un article intéressant:
The Final Cut for 2003 Finalists
________________________________
ITW test results: Of the remaining products, those achieving 100% detection of ITW viruses
with both the on-access and on-demand scanners were:
BitDefender Professional, Frisk's F-Prot for Windows, McAfee VirusScan Home, Nod32, Norman's Virus Control,
Norton AV 2003, Panda AV Platinum, and PC-cillin 2002.
Of these, only Nod32 scored less than 90% on Zoo detection (86.97%).
Both BitDefender and Nod32 had higher false positive rates (.22% and .29% respectively).
The complete test results</A> will be published on the AV-Test.org website.
Performance and Useability: No product will be effective if users do not understand how to use it,
if certain features are absent, or if it has a negative impact on their system.
At a minimum, users should expect smooth performance as well as on-demand, on-access, and right-click scanning.
The right-click scanning should launch the scan and not simply open up the program dialog.
Most importantly, the right-click scanner should scan the file regardless of file-type or its inclusion or exclusion status in the program.
The product should be able to update easily and should allow for full customization of the types of files to be scanned.
Finally, the product should work well on a range of systems and not adversely affect system performance.
***
Narrowing down the competition
______________________________
Proving the importance of up-to-date certification and performance testing, Command AntiVirus,
a historically strong scanner, failed to make the first cut with sub-par detection rates of only 90.91% for boot sector viruses,
98.11% for macro viruses, and 96.55% for script viruses.
Other products disqualified for less than 100% ItW detection included AntiVir Personal Edition,
Avast!, Bullguard, Dr. Web, eScan 2003, F-Secure, MkS_Vir, Power AV XP, Proland Protector, QuickHeal, ViRobot Expert,
VirusBuster, Virus Utilities and V3 Pro Deluxe 2002.
Certain products passed the testing, achieving 100% detection of ItW threats,
but had glitches in their programs that could lead to infections not being removed properly,
system instability, or otherwise leave users at heightened risk. For example,
AntiVirenKit (AVK) was unable to delete or rename the VBS/Stages worm, leaving the infected file accessible on the system.
Similar problems occurred with Kaspersky Antivirus and RAV.
ZeroKnowledge, purveyors of Freedom AntiVirus had a more insidious glitch.
If a large number of infected files had been found, every file scanned thereafter would be deemed to be infected and handled accordingly.
In other words, if the action were configured to delete infected files, perfectly legitimate and clean files would be erroneosly accused and summarily deleted by the scanner.
Such a glitch could lead to a non-functioning system or other after-effects more dastardly than the original infected files intended. It, too, was disqualified.
Both AVG and eTrust AV (VET engine) performed poorly on zoo detection, at 68.22% and 78.34%, respectively.
It should be noted that eTrust AV with the CA engine performed admirably,
with one of the best zoo detection rates of 94.39%.
Despite that, eTrust AV (CA engine) was also disqualified due to numerous installation and program glitches,
one of which affected the on-access scanner in Windows XP.
Sophos was also disqualified due to a bug affecting Windows 2000/NT/XP which prevented its scanning directories on Linux/Samba servers containing an "!" in the directory name
***
The Final Cut for 2003 Finalists
________________________________
ITW test results: Of the remaining products, those achieving 100% detection of ITW viruses with both the on-access and on-demand scanners were:
BitDefender Professional, Frisk's F-Prot for Windows, McAfee VirusScan Home, Nod32, Norman's Virus Control, Norton AV 2003, Panda AV Platinum, and PC-cillin 2002.
Of these, only Nod32 scored less than 90% on Zoo detection (86.97%).
Both BitDefender and Nod32 had higher false positive rates (.22% and .29% respectively).
The complete test results</A> will be published on the AV-Test.org website.
source:
http://antivirus.about.com/cs/soft [...] 003pix.htm
