I'd like to ask you to add some way to tell Hunspell that two characters (or character combinations) are equivalent.
For example, in Catalan we have the letter "ŀl". It is usually written as "l·l" (using a middle dot), but now with unicode there is the "ŀ" character which should be used for it (although certain operating systems provide no convenient way to write it) but as the "l·l" way is more widespread the Catalan dictionary contains the words written that way, and if you write them with "ŀl" OpenOffice.org and other applications mark them as incorrect, which I find highly annoying.
Unless I'm missing something, fixing this would require duplicating all entries containing "l·l", listing them also with "ŀl", so my request is for Hunspell to provide some way to tell it that "ŀl" and "l·l" are equivalent, so that it would accept words written using "ŀl" even if they are with "l·l" in the dictionary, and in the word suggestions it would show only one version (in this case "l·l", or perhaps detect which of both character was used most in the document and use that one.. or if the error is not the "ŀl" being missing, show the form that the user has already typed).
This may also be useful for other languages, eg. German, where "ue" is accepted as an alternative way to write "ü".