A jebga...tako je to, brand je pola profita :-)
A na te "testove" u časopisima nemoj se previše obazirat...rade ih ljudi koji doslovno _nemaju pojma_ o onome što rade. Pola članaka su ionako prevodi tekstova sa interneta, a ostatak vrvi sa toliko gluposti da je to strašno.....
Jednostavno su se ti časopisi uspostavili kao neka vrsta autoriteta, pa većina ljudi koja ni ne zna puno sve što tamo pročitaju uzima zdravo za gotovo. Tako je, velim, Norton užasno popularan u USA, ponajviše zbog
plaćenih reklama i testova. Pogle recimo ovo:
http://www.nod32-si.com/awards/cnet_zdnet.htm
http://www.nod32-si.com/awards/zdnet_trash.htm
Također bih se osvrnuo na ovu izjavu:
http://www.virusbtn.com/issues/virusbulletin/2004/10.pdf
Citat:
The results are often published of antivirus product tests that have no scientific basis, have seriously flawed premises, or demonstrably incorrect methodology. Worse, some tests have all of these faults.
Unfortunately, this is often true of tests conducted by popular computer magazines, whose reviewers seem to think that testing against ``a few samples we collected from the Internet and our email'' constitutes a valid test of an
antivirus product's detection capabilities. Sadly, it has also been true of some of the tests conducted by recognised AV testing bodies for publication by such magazines -- whose interpretation of the results has been wildly misguided.
Reviews in consumer magazines are far more widely read than those in antivirus industry journals and, in general, their readers are not qualified to understand the (in)significance of the test results. As a consequence, the public's confidence in antivirus products is at stake.
A classic example, circa 2000, is that of CNET's testing utilising the infamous Rosenthal Utilities (RU). The Rosenthal Utilities generated benign (i.e. nonviral) files, the data part of which contained a portion of a virus. The executable part displayed a message on the screen. It should have been obvious that the detection of any RU-generated file constituted a false positive. However, CNET rated the products in the test according to their detection of the RU files. Two years later, CNET was not only still making the same mistake [1], but compounding it by altering real viruses such as VBS/Loveletter (i.e. creating new viruses) and trashing products that `failed' their tests. Eventually, partly due to pressure from the industry [2, 3], CNET phased out the use of RU test files.
A vjeruj mi da su BUG-ovci i PC-Chipovci 100 puta nestručniji od CNET-ovaca, ma koliki ovi bili debili.
Konkretno što se tiče Nortona, izolirao sam snippet od 20 instrukcija koje ništa konkretno ne radi, a dotični ih detektira kao "virus". Samo ih okači u .text/.code sekciju, ne moraju se uopće izvoditi...i eto ti "virusa" :)