On the API approach mozilla can pass an XML document with the text to be
checked and the procedure to perform.
The middleware or the mozilla spell checker when it arrives can take the
text and perform the procedure (spell, thesaurus, grammer etc.)
It then passes back an xml page with word offset, word count,
alternative solutions and type of error (spell, grammer etc.)
A small dumy spell checker sits and catches this and passes back an
empty page and the spell cheker does nothing.
Someone writes an adapter that fits M$ spell or some web based
spellchecker, and it can replace the dummy.
Mozilla evenualy comes up with a solution and pluggs it in place of the
dummy.
So we have a way to get spell checking done soon but not total
availability untill later.
Asa Dotzler wrote:
> nospam@nospam wrote:
>
>> All mozilla needs is an API for the spell checker.
>
>
>
>> Then a small piece of middleware can be written to link it to any
>> available spell checkers.
>> Then in time the mozilla team can write a great spell checker and use
>> the existing API.
>> That way no single platform spell checker code has to be written into
>> mozilla but a rout to quick fixes exists in the interim.
>
>
>
> Your "quick fixes" suggestion wouldn't happen to come with a patch,
> would it?
>
> --Asa
>
