Tasks for future versions of the GreekOCR toolkit
-------------------------------------------------

 - optionally enable grouping for the separatistic approach

 - add new wholistic recognition which attaches accents to
   characters with Gamera's grouping algorithm

 - do thorough tests for the different approaches and
   measure the performances on various documents

 - Why does the recognition result contain spurious additional lines?

 - integrate the improved attachment of diacritical signs to
   characters into the basic OCR toolkit

 - disambiguate between quotes and accents
