My crusade for web content filtering

Brian Cluff brian@snaptek.com
Mon, 31 Jul 2000 23:18:11 -0700


> That's all I'm looking for.  I want to scan the data stream for key words.
> If I get a hit, I want to rewrite the page to something like:
> <html>
>   <body>
>     This page is blocked.   :P
>   </body>
> </html>
>
> This is not possible with squid, true?

Unless they have added something very recently to squid.  They only way I
know of with squid to block pages is with URLs.

BTW,  you will find that with content blocking, all you will be blocking for
the most part is the pages with dirty words on them.
If your intent is to block porno sites, you will be leaving every single
picture open to the world.  Your best bet is a hybid system that uses a list
of urls and does content filtering.  That way nobody will find any easy ways
to get around the filter.
    Plus you will need an list available so that you can block url
anonymizers.

You also might want to reconsider a server side content filter if haveing a
decent is at all important to you.  All the stuff I have done with them in
the past cause either a severe slow down in the service or at the very least
a very very annoying pause between every page and images.

Lemme know if you can't find a content based filter.  I have a squid based,
list based, filter that I wrote that I have been using with about 700
computers behind. for almost 2 years, and other than the ethernet driver
crashing on me once, it is very very stable.  The best part about it is that
it come with a fairly large list.  The only thing that I would ask is that
you contribute any sites that you add to your list, back into the master
list so that anyone else that uses the filter will gain.
    Everyone's biggest complaint seems to be that all the filters out there
have a closed list.  I figured that I could take away that complaint with
the fruits of the last 5 years of my life.  (I developed the filter that the
local company USA Safety Net based their business off of)

Brian Cluff