Sunday, June 14, 2009

Twitter Litter

Yesterday's New York Times ran a piece called "Hey, Just a Minute (or Why Google Isn’t Twitter)." The article is oriented around answering the following question about how search and microblogging may be related if both depend on ranking and recognizing trends and patterns in user behavior:

Google moves faster than some of its critics think. But even if didn’t, the more important question is this: Do we really want Google’s search engine to swallow Twitter’s output as fast as it comes, without filtering, analyzing and ranking by authority?

Unfortunately, as Virtualpolitik pal Ian Bogost has discovered, Twitter and Google are already intimately connected. In a posting called "Cascading Failure" he describes the ways that Google has become a de facto authenticator of content, in ways that often render legitimate content-creators unable to access or post their own content when they are mistakenly labeled spammers or malware distributors or in Jim Zwick's case, which I write about in the Virtualpolitik book, page hijackers and copycats. Bogost describes his problem here:

Early in the morning of June 10, my web host was compromised, and a script was run across the entire server on which my site is hosted. The exploit installed hidden links, via iframes and javascript document.write commands, which redirected invisibly to malware sites. This is a relatively common way for malware attacks to begin. When users mistakenly or unknowingly navigate to a web page with exploits, new processes might be started on the host computer, which could later be used as trojan horses to distribute additional malware or viruses, or which could steal sensitive information like passwords via keystroke logging.

Google attempts to protect people from malware by using their indexing system to detect malware on sites, and to mark them as potentially dangerous. You can see this in Google search results marked "This site may harm your computer."

All the popular web browsers, including Firefox, Safari, and Chrome, rely on Google's reports of unsafe sites in its internal browsing system. When a user tries to visit such a site, these browsers display a warning message . . .

Because my site is popular enough to be frequently indexed by Google, their system had already flagged the site as a malware distributor before I was even aware that the server had been compromised. This was inconvenient to say the least, because removing the warning requires one to notify Google that the site has been cleaned, in order that Google can initiate a new check of its contents.

Take close note of this process: one must sign up for a Google account in order to be able to rescue one's site from having been marked as unsafe by Google.

Unknown to me, Twitter also uses the Google unsafe warning as a way to flag accounts as spammers or malware distributors. Before I'd even completed restoring the hacked files on my site, Twitter had suspended my account, because my profile links to this site.

With Bogost's Twitter feed suspended, he is hamstrung at a succession of upcoming Twitter-centric conferences and hampered in staging his annual June performance of "Twittering Rocks" in honor of Bloomsday and the work of James Joyce. Twitter's customer service queue entails a daunting thirty-day wait time.

Labels: ,


Post a Comment

<< Home