The Times: Tech column - VeriSign mischief/spellcheck hell
The internet was stolen last week. A faceless multi- billion-dollar corporation hijacked millions of web addresses for its own financial benefit, forcing ordinary surfers to visit its web pages, from where they are lured towards those of its commercial partners. As if this audacious land-grab were not scandal enough, the hijacking was done in a way that makes it harder to stop spammers sending you junk e-mails from nonexistent web addresses. It may also make it harder to keep your web-surfing habits from companies that will pay to target you.
The corporation responsible for this outrage is called VeriSign, and it happens to run the biggest registries of internet domain names, those ending in .com and .net. Two weeks ago this column mentioned with concern that VeriSign was considering redirecting any misspelled web addresses with these suffixes to pages on its own servers. Suddenly last Monday it went ahead with its plan, which redirects anyone who mis-types a .com or .net address to a VeriSign page which recommends (paid-for) options.
"Typo-squatting", as the practice is being called, has been turned into a huge commercial opportunity. VeriSign says that 20 million web addresses are mistyped every day, so there is clearly money to be made. The $4 billion company argues that the business, called SiteFinder, will simply improve "the user web- browsing experience", but internet service providers and web administrators are fuming. As well as breaching all sorts of web technical standards, they say, it makes it harder to stop spam and more difficult to know when you have mistyped someone's e-mail address.
There are also privacy concerns: a "web bug" has been found in the SiteFinder source code, which places a cookie in your computer that lasts for five years.
Though VeriSign says that it does not harvest personal information, privacy consultants say that it could, in theory, use data that it already holds about millions of people and match it against their surfing habits.
Smaller websites are threatening to sue the company for anti-competitive practices, and web administrators are designing "patches" that can bypass the system. But Icann, the body supposedly in charge of domain names, is merely asking VeriSign to "suspend" SiteFinder voluntarily until these issues can be addressed.
That is not enough: the US Government must end the practice now to stop the web becoming one company's commercial sideline.
++++
A FASCINATING linguistic theory has been zapping around the internet over the past week. "Aoccdrnig to rseearch at an Elingsh uinervtisy," thousands of websites are reporting, "it deosn't mttaer waht oredr the ltteers in a wrod are in; the olny iprmoatnt tihng is taht the frist and lsat ltteers be in the rghit pclae.
"The rset can be a toatl mses and you can sitll raed it wouthit a porbelm. Tihs is bcuseae we do not raed ervey lteter by itslef, but rthaer the wrod as a wlohe." Isn't taht itnerseting?
Some versions locate the study more specifically at the University of Cmabrigde - sorry, Cambridge. Yet one small detail bothers me: I cannot find any evidence of the original survey, conducted at Cambridge or any university.
Back in 1976, a PhD student at Nottingham explored randomised letter orderings, and in 1999 Nature published an American paper dazzlingly entitled "Cognitive restoration of reversed speech". But it is impossible to locate the source of the current memo (one of those "idea viruses" that the internet propagates at vast speed). Is it a hoax? All we can prove scientifically is that a plausible verbal trick will jump around the world in minutes. Even one that sends The Times's spellcheckers into meltdown.
(The Times, September 30 2003)
Read more!




