Op Sa, 2006-05-27 om 13:13 -0300, skryf Bruno Sant'Anna:
...
> A list of what?
>
> a list containning objects, for example;
>
> object mistake
> {
> int startpos; // start position of the sentence
> int endpos; // end positon of the sentence
> string guessed_sentence; // the right sentence guessed by the
> grammar checker
> string rule_tip; // the grammar rule comment
> boolean checked; // flag if the user want to ignore it or not.
> }
>
Perhaps we rather want to indicate the start and end positions of the
perceived erroneous part, rather than the sentence. The grammar checker
might isolate the problem to two consecutive words with no need to
underline the whole sentence containing the problem. Suppose we have:
"Students is terrible at grammar these days."
The error really is limited to "Students is" (or arguably, "is" only).
This helps the user by showing a smaller part that must be reviewed.
Another advantage to this is that underlining a small phrase containing
the perceived error will probably give the user a hint as to how the
grammar checker parsed the sentence. Say the grammar checker gets it
wrong in
"Students is the word used for those people."
and we only underline "Students is" it might help the user to realise
that the grammar checker is the one in the wrong, because the error
context is too small. (It might be that only us programmer types are
interested in analysing the behaviour of software in this way ;-)
Furthermore, it might after all be useful to return the sentence
boundaries along with the error boundaries. This might help the user if
it indicates that sentence boundaries were not calculated correctly.
Perhaps a missing full stop, question mark, quote, or something like
that confused the grammar checker. This is perhaps only useful in an
interactive checker that can indicate the sentence as well as the error.
Furthermore, perhaps it is meaningful to allow for more than one
suggestion to be returned from the grammar checker. Suppose the mistake:
He will probably grows.
We might want to suggest:
"He will probably grow." as well as "He probably grows."
This is probably not the best example, but I imagine that there will be
situations where we might want to suggest more than one correction
because it could go either way. We might also want to associate a
different tip with each suggestion, I'm not sure.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
