Eric,
Eric Ruttan wrote:

As a public service announcement...
Please change the post topic to fit your post.  It just makes sense.  Get
creative!  Make a new Subject Title!

A question about belts; what are the race guys doing?  Do they see belt
failures? Is Tracy using a belt?  Ever failed?

I admit I feel more comfortable without belts, but we should be deciding
this on facts not feelings.  I don't have any facts.

I understand many automotive overhead cams use belts.  A failure will (if
interference engine) bend valves and break pistons.  These belts must be a
well known quantity.
  

Manufacturers insist on replacing them every 60k mi or so.  They are covered and thus protected from collateral damage from, say, V-belt failing or other FOD.

I perceive the "danger"of a belt is the potential cascading failure.  Belt
failure takes out alt and water pump and air-conditioning (if your John
Slade).  But i guess, if this case is sufficiently rare then a secondary alt
and water pump not effected by this event will serve as a back up allowing
us to complete the flight.
  

I would describe a "cascading" failure of a belt as primarily failing everything (water pump, alt, AC) that it is connected to and then flailing around taking out crank angle sensor, oil pressure sender, etc. that is NOT driven by the belt.

To argue the other side of the coin, not using the automotive water and alt
system is throwing away cheap well engineered and proven, easy to use system
away, to get something we "feel" is better.  

The case for EWP for example is performance.  PL insisted that an EDWP absorbs over 10 hp at 6000 rpm.  He is probably damned close.  He then made the unfortunate leap that therefore an EWP must absorb the same power.  Not true.  EWP conservatively absorbs 14V x 5 A = 70 W =~ 0.1 hp.  He was off by about two orders of magnitude or about 9.9 hp.  Don't know about you  but I can always use an extra 9.9 hp.

Remember, the most common failure of auto conversations is the supporting systems.
  

EXACTLY.  But do you regard that as a reason to stick with 1920's technology?  Or do you, like me, regard that as a reason to design and develop these systems very carefully?

Eric