Re: Software Translation of Lojban (was: Re: [lojban] eurolinux proposing lojban for community patent

[Richard Curnow <rpc@digitalkingdom.org>]

On Thu, Dec 13, 2001 at 12:56:00PM -0500, Invent Yourself wrote:
> 
> > > Good question. I can't answer until the code is written. Take a look at
> > > the English output from jbofi'e/cmafi'e.
> >
> 
> I say the English isn't yet quite at the level of Babelfish. It's more

Oh, you are being far too kind to jbofi'e!! 

> like a word-for-word; a non-Englishistic word order of the Lojban would
> destroy the output. The first part of this project is to write a

Yes, the English that's generated is pretty much a word-for-word
substitution.  I did it this way from the outset because I wanted the
tool to be useful for learning what the words meant.  I thought a
bridi-to-sentence translation would lose the association of which words
meant what, making the tool less useful for actually learning the word
meanings.  Plus, it would have been much harder to write.  There are a
few special cases that are handled, e.g. {le mi zdani} is converted to
{the house of me}, but these are hardly enough to make the output flow
as English.

> pre-processor that rearranges the Lojban to an Englishistic word order.

tugni

> 
> Next, the English glosses must be context-dependent.

i ji'a tugni cai

Another big issue is that currently, the sumti are tagged with a
reminder of what their place in the selbri mean.  These would need
writing as prepositions instead to get useful English out.

-- 
Richard.
---------------------------------+-------------------------------------
Richard P. Curnow                | C++: n., An octopus made by
Weston-super-Mare, UK            |  nailing extra legs on a cat.
http://www.rrbcurnow.freeuk.com/ |