JetDogRush

Yes! I'm the lead developer for the ground control software for the Mars rovers. The missions are expanding the bounds of human knowledge, and doing my job right means we get more/better science data back from the rovers!

You're right - you can see the cable harnesses to the drive actuators in the original article. A major concern would be if a big flake bends inward and manages to sever those wires.

Surprisingly the wheels only weigh a few pounds each! (I still can't think in metric and I don't have a real number.)

They were primarily constrained by packing volume during cruise (https://mars.nasa.gov/msl/mission/spacecraft/cruiseconfig/) and then by shock loads as the securing pyrobolts blew during landing (https://youtu.be/Ki_Af_o9Q9s?t=4m14s).

True! The actual induced strain is not particularly high (the wheels do not visibly deform) but it's the cycle count that will get you. The chevron-shaped grousers also act as stress concentrators, which is why the cracks are all starting at those tips.

Not quite - primary mission was two years (one Martian year). We are in Extended Mission #2.

If only Curiosity were as fast as the KSP rovers... top speed is somewhere around 100 meters/hour, and usually more like 40!

There are two major factors - fatigue (think bending a paper clip repeatedly) due to driving over tons of small rocks, and the the mechanics of having six wheels contributing additional damage to the front and middle wheels. Think of pushing a rolling suitcase versus pulling one - when hitting an obstacle it tends to dig in instead of easy popping up. The front and middle wheels are in front of their pivots, and so tend to get driven in to rocks by the other wheels.

This seems dramatic, but there is a ton of life left in the wheels. Some of my colleagues performed life testing on the wheels after we first discovered the cracks and found the rover is capable of driving around on the titanium support struts even after the aluminum wheels have completely disintegrated. We've got a long way to go yet!

As others have said, the light time is an issue, but it's also not a constant data stream - we uplink data directly to the rover, but can only get data downlinked by relaying it through the MRO and Odyssey orbiters. That happens 2-3 times a day (or sol).