ONI Testing Results
Episodic filtering, in addition to routine blocking of Web sites considered blasphemous
or threatening Pakistan’s internal security, continued through the end of Musharraf’s
term as president and into the tenure of the new civilian government. In December
2007, the enactment of the cyber crimes ordinance was followed shortly by the
reported blocking of several hundred anti-government blogs. 60 In late December
2008, upon the recommendation of the IMC, the PTA issued an order to block six
URLs: three from dictatorshipwatch.com, a Web site created after Musharraf declared
a state of emergency; one from makepakistanbetter.com, a social and political discussion forum; one from friendskorner.com, another discussion forum; and one from
buzzvines.com. 61 The common thread among the Web pages appeared to be information about Punjab Governor Salman Taseer, 62 rather than any type of content under
the authorized purview of the IMC.
OpenNet Initiative testing was conducted on Cybernet, LINKdotNET, the PTCL, and
Micronet in May through July 2008. There is no uniform method of blocking among
the four ISPs tested. Cybernet blocks by setting a DNS lookup failure; both LINKdotNET
and Micronet employ block pages—LINKdotNET by means of Squid Proxy and Micronet using both proxy server and Squid Proxy. The PTCL had appeared to rely on DNS
poisoning by redirecting to an IP address they own containing no content, but during
later rounds of testing switched to a block page using a proxy server.
A comparison of ONI testing results from 2006–2007 and 2008 indicates that despite
high-profile filtering incidents, ISPs may actually be blocking less content. For example, in contrast to testing results from 2006 to 2007, where the PTCL implemented a
limited block on pornography and religious conversion sites, 2008 testing found no
evidence of filtering in these categories.
A form of collateral filtering, the blocking of additional content that is unintended
and caused by imprecise filtering methods, 63 has long been a feature of Internet censorship in Pakistan. For example, in March 2007, in an attempt to comply with a
Supreme Court order to filter blasphemous content ‘‘at all costs,’’ the PTCL implemented a blanket IP address block at their Karachi PIE exchange that lasted for
four days and impacted the Akamai servers, leading to disruptions in accessing Google,
Yahoo, BBC, CNN, ESPN, and several other major Web sites. 64 In January 2008,
several bloggers reported the blocking of the Blogspot.com and Wordpress.com
domains. 65
One of the most severe examples of collateral filtering took place in February 2008,
when a government attempt to block YouTube in Pakistan made the entire Web site
inaccessible to most Internet users around the world for up to two hours. 66 On February 22, the PTA issued an order to block access to a single video, while listing three
IP addresses. 67 The film Fitna by the Dutch parliamentarian Geert Wilders, which
Pakistan
495
Episodic filtering, in addition to routine blocking of Web sites considered blasphemous
or threatening Pakistan’s internal security, continued through the end of Musharraf’s
term as president and into the tenure of the new civilian government. In December
2007, the enactment of the cyber crimes ordinance was followed shortly by the
reported blocking of several hundred anti-government blogs. 60 In late December
2008, upon the recommendation of the IMC, the PTA issued an order to block six
URLs: three from dictatorshipwatch.com, a Web site created after Musharraf declared
a state of emergency; one from makepakistanbetter.com, a social and political discussion forum; one from friendskorner.com, another discussion forum; and one from
buzzvines.com. 61 The common thread among the Web pages appeared to be information about Punjab Governor Salman Taseer, 62 rather than any type of content under
the authorized purview of the IMC.
OpenNet Initiative testing was conducted on Cybernet, LINKdotNET, the PTCL, and
Micronet in May through July 2008. There is no uniform method of blocking among
the four ISPs tested. Cybernet blocks by setting a DNS lookup failure; both LINKdotNET
and Micronet employ block pages—LINKdotNET by means of Squid Proxy and Micronet using both proxy server and Squid Proxy. The PTCL had appeared to rely on DNS
poisoning by redirecting to an IP address they own containing no content, but during
later rounds of testing switched to a block page using a proxy server.
A comparison of ONI testing results from 2006–2007 and 2008 indicates that despite
high-profile filtering incidents, ISPs may actually be blocking less content. For example, in contrast to testing results from 2006 to 2007, where the PTCL implemented a
limited block on pornography and religious conversion sites, 2008 testing found no
evidence of filtering in these categories.
A form of collateral filtering, the blocking of additional content that is unintended
and caused by imprecise filtering methods, 63 has long been a feature of Internet censorship in Pakistan. For example, in March 2007, in an attempt to comply with a
Supreme Court order to filter blasphemous content ‘‘at all costs,’’ the PTCL implemented a blanket IP address block at their Karachi PIE exchange that lasted for
four days and impacted the Akamai servers, leading to disruptions in accessing Google,
Yahoo, BBC, CNN, ESPN, and several other major Web sites. 64 In January 2008,
several bloggers reported the blocking of the Blogspot.com and Wordpress.com
domains. 65
One of the most severe examples of collateral filtering took place in February 2008,
when a government attempt to block YouTube in Pakistan made the entire Web site
inaccessible to most Internet users around the world for up to two hours. 66 On February 22, the PTA issued an order to block access to a single video, while listing three
IP addresses. 67 The film Fitna by the Dutch parliamentarian Geert Wilders, which
Pakistan
495