adblock.txt Standard
By Bill Dimm
April 20, 2001
Motivation
Most content sites have "Terms of Use" statements that prohibit
"modifying" or "creating derivative works" from the site content.
Much like Napster, ad-blocking software acts as a tool allowing
people to violate content owners' rights and deprive them of income.
Ad-blocking software should face restrictions similar to those
imposed upon Napster in the courts. Ad-blocking software
should not block ads displayed on sites which have communicated
that blocking is prohibited. This proposal specifies an automated
mechanism for sites to communicate that ad blocking is prohibited.
Proposal
Web site owners should communicate their position on the use
of ad-blocking software on their site by providing an adblock.txt
file in their top directory in much the same way that they
provide a robots.txt file to indicate what parts of the
site search engine spiders and other robots are permitted to access.
The format and meaning of the contents of adblock.txt are
the same as for robots.txt as described in the
Standard for Robot Exclusion
except that where robots.txt refers to "User-Agent"
the relevant item for adblock.txt is the name of
the ad-blocking program. If no adblock.txt file
is supplied by the web site, the ad-blocking software is
free to block ads on that site.
Here is an example adblock.txt.
Use of this format allows adblock.txt a great deal
of flexibility in specifying permitted ad-blocking on a site.
This proposed standard is, in fact, probably far more flexible
(and complicated) than is really necessary. It would probably
be quite sufficient (and a huge improvement over the current
situation) if web site owners could simply specify that all
ad-blocking is either permitted or prohibited for an entire site.
There are two reasons for proposing this more complicated
standard for adblock.txt. First, the format is one
that most webmasters are already familiar with from robots.txt.
Second, some site owners may be generally willing to permit
ad-blocking on their site while wanting to prohibit certain
very abusive (e.g. software that replaces ads with something else)
or aggressive (e.g. software that modifies and can break JavaScript
code) programs. Ad-blocking software authors who do not
want to implement a parser for such a complicated adblock.txt
can follow the much simpler and more conservative procedure
of turning-off ad-blocking if an adblock.txt file
is present on the site (regardless of its contents).
Implementation
Most ad-blocking software seems to use a custom proxy server
to filter HTTP requests based on the URL of the requested object.
Once the software has determined that the request was for
something it wants to block (and is thus an advertisement rather
than an entire web page), it can determine the web page that
the advertisement belongs to from the Referrer field in the
HTTP header. Once this is determined the corresponding
adblock.txt can be requested if it has not recently
been checked.
Comment or vote on this proposal.
| 
