Promise to find a wall this afternoon, IF the weather lightens up AND I'll be out for the landscapes on Sunday. Thanks both of you, your comments are much appreciated.
A wall is not a good test target, especially with wide-angles. If the lens has some field curvature, which is very unlikely to be an issue in practise, it may show up as a problem when it isn't. Wide-angles are more prone to field curvature. The other thing is that positioning is critical and you must be
exactly square to the wall. With a wide-angle you'll be very close (another less than ideal condition) and it's actually very hard to get the necessary precision - impossible hand-held. If the camera is just a mm out, that'll mean all the corners will all be at different distances completely invalidating the results.
Here a much easier and more reliable test. Good light, decently fast shutter speeds at lowest f/number, keep the sun behind you and choose and a subject up the road like a road sign or car number plate. Focus carefully with centre AF point, then lock focus (switch AF to M, and be careful not to nudge the focusing ring). Then shoot four pictures with the target positioned in all four corners of the frame. It must be exactly the same distance from the corner in all cases, and make sure the light doesn't change - watch for clouds.
Compare images - you can often do this well enough by zooming in on the LCD. What you're looking for is a consistent standard of sharpness, not necessarily a high standard - this is a very tough test so don't expect that. If all four corners look pretty much the same, then the lens elements are properly centred and that covers 90% of potential problems (eg if the lens has taken a knock). If you have to look twice to check the differences, then there's unlikely to be a significant problem. All lenses vary a bit, but if there is something wrong it'll be obvious
