Princeton University computer scientist Edward Felten has a knack for poking holes in technology.

Last week, the tech industry hit back and, in the process, transformed Felten into a symbol of the fight for academic freedom. Bowing to legal threats, Felten and seven colleagues announced they would pull their research paper detailing how they broke audio watermarking security on digital music. "We chose not to fight the battle at this time, but we are determined to find a way to continue fighting and eventually publish our paper," says the 38-year-old associate professor.

At issue is the 1998 Digital Millennium Copyright Act, which the industry insists prohibits Felten and his colleagues from publishing their work. The law makes it illegal to circumvent copyright protection technologies.

This isn't the first time Felten has taken on powerful adversaries. In 1996, he and his students found security flaws in Java. Three years later, he and another student team humiliated Microsoft when they discovered that the company had doctored a videotaped software demonstration it used as evidence.

Felten's latest tech tangle began in September after the Secure Digital Music Initiative, an industry group, threw down a challenge to the code-cracking community to break SDMI's watermarks. Felten put together a team made up mostly of academic researchers and found an answer.

Officials at the SDMI Foundation disputed Felten's claim; the foundation's beef was that Felten's code degraded the audio quality. But that didn't keep it from joining the Recording Industry Association of America last week in threatening to sue Felten's team if it presented its work at an academic conference.

Felten says he was interested in whether watermarking technologies - supposedly inaudible additions to digital recordings - could protect music copyrights in the digital age. "Today I think they are not viable," he says, conceding audio watermarking may improve because of research like that done by his team.

Team member and Rice University professor Dan Wallach describes Felten, his former Ph.D. adviser, as "fearless." But Wallach says he and his colleagues agreed to withdraw the paper for fear of losing their homes in litigation. Still, Felten vows to find a way to publish his code-cracking research. "I suppose I'm not so worried about who is angry at me," he says.

Keith Perine contributed to this report.