[ODE] multiple contacts and sticking on the cusp

[Martin C. Martin]

Hey Russ,

ODE never ceases to amaze me with the level of support you give, Russ. 
Tracking down things like this, documenting them and fixing those LCP bugs
is a great example.

Since many people simply take the examples and change what they need to
change, and don't change anything they don't understand; and since the
difference between the two friction models is subtle, should the examples
be changed to use dContactApprox1?  For that matter, should it be the
default friction model?  At the very least, when discussing contact
parameters in the manual, should it say "if you don't know which friction
model to use, start with dContactApprox1?"

Just a thought,
Martin

Russ Smith wrote:
> 
> oh yes ... other people doing vehicle stuff may want to update CVS and
> use the dContactApprox1 mode too. this problem affected all rolling
> bodies that had multiple frictional contacts at different angles (i.e.
> car wheels going over bumps, up ramps etc).
> 
> let me know of any problems.
> 
> russ.