Knocking Out Norton AV Early

Dwight Silverman on his forthcoming look at antivirus applications: “I am not going to focus on their malware-stopping effectiveness, but rather how well they get along with other programs, user-friendliness and their effects on system performance.”

I predict a last place finish for Norton AntiVirus.

On a related note, I am a long-time Consumer Reports subscriber.  I don’t have any strong feelings one way or another about creating viruses to test anti-virus applications, but most of them claim to be able to identify new viruses by their behavior.  And how else could one test that claim?

UPDATE: Randy Abrams answers my question in a Comment and in this post.

Tags: ,

0 responses to “Knocking Out Norton AV Early

  1. Here is how you test the effectiveness of an anti-virus product’s ability to detect unknown viruses. It is called retrospective testing. You stop undating scanners for a period of time, say three months. At the end of that time you scan samples of viruses that have been discovered in the past three months. Now you are testing how products actually did against the real threats that users are really encountering. Consumer Reports really missed the boat by testing in the manner they did. For more information on retrospective testing I recommend http://www.mcafee.com/common/media/vil/pdf/imuttik_VB_conf_2001.pdfThe professional anti-virus testers at av-test.org and av-comparatives.org recognize the scientific superiority and accuracy of retrospective testing and use it.Cars, refrigerators, stoves, yeah, I’ll look at Consumer Reports. Anti-Virus Software? No way. during the seven years I worked at Microsoft in charge of preventing the release of infected software I never once saw a magazine do an anti-virus software review of any quality unless they went to an outside test organization for all of the important testing.