I have used Eric's Cookie Jar proxy to filter out cookies (it's located at http://www.lne.com/ericm/cookie_jar/ ) and modified it to replace advertisement images with a small jpeg image of a tin of spam. (Cookie Jar's behavior is to simply close the connection which results in a "broken" icon in my browser. The user is unable to discern the difference between a broken picture vs. a piece of intentionally ignored spam.) I did it by returning the spam can instead of an advertising banner when requested, but I probably could have accomplished the same thing by modifying the incoming HTML to have an IMG SRC of
Junkbuster, which I've been playing with, too has just a list of URL patterns that are blocked. When it encounters such an URL, it displays this hardcoded spiel: char CBLOCK[] = "HTTP/1.0 202 Request for blocked URL\n" "Pragma: no-cache\n" "Last-Modified: Thu Jul 31, 1997 07:42:22 pm GMT\n" "Expires: Thu Jul 31, 1997 07:42:22 pm GMT\n" "Content-Type: text/html\n\n" "<html>\n" "<head>\n" "<title>Internet Junkbuster: Request for blocked URL</title>\n" "</head>\n" WHITEBG "<center>" "<a href=http://internet.junkbuster.com/ij-blocked-url?%s+%s+%s>" BANNER "</a>" "</center>" "</body>\n" "</html>\n" ; Notice something unpleasant!!! See the 2 %s's? They are filled with 1) the URL you tried to access, 2) the pattern from the blockfile that matched. But they're not just displayed in your browser; only if you click on the junkbuster logo, they are passed to a cgi script at the junkbusters site, which displays them and possibly logs them. Not nice at all. I'd much prefer a filtering proxy that allowed different actions for different patterns in the block file: at the very least, the choices should be "404 not found", "403 verbotten", a 1x1 jpg, and a user-specified replacement text/url.
http://localhost/spam.jpg. I was really trying to write code to eliminate the entire spam anchor, but found that identifying the matching start and end of the spam anchor (in PERL) was more work than it was worth, and was breaking some HTML.
The ability to edit the incoming text and pictures to delete embedded ads would be awfully nice. Consider for example the html dynamically generated by wired's cgi script when you access http://stocks.hotbox.com. (take a look at it). Wouldn't it be nice if the proxy could be told to delete or replace all the text between <!-- ad strip --> and <!-- end ad strip -->??? Occasionally an ad is embedded in a picture together with some useful information. Take a look at http://www.bloomberg.com/gifs/iview.gif. it has both useful information that should be displayed and sponsor informaion that should be mercilessly censored - replaced by a user-specified graphic or at least a black rectangle.
Anyway, yes, keep posting your lists of evil URLs. I have no moral problem with trashing those unwanted ads.
Instead of spamming the mailing list, I've put up a web page at http://www.geocities.com/CapeCanaveral/Hangar/6354 (let's see if they pull the plug on it)
Would there be any interest if I put an anootated netscape cookie file (cypherpunks cookie potluck) someplace like geocitie.com?
I have also placed some cookies for sharing on them same web site.
Also a composite blockfile for junkbusters and other filtering proxies.
A really neat feature for ad-filtering proxies would be to be able to place a specified picture (or at least a black rectangle) over a jpg or gif file being received from a URL that matches a specified pattern. I'm not sufficiently handy with graphics file formats to tell how easy this would be.
Death to the Usenet Cabal! --- Dr.Dimitri Vulis KOTM Brighton Beach Boardwalk BBS, Forest Hills, N.Y.: +1-718-261-2013, 14.4Kbps