Profanity filter in chat application help!

[starblood]

I'm running an ASP chat application on the site I administer and it has a profanity filter that works along the the line of:

msgLine=Replace(msgline, "bugger", "b****er&quot
msgLine=Replace(msgline, "swine", "sw**ne&quot

.......but obviously with worse words than these

So as users get wise and start using "b.ugger" and "bu_gger" I will have to add new lines to the code to ban those words. What I'd like to do is kick them out if they swear - a solution has been suggested by a colleague:

if msgLine<>Replace(msgline, &quot;bugger&quot;, &quot;b****er&quot then
msgLine=Replace(msgline, &quot;bugger&quot;, &quot;b****er&quot
Response.Redirect &quot;swear.asp&quot;

and this works but is going to prove unweidly and cumbersome if I have to use it for every bad word. Anyone got any suggestions on how to do it more elegantly? Treat me as an idiot when it comes to explaining as I am not a programmer by a long way!

 

[mylesorme]

Trouble is computers are even dumber than any human, so we have to tell them what a profanity is and then what to do about it. There are some dictionaries of swear words, but you look as if you are on the right tracks.

There are more sophisticated approaches that are way beyond the scope of asp and databases - check out automomy for inspiration.

For a simple implementation alng the lines you are using check out WebWizGuide.com - they provide an open source forum that has this facility.

 

[starblood]

Thanks Mylesorme, much appreciated!

 

[Salibas]

When it come to string search and replace, the command

replace

is not efficient at all!

Use the magical object

RegExp

. I have made a chat area that accepts smilies of any kind, like or

here how it is used:

 set regEx = new RegExp
 regEx.Global = true     'make search to ALL matches
 regEx.Pattern = "b\Wu\Wg\Wg\We\Wr"
 msgLine = regEx.Replace (strString,"b****r") 'make replacement
 
'to test match only
 if reqEx.test (strString) then response.redirect "swear.asp"

focus on the Pattern part. you must make a combination that covers all possible cases that you expect people will use. Like b.ugger and b_ugger ....

Salibas

http://www25.brinkster.com/salibas

 

[starblood]

Thanks Salibas, I'm onto it!

I've actually managed to cobble together a bad word database by adapting a guestbook script at WebWizGuide.com, as recommended by Mylesorme (amazed myself as I am not a programmer as previously stated and most of my dealings in code have been HTML and Javascript)

Anyway, if anyone wants to see the chat application in question, it is here:

http://www.s-cheshire.ac.uk/chattrap/login.asp

logon: guest
password: guest

It is quiet at the moment, students are on holiday!