broadcast-oriented censor. Combined with the absence of scarcity in terms of the number of
channels or spectrum and the fast-dropping cost of accessing Internet from a wide range
of devices including shared terminals and mobile devices, the changes in the online environment give an edge to the online publisher against the state’s censor in the medium- to
long-run.
Along with Wikipedia, Weblogs offer a poignant example of these growing challenges for
the censor. No current filtering regime appears designed to address content developed on
blogs, podcasts, and Wikis and accessed via Really Simple Syndication (RSS) feeds in aggregators, next-generation peer-to-peer networks, BitTorrent, and so forth. The most effective
model demonstrated to date may be China’s moves in the past few years to require blog service publishers to block keywords in blog posts, though even this approach can be only a partial means of blocking subversive content over time. Chinese bloggers routinely turn to
broadly understood code words to evade the censorship built into the tools.
As online content changes very quickly and can be accessed through new means, the process of prescreening content and establishing a blockpage—akin to updating one’s static
virus definitions as new viruses are isolated and defined—breaks down. The process must
become an heuristic one to function properly, if at all. Multimedia content, which is harder to
screen and is accessed in different ways than through the World Wide Web, poses similar
challenges for filtering regimes. Those states that are intent on filtering the Internet will have
to adapt quickly if they intend to keep up. These adaptations might take the form of more aggressive filtering, or a shift to surveillance of user behavior with legal sanction for those who
receive or transmit forbidden material.
In light of the prevalence of structural-based blocking in the states we studied, the trajectory
of the Internet to a more dynamic environment will continue to create new problems for filtering regimes. The use of Weblogs by citizens—human rights activists, for instance—as a
means of self-publishing is sharply on the rise in many cultures around the world. The general
trend on the Internet is the divorcing of content from structure through the syndication of blogs
via RSS and similar technologies. Syndication allows the text of a blog to be easily reproduced
on other Web sites anywhere
34 in a way that circumvents filtering—since the retrieval of content from a blocked URL is done by the site the user is visiting, potentially located in a country
with little or no filtering, instead of by the user’s machine. While such mirroring of content has
always been possible on the Internet, syndication represents a dramatic decrease in the
amount of time and level of technical skill required to easily replicate content. In many ways,
this freeing of content from structure mirrors how large sites are internally managed. The reason that CNN can easily display the same article at multiple URLs is that the text of the article
can be retrieved from a single location, eliminating the need to separately create each HTML
page on which it displays. Through this means, the acceptance or rejection of a large site in
its entirety may in itself be a partial reaction to the problem created for URL-based structural
filtering when content is not strictly tied to location.
52
Jonathan Zittrain and John Palfrey
channels or spectrum and the fast-dropping cost of accessing Internet from a wide range
of devices including shared terminals and mobile devices, the changes in the online environment give an edge to the online publisher against the state’s censor in the medium- to
long-run.
Along with Wikipedia, Weblogs offer a poignant example of these growing challenges for
the censor. No current filtering regime appears designed to address content developed on
blogs, podcasts, and Wikis and accessed via Really Simple Syndication (RSS) feeds in aggregators, next-generation peer-to-peer networks, BitTorrent, and so forth. The most effective
model demonstrated to date may be China’s moves in the past few years to require blog service publishers to block keywords in blog posts, though even this approach can be only a partial means of blocking subversive content over time. Chinese bloggers routinely turn to
broadly understood code words to evade the censorship built into the tools.
As online content changes very quickly and can be accessed through new means, the process of prescreening content and establishing a blockpage—akin to updating one’s static
virus definitions as new viruses are isolated and defined—breaks down. The process must
become an heuristic one to function properly, if at all. Multimedia content, which is harder to
screen and is accessed in different ways than through the World Wide Web, poses similar
challenges for filtering regimes. Those states that are intent on filtering the Internet will have
to adapt quickly if they intend to keep up. These adaptations might take the form of more aggressive filtering, or a shift to surveillance of user behavior with legal sanction for those who
receive or transmit forbidden material.
In light of the prevalence of structural-based blocking in the states we studied, the trajectory
of the Internet to a more dynamic environment will continue to create new problems for filtering regimes. The use of Weblogs by citizens—human rights activists, for instance—as a
means of self-publishing is sharply on the rise in many cultures around the world. The general
trend on the Internet is the divorcing of content from structure through the syndication of blogs
via RSS and similar technologies. Syndication allows the text of a blog to be easily reproduced
on other Web sites anywhere
34 in a way that circumvents filtering—since the retrieval of content from a blocked URL is done by the site the user is visiting, potentially located in a country
with little or no filtering, instead of by the user’s machine. While such mirroring of content has
always been possible on the Internet, syndication represents a dramatic decrease in the
amount of time and level of technical skill required to easily replicate content. In many ways,
this freeing of content from structure mirrors how large sites are internally managed. The reason that CNN can easily display the same article at multiple URLs is that the text of the article
can be retrieved from a single location, eliminating the need to separately create each HTML
page on which it displays. Through this means, the acceptance or rejection of a large site in
its entirety may in itself be a partial reaction to the problem created for URL-based structural
filtering when content is not strictly tied to location.
52
Jonathan Zittrain and John Palfrey