Language-specific spelling suggestions ('Close words' list)

[det]

When a word being looked up does not match any dictionary entry, GoldenDict often shows a list of "Close words." This is very helpful feature for misspellings. The list seems to be a set of words beginning with the typed string. In many cases additional rules for the list of close words would make this feature much more helpful.

For example, in Bengali there are vowel pairs that have the same sound: উ and ঊ, ই and ঈ. Similarly several consonants have the same sound: e.g. স, শ, and ষ are all used for the 'sh' sound. Words containing these letters are very often misspelled. (In fact it is often hard to track down the 'correct' spelling, as dictionaries and other authoritative sources often differ.) The situation is similar for Hindi, Nepali, and other languages with Sanskrit-derived scripts. In these cases, spelling suggestions that consider these interchangeable character sets would make finding a word a lot easier.

Is this something that could be added by users ...or considered for development plans? I guess this is a feature request.

[det]

No comments on this one? And any comments the 'close word' suggestions made? Would there be anything in that could be done in the .dsl dictionary to facilitate better suggestions?

[ikm]

Grab a hunspell/myspell spell-checking dictionary for Bengali, go to Edit|Sources|Morphology and enable it there, then put the resulting dictionary to your Bengali dictionary group. Problem should be solved.

[det]

After a bit of looking, I tracked down a hunspell dictionary, and that improved things greatly. Thank you!