any one first
national language in an electronic storage means as a first text file

2. said first step including a first substep that parses said source
text into individual sentences, with each sentence in its own separate
subfile

3. a second step of interpreting said sentences into a created
internationally universal language which serves as an intermediate
pathway, as the word making up said sentences are translated into the
created internationally universal language, they are entered into
individual files consisting of the word plus a grammatical syntax tag,
such tag consisting of a designation of the word's grammatical
posture, i.e., noun, pronoun, verb, adverb . . . etc., and

4. a third step of interpreting said tagged words from said created
internationally universal language into any one of said second natural
national target languages; said interpretation steps being totally
reversible within said intermediate pathway of said universal
international language, there being no requirement of reworking of
said intermediate pathway language to accommodate changes in source or
target languages.

--------------------------end of
quote------------------------------------

This outfit, which goes by the name of Toltran, has hired a law firm
to try and enforce this patent.

I quote their lawyer's letter verbatim (posted on sci.lang and
soc.culture.esperanto):

Dear Dr. Martin

It has come to my attention that your website is offering a
multi-lingual dictionary programme that goes by the name "ERGANE".
It uses the artificial language Esperanto as an auxiliary language.

I direct your attention to the intellectual property right owned by
my client TOLTRAN LTD., U.S. Patent No. 4,864,503, a copy of which is
enclosed herewith for your study.

Since there appears to be an encroachment of Toltran's patent by
ERGANE, you are hereby advised to immediately cease and desist and
remove this
material from the offering you are making in your website.

Failure by you to (1) immediately remove this material from your
website, (2) publish acknowledgement of your error, in your website, and
(3)
advising the persons who have down-loaded this material from you that
the use
thereof will infringe our patent and to cease using and destroy the
downloaded
material; can result in further action by my client.

Your immediate response is requested.

Sincerely yours,
Laff, Whitesel, Conte & Saret, Ltd.
J.R. Halvorsen, Of Counsel

---------------------------end of
quote----------------------------------------------

It is evident, at least to me, that this charade must be stopped
forthwith.
Meantime, I decided to test Toltran, and e-mailed them this:

Subject: US company tries to prohibit freeware dictionary programme
Date: Fri, 10 Oct 1997 15:07:09 -0700
From: Jacques Guy <j.guy@trl.telstra.com.au>
To: rkt@wwa.com
CC: gvwilgen@worldonline.nl

That was the subject line of a post on sci.lang and
soc.culture.esperanto, go find it.

I suffer fools lightly, for they provide me with
amusement. You, sirs, amuse me. Now that I know about
Toltran and Travlan, and having suggested to the
Travlan people an exceedingly simple solution to
eschew your swinish threats, I shall give them every bit
of cooperation to ensure that their product so outshines
yours that... good-bye Toltran. After all, haven't
I been in computational linguistics for 20 years
now? How dull life would be without fools!

------------------------------end
quote-------------------------------------

Here is the highly literate answer from Toltran:

Subject: Re: US company tries to prohibit freeware dictionary programme
Date: Fri, 10 Oct 1997 06:38:35 -0700
From: Richard Tolin <rkt@wwa.com>
To: j.guy@trl.telstra.com.au

LISTEN YOU STUPID FUCKING IDIOT, YOU CAN SUCK
MY ASS FOR ALL I CARE YOU BENT SON OF A BITCH.
YOUR LIFE MUST BE SO FUCKED.YOU LITTLE TWISTED SHIT.
LIKE MY ENGLISH ! ASSHOLE.

--------------------------end quote-----------------------------

You can see the calibre, size, and level of the staff.
Lawyers do not seem very particular about the
clients that hire them. As I wrote to Toltran, I did
suggest an exceedingly simple solution to the Travlan
people (I am sure you can all guess what it is). I
also suggested that they claim a patent there. As
they distribute ERGANE for free, no harm done.

However, I urge you all to ponder this matter very
seriously. I, for one, am not in the mood to let
this pass unchallenged.

Jacques B.M. Guy
Telstra Research Laboratories
770 Blackburn Road
Clayton 3168 Australia