Download PDF version of this article PDF

Taking a second look at the news so you don’t have to

Carrying the Torch of Anti-Spam Vigilantism

In a past column (September 2005) we reported on the “vigilante” anti-spam tactics adopted by Blue Security. The company’s software, named Blue Frog, fought spammers directly through tracking, warning, and ultimately flooding them with complaint e-mails if they continued spamming e-mail addresses included on the “do not intrude” list. Some saw this as essentially a DoS (denial-of-service) attack and criticized the company for its tactics. But users—more than 500,000 of them—were undeterred: Why not fight fire with fire?

Well, because you will lose. This was the lesson learned by Blue Security, which recently experienced a DoS attack by a spammer apparently fed up with the company’s anti-spam tactics. The attack brought down the company’s Web site and ultimately succeeded in shutting down the Blue Frog service. So much for fighting fire with fire.

Or maybe not. Another project, Okopipi, has risen out of Blue Frog’s ashes determined to revive and refine the Blue Frog anti-spam technology. Okopipi will be distributed, open source software that works in much the same way Blue Frog did. The difference is that the project will be based on a P2P network and won’t disclose the addresses of its servers, making it more robust than Blue Frog and therefore more resistant to retaliation—in theory, at least. Time will tell if the Blue Frog still has legs.


Taking Constrained Programming to the Next Level

As illustrated by one of this month’s special report articles, ubiquitous computing could revolutionize the way we live and work. The key enablers of ubiquitous computing technology are the tiny microprocessors and sensors that can be embedded into our physical world. But these small platforms, where power, memory, and processing are all limited, pose a challenge to programmers, who usually must rely on lower-level languages such as C to write efficient code. The coding is often tedious and the resulting software limited in complexity. In addition, code optimized for a particular platform needs to be rewritten to run on other platforms.

Fortunately, this could change. A group of Spanish researchers have created Ciao, an open source programming, analysis, and optimization toolkit for ubiquitous computing systems. Ciao uses a high-level CLP (Constraint Logic Programming) language to make programming in constrained environments easier. It also has high-level constructs that enable software written with it to be more complex. The key advantage, however, is that the code can be interoperable on multiple platforms—only when diverse embedded systems can easily interact will ubiquitous computing really take off.


The Human Intelligence Revolution

In the rush to automate, virtualize, and simulate, the value of human intelligence is sometimes forgotten. A trend to use technology to tap real human intelligence has developed, however. Services such as Google Answers, which allows users to pose questions to researchers who provide answers for a fee, and Amazon Mechanical Turk, which distributes requests for simple human intelligence tasks to a pool of workers (see “AI Gets a Brain,” by Jeff Barr and Luis Felipe Cabrera, ACM Queue, May 2006), reveal an increasing acceptance among technologists that some jobs are still best performed by humans.

The latest service that uses technology to serve up human intelligence is Illumio from Tacit Software. It works by tapping expertise within a network of users. A user anonymously poses a question to the service, which then uses desktop search products such as Microsoft’s or Google’s to scan users’ hard drives and determine who from the network would be the most qualified respondent. Despite Tacit’s guarantee of privacy, a key obstacle could be users’ willingness to install desktop search products, without which the service will not work.



Originally published in Queue vol. 4, no. 6
see this item in the ACM Digital Library


© ACM, Inc. All Rights Reserved.