to some degree the redundancy in the translation codons is there because the mating tRNA isn't always an exact match.
We like to think of it as a nice clean process, but in reality it's a VERY messy process involving lots of fuzzy factors and probabillities. The way that the ribosome works is by having the tRNA "crash" randomly into the opening. If it's the correct tRNA it will bond to the DNA and stay there. Ones that don't fit are repelled by molecular forces. There is such a small difference between the bonding potential of GUU and GUA that if they were assigned to different amino acids there would be many many more mistakes, resulting in miscoded proteins.
redundancy in wetware is usually there for a reason!
Unix soit qui mal y pense [Unix to him who evil thinks?]