Grammar check module?

[skydesigns]

I want to prevent people from posting garbage link this in my forms:

sffstdtysah addafgaytys

Maybe it can be done simply with regex, but I haven't been able to figure it out. I guess it would be easier to define what constitutes and word, rather than what is "not" a word.

Is there a Perl module available for this sort of thing?

Thanks!

Matt Parker

 

[NEVERSLEEP]

what if the 'good' link is

[URL unfurl="true"]http://255.255.255.255/[/URL]

???
or

[URL unfurl="true"]http://www.tdtg.com/[/URL]

as u said 'Maybe it can be done simply with regex'

use regex to see if there is at least (in the link):
- twice /
- http: or ftp:

someone knowledge ends where
someone else knowledge starts

 

[skydesigns]

I'm sorry, I typed "link this" when I meant "like this" :-(

Well, it's not links that I'm concerned about. The fields that I'm having trouble with are text fields where people are suppose to put comments, describe themselves, etc. They are required fields - and I already check for min/max length and "word length", but that doesn't stop anyone from just hitting a bunch of keys, typing in some garbage and submitting the form.

This is a difficult one, because numbers and other characters are allowed in these fields. I just want to try and force people to use "real" words, rather than just a string of characters. This seemed like it may be beyond the scope of a set of regex's, and that's why I was hoping there may be a module for just this purpose.

 

[NEVERSLEEP]

even if there were a "english checker" module
this wont be very "user friendly"
people will get pissed if they make "english errors"
or "fast english" like :
r "as" are
u "as" you
etc....

someone knowledge ends where
someone else knowledge starts

 

[icrf]

You could test against a partial dictionary listing. I'm not a linux man by any stretch, but the last system I worked on had /usr/dict/words which was a simple text file of about 43000 english words (<.5MB), one per line. I actually used it as part of a password guesser to catch the people who make bad passwords to my bbs and write them nasty e-mails demanding a change. You could use such a thing to test if at least x number of words in the input area are found in such a word listing. The list could be ordered, so efficient searching options would be available. I might have that file laying around somewhere if you'd like to do something with it. Just my 2 cents. ----------------------------------------------------------------------------------
...but I'm just a C man trying to see the light