Gary - lucky b******
Thinking a bit about Jwenting's comments - which I agree with entirely with relation to film SLRs - I'm wondering if a different set of rules apply to DSLRs.
1 - Fristly, the reason we get a multiplier effect on a DSLR is that the sensor is smaller than a 35mm frame. In effect, the sensor is only using the central portion of the image formed by the lens. Given that the problem with most wide-angles is edge definition and distortion towards the edge of the frame, it could be that faults which would condemn a lens in tests based on a 35mm frame might not even appear in a DSLR based shot - I can't think of any case of a lens not being better towards the center of the image than the edge
2 - Resolution - lens resolving power is a fixed, measurable qualtity ... film is more variable as the emulsion, processing, exposure all affect the ability to resolve fine detail. The DSLR sensor, however, has a known theoretical max resolution - still significantly less than the theoretical max of film. Hence a lens which does less well than others on a bench test may still be capable of providing as good an image as a DSLR can use.
Point is, all lenses are still tested and measured based on 35mm film use. These results may not be entirely useful in relation to the DSLR - it could be very worth while investigating some of the bargain models. It should be possible to go into any decent dealer with your camera and try some shots with top end and lesser lenses on a test target, pop home, check the results on the computer, and make your decision.
It would be interesting to see lens testing done both for the 35mm format and an arbitrary sensor format (say 1.5x).
Colin K. Work, Pixstel