← Back to team overview

tamilspellchecker team mailing list archive

Hunspell spell checking and development

 

i am summarizing our(Amachu,Santhosh thottilingam,sivaji and me) discussion
at MIT a few weeks before.

Santhosh thottilingam was happy about our project and our word collection.

He insited on developing an existing spell cheker -especially Hunspell-for
local language support.Hunspell is internally used by Open Office ,firefox
and many others.

He had tried spell checking  for malayalam with Hunspell,but found many bugs
in Hunspell which is now being fixed by the Hunspell developer.(And other
interesting issue,he found  is strcmp(),isletter() of glibc have bugs when
processing  local languages(tamil,malayalam and ...)   )

So, we need to explore Hunspell.Basically hunspell must contain some
configuration file which is used to decide wrong word.That file has to be
updated for our local language support.


1)Finally,if we can develop Hunspell ,Tamil spell checking will reach all
applications quite easily.
2)We need to refine our word collection ,4.8 lakh words  as it contains some
miss spelled words.

Your suggestions are welcome.


-- 
Yours,
S.Selvam