[Tetum-translators] Re: Natural language translation Project

[Peter Gossner]

Hey bro !
On Sun, 28 Mar 2004 02:45:33 +1030  from a terminal far far
away<cromwell/>  wrote:
>Dear Peter, very interesting, i'm almost tempted to say most
>interesting, this is very good stuff indeed, still going throught it
>but it seems to have at least cracked the tip of the iceberg. 
The NLTK is going to be really useful. If nothing else it shows just how
hard this is !
(I downloaded the data files last night .... don't yet. they are huge.)

There are some things I don't like and I reckon we can go faster than
they do, if only because we can target our processor time...
e.g.s
-  faking a class concept seems silly why . probably a teaching
thing.
-why not tag in proper xml and be done with it. 
- needs a gtk interface for real end users ?
(less important though)
- I am sure we can do a better Database engine. (relational etc)

The demos are great though and the concepts are generic enough for us or
any other translator I think.. I will try to adopt them as I understand
them :)


i.e. I think I will just use the NLTK as a
template and rewrite.


> I'm not sure
>what part you guys want me to play in this but it's time to start
>defining roles because i'm beginning to spread myself too thinly to
>pile much more on the plate.  Despite myself and my self inflicted
>schedule, this project really appeals, i'm not quite sure what i can do
>for it but i'd like to play some part for sure. 

Mate. Just having your smiling face about makes me want to continue :)
Pick something your interested in and claim it !

Lev is our linguist and I suspect will be a great coder.
You should be a great coder as well, especially at the abstract stuff ?
Me I am good for grinding blindly away a guff. (scoping etc) and
managing the project wide interfaces.
(web sites mailing lists etc...aka admin)

> You're latest diagram
>made a lot more sense to me also, thanks for the work on that.

Cool. Lucky Lev caught me before I started the API design.
I will have more stuff this week.

Would you like to be on the tetum list ?


>
>Take care cobber, best regards to the clan,
>
>cromwell



-- Pete  (blows me out how often these are appropriate)
Todays fortune:
<Pahan> exarkun: Are you a brain surgeon?
<exarkun> Pahan: I know where your brain is, and I've used a knife
before, if that's what you mean.     
< http://www.gnu.org/software/tetum/ >
< http://bigbutton.com.au/~gossner >
< address@hidden >

>
>
>On Thu, Mar 25, 2004 at 03:15:21AM +1030, Peter Gossner wrote:
>> Updates to this:
>> 
>> the wget string should have   " -np " added to it else you get the
>> entire site :)
>> 
>> you can get particular sections like so:
>> 
>> (all in one line )
>> wget -k -c -r -np --reject pdf,ps,zip
>> http://nltk.sourceforge.net/tutorial/tagging/index.html
>> 
>> This site is very current (this year) and still being written.
>> 
>> --pete
>> 
>> Guys,
>> (esp crommers !)
>> 
>> I tripped over a one stop shop for the "Spirit" project 
>> (it's even in Python)
>> 
>> http://nltk.sourceforge.net/
>> 
>> also check out the open Natural language Group (bottom of page)
>> I had found most of these quite independently (address@hidden)
>> 
>> has extensive tutorials etc ..
>> 
>>  wget -k -c -r --reject pdf,ps,zip
>> http://nltk.sourceforge.net/tutorial/introduction/index.html
>> 
>> Will get you LOTS of stuff in one hit (enter as one line)
>> 
>> This should help at least as a common info / approach / base 
>> 
>> I am downloading all the code now and have the beginnings of the
>> tutorials etc.
>> 
>> 
>> -- Pete
>> Mesage Composed: Wed Mar 24 16:10:34 UTC 2004
>> Calendar events:
>> Mar 25       Aretha Franklin is born in Detroit, 1943
>> Mar 25       Anniversary Day (Otago)
>> Mar 25       Triangle Shirt Waist Fire, 1911
>> Mar 26       Emerson, Lake, and Palmer record "Pictures at an
>Exhibition"> live, 1971 Mar 26         Popeye statue unveiled, Crystal City TX
>Spinach> Festival, 1937     
>> < http://www.gnu.org/software/tetum/ >
>> < http://bigbutton.com.au/~gossner >
>> < address@hidden >
>>