There has been much discussion lately regarding variations in the forms of word as the context (gender and number) change. It has been pointed out that not all words follow all of the rules. For example, coys is not the plural of coy. I do not fail to see the amount of work that would be needed to do what I am about to propose. This is a serious undertaking, but should not be refused because of the required work. I suggest that the dictionary files are modified so that they contain flags for allowed parts of speech. Each part of speech would have a rule set defined for each language. Coy would be marked with a set of flags that would be read as "only to be used as an adjective." It would take on modifications for adjectives but not the plural endings and other changes used for nouns (coy-s).
As of yet, my grammar checker does very little. As I have said before, I intend to build my grammar checker based on two principles: (1) word order, and (2) verb conjugation. At present, my program can only conjugate the verbs. I would appreciate feedback on or patches to the code (or feedback on the format of the language files). Most plans I've seen for other grammar checkers also depend on word orders. Certain types of phrases are constructed using set rules. The checker would break down the sentence into clauses and (if possible) phrases using punctuation and check whether the sentence is constructed correctly with the rules set out for each phrase type.


Regards,
Luke Myers
Project Lead, ConjuGNU Grammar Checking
http://conjugnu.sourceforge.net/download/nightly/4.0/

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



Reply via email to