In-Reply-To: <EC9405E7-9697-45AD-99E2-A0EDD9D755B4@gmail.com>
References: <EC9405E7-9697-45AD-99E2-A0EDD9D755B4@gmail.com>
Mime-Version: 1.0 (Apple Message framework v730)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <1FC40674-E65E-47B8-B957-719BC621BB83@mac.com>
From: Richard Todd <rwtodd@mac.com>
Subject: [lojban-beginners] Re: Hello from a new lojban learner
Date: Mon, 27 Jun 2005 07:50:43 -0500
To: lojban-beginners@lojban.org
Sender: lojban-beginners-bounce@chain.digitalkingdom.org
Errors-to: lojban-beginners-bounce@chain.digitalkingdom.org
Precedence: bulk
Reply-to: lojban-beginners@chain.digitalkingdom.org
Content-Length: 2162

On Jun 27, 2005, at 12:11 AM, Mitch Tishmack wrote:
> To elaborate on the previous example, use gerku as the meaning of  
> la chien, der Hund, and dog. This enables me to look for links from  
> gerku to other languages and easily build lists of words to ask. If  
> any fellow programmers have a better idea on how they would  
> implement this I am more than happy to listen to them.

You didn't really say what your program is trying to do, but I am  
assuming it is a flashcard type of thing, where it puts up a  
'question' word, and then gives a list of words in the other  
languages.  The user picks the 'answer' word that means the same  
thing as the 'question' word it put up?

If so, then I'd store the words as tuples:
bow  german-for-bow  french-for-bow    #  (as in bow-and-arrow)
bow  german-for-bow  french-for-bow    # (as in take a bow)
dog    ...
fish    ...
turkey  ....

So then, say your program selects the first bow tuple for the  
question.  It selects german to flash on screen, and french for the  
answer set (so that there is a right answer to choose from!).  Then  
in needs to select 3 other random incorrect answers.  To avoid the  
ambiguity problem, all the program has to do is make sure the other  
random answer tuples do not contain any of the same words as the  
question tuple.  So, to complete the example, it could randomly  
select 'dog', and check that none of the dog words equal the bow  
words.  No problem there.  Then, if it randomly selects that second  
type of 'bow', it will notice that the english component matches the  
question tuple, and would move on without selecting a word from this  
tuple.

Make sense?

Putting a lojban word in each tuple to avoid ambiguity would also  
seem to imply that the question word will always be a lojban word.   
(or each question word will always be annotated with a lojban word to  
disambiguate it?)  That may be fine for your purposes, but for the  
program I'm assuming you are trying to write, I'm not sure it makes  
any sense.  There's no reason not to go ahead and make lojban one of  
the languages it quizzes you on, of course!

In any case, good luck!