NODE 0bbc2f13referers and W3 (fwd)
Rich Graves <llurch@networking.stanford.edu>Sun, 18 May 1997 14:59:30 +0800
A friend who webmasters a large site that is implementing referer-specific
content sent me this when I mentioned the cpunks/cryptography thread of a
few months back. I basically agree with the W3C. While user education on the
potential privacy threat is essential, I do not believe that Netscape should
violate published technical standards. There are also privacy and property
issues from the server's perspective.
---------- Forwarded message ----------
Subject: referers and W3
The following text was added to W3's page on HTML on May 8. If nobody
else has noticed this, you might wish to share it with the cypherpunks, I
don't know.
HTML, Hypertext Links and Referer URLs
In a vacuum, semantically neutral technology attracts meaning. If
neutral technology is useful, commerce -- abhorring a vacuum --
uses technology in often-unforeseen ways to build new markets.
Over the past year we have seen attempts to create meaning for
HTML references where none existed before. Links have been argued
to mean revenue streams, approval, a source of liability,
endorsement, and ownership.
The Web depends on hypertext links between resources. Links can
act as references, or to embed material, both of which are
concepts that have been around for a long long time. Whats new
with the Web is the ability to follow links quickly.
When a user clicks on a link, the URL of the page containing the
link is passed to the server along with the requested URL. This
gives Web sites the opportunity to customize the response
according to the page the user clicked on.
If someone jumps into a Web site without entering via the home
page, a content provider may choose to turn them away or to
redirect them to the home page. A smarter Web site could use
partial string matches on the Referer and Requested URLs to
customize the page it returns.
One reason for doing this is to ensure visitors see the banner ads
and promotions, a site is using to create revenue; another is to
avoid broken links when people (or indexing engines) bookmark
transitory Web pages.
The Referer URL can also be used to determine how to honor
requests for images or applets. For instance, you might decide to
only provide images to sites which have signed a license agreement
with you.
http://www.w3.org/pub/WWW/MarkUp/
NODE 9314de33Re: referers and W3 (fwd)
Tom Weinstein <tomw@netscape.com>Mon, 19 May 1997 07:55:22 +0800
Rich Graves wrote:
>
> A friend who webmasters a large site that is implementing
> referer-specific content sent me this when I mentioned the
> cpunks/cryptography thread of a few months back. I basically agree
> with the W3C. While user education on the potential privacy threat is
> essential, I do not believe that Netscape should violate published
> technical standards. There are also privacy and property issues from
> the server's perspective.
Just in case someone hasn't heard about this yet, we support a stealth
preference in 4.0 to turn of referer headers. If you wish to disable
referer, you can place the following line in your prefs.js file:
pref("network.sendRefererHeader", false);
There is no UI checkbox for this preference because it was added after
we froze the UI.
--
You should only break rules of style if you can | Tom Weinstein
coherently explain what you gain by so doing. | tomw@netscape.com
NODE 381c7377Re: referers and W3 (fwd)
Rich Graves <llurch@networking.stanford.edu>Mon, 19 May 1997 08:26:31 +0800
Yes, I'm aware of that. I don't think it's a good idea. If you don't agree
with the standard, work within the W3C to change it. Especially if it's an
undocumented cypherpunk-only feature, that bugs me.
-rich
On Sun, 18 May 1997, Tom Weinstein wrote:
> Rich Graves wrote:
> >
> > A friend who webmasters a large site that is implementing
> > referer-specific content sent me this when I mentioned the
> > cpunks/cryptography thread of a few months back. I basically agree
> > with the W3C. While user education on the potential privacy threat is
> > essential, I do not believe that Netscape should violate published
> > technical standards. There are also privacy and property issues from
> > the server's perspective.
>
> Just in case someone hasn't heard about this yet, we support a stealth
> preference in 4.0 to turn of referer headers. If you wish to disable
> referer, you can place the following line in your prefs.js file:
>
> pref("network.sendRefererHeader", false);
>
> There is no UI checkbox for this preference because it was added after
> we froze the UI.
>
> --
> You should only break rules of style if you can | Tom Weinstein
> coherently explain what you gain by so doing. | tomw@netscape.com