[Project-ideas] gsoc 2012 project idea---A Glossary Tool to index all available localizable strings for Bengali across various FOSS projects and, other available corpus

[Subhobrata Dey]

Well,i did study the input strings for translation tools like lokalize.It
uses a glossary(i think this is the one that this project aims to build) to
translate the input string into the language mentioned. It uses
paragraph-by-paragraph translation.What i found out is that lokalize can
directly substitute a particular part of the input sentence in native
language.Are you talking about these approaches taken by common translation
tools like lokalizeor something else??If so,i feel my approach can work
better.Or else the lokalize algorithm can be used.Can you please elaborate??
Thanking you,
subhobrata

On Fri, Mar 23, 2012 at 10:12 PM, Runa Bhattacharjee <runabh at gmail.com>wrote:

> On শুক্রবার 23 মার্চ 2012 02:50 অপরাহ্ণ, Subhobrata Dey wrote:
>
>  In this way,there are these many (or may be some more) variants of the
>> verb
>> 'come'. So, if the virtual translator can identify the subject of the
>> verb &
>>  the tense,then it can use the glossary to correctly identify the variant
>> of
>>  the verb for the given sentence & translate it. The glossary will become
>> self-learned now as well.Whenever it encounters an unknown word,it'll
>> first
>> detect if it is a verb/noun.If verb,then create the variances accordingly
>> &
>> store them.If noun,then just find the bengali term for that & store them.
>>
>
> Have you taken a look at strings/content in the files and documents that
> are
> used for technical translation?
>
> regards
> Runa
>
> --
> blog: http://arrbee.wordpress.com
> irc: arrbee or runa_b on Freenode
> http://fedoraproject.org/wiki/**User:Runab<http://fedoraproject.org/wiki/User:Runab>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ankur.org.in/pipermail/project-ideas-ankur.org.in/attachments/20120324/7d01c244/attachment-0003.htm>