A weekly look at recent science and technology research conducted at Princeton.
Can Computers Get fAsTer?
DNA-powered molecular computers are slowly but surely winding their way up the computational ladder. EEB researchers Michael Livstone, Danny van Noort and professor Laura Landweber explained what advances have been made and where the technology is going in an article published last month.
The first computational problem solved using DNA was performed in 1994. The problem was: "Given a set of cities, a starting point, an ending point, and the (one-way) routes connecting them, is it possible to visit every city exactly once?" The problem was solved with DNA by representing cities as unique 20-nucleotide sequences of single-stranded DNA and by representing the connections between cities as complementary sequences that could attach to the end of the starting city's nucleotide sequence and the beginning of the ending city's nucleotide sequence.
By mixing the cities and paths together, DNA sequences formed that represented all possible paths through the cities. In one fell swoop, the solutions present themselves. In molecular computers, all of the "processing" takes place in parallel. A conventional computer would go through each possible route — serially — until an appropriate route is found.
Conventional DNA-analysis techniques, including PCR and gel-isolation, were then used to pick out the DNA strands that began and ended with the nucleotide sequences for the starting and ending cities of the path and that were 140 base pairs (7 cities) in length. Those sequences can then be examined to determine all possible paths that satisfy the problem.
There are practical limits to molecular computing. A 200-city version of the same problem would require an amount of DNA more than the mass of the Earth, the authors reported.
Trends in Biotechnology, "Molecular computing revisited: A Moore's Law?" March, 2003.
Digital rights
Digital rights management (DRM) technologies are systems that control how copyrighted works can be used. A DRM system might, for instance, allow you to play a music file on your computer — a legitimate use of the music — but prevent you from sharing it with others — possibly an illegal use of the music. The music industry would like its customers to use DRM-controlled computers and music players so that no one would be able to violate copyright law.
Computer science professor Ed Felten argued in an article published this month that DRM systems may not ever be able to fairly enforce copyright law. The problem for DRM, he wrote, is the gray area of "fair use." The catch is that only a judge can determine whether an instance of copying is covered by fair use, and the best a DRM system could do is try to predict what a judge would rule. It would be a "judge on a chip," he wrote.
Felten further argued that no computerized system available today could reasonably determine whether something was or was not fair use. Understanding the circumstances and evaluating the effect on the market for the original work — which need to be considered — are well beyond the current limits of artificial intelligence.
As a result, technology is no substitute for legal enforcement of the rights of both copyright holders and end users. Communications of the ACM, "A skeptical view of DRM and fair use." April 2003.
Untied tubes
If you've been sterilized (hey, you never know) but have since changed your mind, take comfort in knowing that nearly 2 percent of women and over 2 percent of men obtain a reversal after sterilization in Quebec. James Trussel and Allison Hedley GS, director of and graduate student in, respectively, the University's Office of Population Research, along with a researcher from Quebec, studied the last 19 years of payment records in Quebec's state-run health insurance system. They looked at the statistics for men undergoing vasectomies and women undergoing sterilizations.
Roughly 20 percent of men and women who had reversed a sterilization then obtained a subsequent vasectomy or sterilization.
Obstetrics & Gynecology, "Sterilization failure, sterilization reversal, and pregnancy after sterilization reversal in Quebec." April, 2003. — Princetonian Staff






