1998-01-24 - Re: Why no “Banner Ad Eaters”?

Header Data

From: dlv@bwalk.dm.com (Dr.Dimitri Vulis KOTM)
To: cypherpunks@toad.com
Message Hash: 005ca5df7b5e1a822f9d6cb33968ece9652c22cebbaad385976e9fd699001146
Message ID: <5DiTJe33w165w@bwalk.dm.com>
Reply To: <199801240418.FAA17881@basement.replay.com>
UTC Datetime: 1998-01-24 18:08:45 UTC
Raw Date: Sun, 25 Jan 1998 02:08:45 +0800

Raw message

From: dlv@bwalk.dm.com (Dr.Dimitri Vulis KOTM)
Date: Sun, 25 Jan 1998 02:08:45 +0800
To: cypherpunks@toad.com
Subject: Re: Why no "Banner Ad Eaters"?
In-Reply-To: <199801240418.FAA17881@basement.replay.com>
Message-ID: <5DiTJe33w165w@bwalk.dm.com>
MIME-Version: 1.0
Content-Type: text/plain



nobody@REPLAY.COM (Anonymous) writes:
> Tim, I'm surprised. WebFree for the Macintosh does this very thing. It
> matches on URL substrings and simply removes these things wholesale from
> your browser window. No broken image boxes, nothing. It also kills cookies
> and stops animated GIFs from leaking through, if you want to nuke either of
> those things.

Of course, the person putting up the web page might well argue that you're
violating his copyright in some bizarre way.  The image on your screen with
the ad banners stripped is not what the author intended; especially if you're
helping third parties to strip ads. IANAL... thoughts?

I occasinally use the Altavista search engine in text mode
(www.altavista.digital.com) and it has ads (for amazon.com) right in the
middle of the text, in the same html file as the search results.
I suppose the next logical step in the anti-ad proxy is to edit the dynamically
edit the incoming html and to delete or replaces specified strings; perhaps
a while "intelligent agent" that will filter in only the information you've
requested, keeping out *all* the graphics and other crud.

While I was researching this subject, I came across a web page of some
student at Rutgers. It displays 4 advertizing banners and uses java/javascript
(obnoxious technologies that I disable) to make sure that you've clicked
all 4 (presumably he gets some credit for it); only after the javascript is
convinced that you've clicked the ad banners can you get in. Yeah, right.
But if ad filtering becomes more commonplace in the future, i can imagine
crypto being (mis)used to enforce ad viewing: "You no clicky my sponsor ads,
you no see my content."

I disagree with whoever said that the technology for deleting ads is the
same as the technology for deleting porn. Porn publishers very much don't
want their content to reach minors or anyone who's offended by it, and
mostly coopereate with rating agencies.  Porn isn't generally mixed with
desirable content on the same pages. You can just block the entire page
containing some porn (or, with vancouver-webpages' pics ratings you can
block pages containing "disparaging remarks about the environment", kewl).
Advertisers don't give a damn if their ads are seen by those who aren't
interested in them, and will use technology to make it harder for you to
filter out their ads (or to get to the content without seeing sponsor ads).

When MS IE 3.0 first came out, it supported pics ratings. So I got very
excited and wrote my own little pics ratings server (a better ones is
available from wc3, by the way) to filter out banner ads. Unfortunately,
after I wrote it and started testing t, I doscovered that IE 3.0 only
asks the pics server about eh top-level page; it does NOT ask the pics
server about the urls references in "IMG SRC=" and the like. I guess
the reasoning is that if the top-level page is acceptable, than anything
embedded on it is acceptable, which may be true about porn but not true
about ads.

Overall, this whole advertiser-sponsored model sucks. I'd rather be paying
(electronic cash, of course) for the resources I use.

> The URL is: http://www.falken.net/webfree/
>
> Junkbusters (http://www.junkbusters.com/) offers a personal proxy server
> kind of thing that lets you mash just about everything: HTTP headers, user
> agent info, referer URL, cookies, etc. It runs on either NT or Unix variants.

It runs udner OS/2 too!! I've got it to work under NT. it may work under
Win95 as well. (Anyone wants an executable for M$ platforms?)

> One could also use it to get around a firewall configured to reject
> non-standard browsers, but That Would Be Wrong, as Nixon once said. It lets
> you specify which sites or substrings to look for, and which ones you trust.

I've been using junkbstr and am pretty happy with it. One disadvantage of
using any proxy server, as opposed to filtering in the browser, is that the
browse is one step removed from the TCP/IP.  E.g. if you try to connect
to a non-existent domain without a proxy, the browser goes to DNS and tells
you that it can't resolve it. Going via a proxy, the proxy server goes to
DNS and sends a page back to the browser sayng it couldn't resolve it, which
isn't as clean.

Although I'm not doing it much, the following worked out fine for me:
junkbuster running on an NT box connected to the internet via ppp; browsers
running on 4 win95 boxes connected to the NT box via ethernet, using
the NT box as the proxy server.

I've exchanged e-mail with someone about the list of URL patterns to block...
I have no time for this myself, but someone ought to put up a web page
tracking what ad urls people block. That would be a nice public service.

> It also does proxy chaining.

It would be nice to have it chain automatically through more than one proxy.

Another nice little feature of junkbstr is it ability to send "wafers" -
your own cookies, which will hopefully confuse whatever tracking software
exists on the server.  I send noise with the following wafers:

NOTICE=Please send no cookies
AnonTrack=X
Apache=X
ASPSESSIONID=X
CFID=X
CFTOKEN=X
DOL=X
DTRACK=X
EGSOFT_ID=X
GeoId=X
GeoStitial=X
group_discount_cookie=T
GTUID=00.00000.0.0.0000.0000
ink=X
JEB2=X
MC1=ID=X
p_uniqid=X
PFUID=X
registered=YES
RMID=X
s_uniqid=X
session-id-time=X
session-id=X
SWID=X
UID=X
Urid=X
userid=X

---

Dr.Dimitri Vulis KOTM
Brighton Beach Boardwalk BBS, Forest Hills, N.Y.: +1-718-261-2013, 14.4Kbps






Thread