As testing for this comparative got underway, something of a milestone in the history of the Windows XP platform was reached – on 30 June, most versions of the operating system ceased to be sold via most OEM and retail channels. Licensing will continue to be available for ‘System Builders’ until January 2009, and in April official support for the platform will be downgraded to an ‘extended’ period set to continue until 2014. These first steps towards putting the platform out to pasture seem somewhat premature, given its continuing popularity and massive market penetration.

With its slicker, more advanced successor Windows Vista now well past its launch stage and settled in as the default (and in many cases only available) operating system for new PCs, Windows XP has maintained its dominance as the platform of choice for the majority of PC users. Looking at a selection of studies of platform usage, XP’s figures are declining very slowly, currently estimated as being in use on around 75% of systems while Vista has crept up to 15%. Many businesses continue to run XP on their workers’ desktops, even where this entails removing Vista from new purchases. At this rate, XP looks set still to be the most widely used Windows version when the next new release, the successor to Vista currently going by the title ‘Windows 7’, hits the shelves – currently scheduled for around two years’ time.

Adding further to the longevity of XP is the latest service pack, released a few months ago and added to the Automatic Update system during July. The update contains a number of new features, many of which are related to security, authentication and encryption, but for the majority of users is expected to make little obvious impact. In the weeks following initial release of the service pack, a number of issues were spotted arising from clashes between various aspects of the update and a selection of third-party anti-malware and security products, but most were quickly resolved. This test should see products at the top of their game, on a mature and stable platform, but as usual there is no knowing just how the range of updates will affect the products during the in-depth grilling applied on the VB test bench.

The toughness of this month’s test was kept to a minimum thanks to an early deadline (intended to allow adequate time to deal with the anticipated glut of entries), which meant that the release of the May 2008 WildList narrowly missed the cut-off date for this month’s test. The test sets were frozen on 20 June, using the April WildList for the core certification set, with the product submissions taken and frozen on 24 June.

The false positive set saw its usual expansion with new files and packages, and the other test sets were also extended somewhat, most notably the polymorphic set which saw several new items introduced in fairly limited numbers. This will be added to over the next few months as further generations of samples are replicated and verified.

The legacy set of older and more obscure items was left out of this test, something which has been planned for some time. Interest in such items continues to fluctuate, with a surprising number of macro and even DOS viruses still cropping up on the prevalence reports we gather, and this set may occasionally be resurrected for server tests where it has more relevance. In its place is a new set of trojans, an introductory selection of several thousand samples gathered over the course of the last six months or so. This move heralds a planned expansion in this direction for the VB sets, and we hope to have further improvements in the upcoming tests.

With an entirely new set of samples to measure detection against, a new platform on new hardware and a selection of new products, I expected the month of testing to be eventful, so I quickly got down to the lab and started testing.

-----------------------------------------------------------------------------------------------