I have a minolta 28-75mm f2.8 and a sigma 28-200mm f3.5-5.6, the minolta f2.8 is very good in low light, I just compare these two lens both set at f5.6, aperture priority on my KM5D, the picture of the minolta len is darker with shutter speed of 1/20" and the sigma is very good at 1/13", I don't know why the KM5D did not detect the light and set the exposure correctly on every len? But when I use the M mode to set both len at f5.6 and 1/13", the picture of both lens are the same. Looks like that the sigma has done a better job than the minolta. I then tried the minolta 50mm f1.7, it has the same result of the the minolta 28-75mm when I set to F5.6, the KM5D sets the speed at 1/20" instead of 1/13" and the picture is little too dark.
Metering is a complicated system that I think most of us take for granted most of the time ( I know I'm always irritated when the cameras metering isn't what I expect). I do often shoot with bracketing (from -1 through +1ev in 1/3 stop increments) when shooting in difficult ambient light circumstances. This way I can choose the closest to desired exposure before editing. This isn't too inconvenient since I don't need to process film before deciding which shots to keep and which ones to dump.
One other thought- In the film days, the film processor would correct many of our metering errors before we saw the finished photo. Today, we see the shot just as we captured it, making it pretty needs our input.
I try to frame and compose most shots so I don't have to do much processing, but find that I haven't captured exactly what I wanted most of the time.
You're probably right about that KM 28-75 but it's a lot like that Tamron 28-300, everyone that has one loves it. I think I'd be smarter to wait for the Sony 16-80 due to it's superior focal length for digital, but you know how lens envy is....:G