Re: Multilingual Wiktionary?

Salve Yann,

Am Mittwoch, 18. Februar 2004 22:33 schrieb Yann Forget:
> When will a multilingual wiktionary be put in place?
> This was already requested several times for German and French languages.

before starting other languages I would prefer to think about how a system 
would be more perfect than now. The existing Wiktionary would many make sence 
to share < 10.000 vocabulary, but when we want  to work with >> 250.000 words 
it could be much better than today.

I found today a German project to creat a dictionary of the German languages 
of the 20th century:
http://www.dwds.de/
http://www.dwds.de/pages/pages_woebu/dwds_woebu.htm

I have doubts that a WikiWiki like in the Wikipedia would be the best system 
for creating a dictionary with several people:

If you looking into the wikipedia, for example Europe, than do you already 
know > 25% of this article - if someone has manipulate this article, you 
probably will reconice this manipulation if it was only one part, the chance 
for you is 25%, for the next visior > 40% and for the third > 50%...

One of my wish for the wikipedia is a statistic how many registert/unregistert 
users has visit this articel after last change.

But when you lookup the wikitionary, you know nearly nothing about a word, 
especialy when you translate it from your mother language into another (for 
example I have used "blinds" instead of "blind persons". How can I trust 
wikitionary?

Before one articel become excellent in the wikipedia 1->15 people has worked 
on it. Word to Word translations could be right or wrong, so mainly only on 
person has this word on his viewlist.

When I would contribute to a free dictionary, I would work more systematic, I 
would take a field of my study field and create a list of 100-50.000 words.
- To these world I would add a short explanation and example sentence.
- Then I would sort them into subfields
- I would try to scan  and make statisticaly research how often this word is 
used-
-This list should be checkt by several persons

When there is an english, netherlands or frensh list, too, I would try to link 
these world with the explanations and a contexexampel to one or more word of 
my mother language.
This linking should be done by several people and in both direction, 
seperatly. So for every worldfield > 10 people would do this linkink, than we 
could do a statisticaly analyse again, which translation is clear, and where 
not everybody choose the same and start to discuss these words.

This algorith is far away from a wiki concept for creating webpages and I like 
to hear other ideas and opinions. BTW does there exist literature about how 
to create a dictionary?

So Yann, are you totaly happy with the Wikitionary system, or does you have 
ideas how it could be done better?

Looking forward to read your ideas, arguments - greetings
rob

PS: I don`t think the wikitionary programmer did a bad job, but for building a 
good dictionary (and better than already existing ones like dict.leo.org) we 
should take the chance now to brood (or search) new ideas.