[kde-edu]: kvoctrain: fundamental change in concept of vocabulary structure

Marc Heyvaert marc_heyvaert at yahoo.com
Mon Nov 15 20:00:16 CET 2004


Kvoctrain has one 'original' language and one or more
tranlated languages. I wonder if it would be possible
to get rid of this restriction, so that the original
language could be 'set' at will.

I know that in principle the following does not hold
between languages : A translates into B, A translates
into C so B translates into C. This is not generally
true, but it is in a lot of particular cases. There
are a lot of specialized vocabularies that could be
build in such a way the the file containes the
lexicology in 3,4, 5, even more languages and if we
only can learn from the original language into one of
the others this would mean that we have to duplicate
the files...

Another example is the relationship between an
original language, say English and Chinese. Chinese
can be written in simplified characters, traditional
characters, pinyin, wade-giles, etc. One should be
able to learn from any 'language' into any other. How
can we do this with kvoctrain today except by
splitting the file?

Incidentally this example shows another flaw. As far
as I can see we can only define 'real' languages,
based either on KDE or iso. I thought of giving
'esperanto' here as an example of a language that was
unavailable, but it is :) But one could imagine
dialects and other specialized languages that are not
defined, and I can't see (yet?) what a possible
workaround would be. 

If we want to make 'setting' the original language
optional (this could be saved at file level), there
will be an impact on the 'levels' because the number
of levels per language will grow exponentially with
each language that is added. But perhaps we can
discuss a restriction in the number of languages, or
test to see at what level the combo of number of
entries/lessons and number of languages becomes



Do you Yahoo!? 
Check out the new Yahoo! Front Page. 

More information about the kde-edu mailing list