Hi Jon-Paul,
thank you for the examples. I would like to put them into one of the cook books we discussed.
So in test 1 you are doing a actual calculation on a ellipsoid, why do you think it is 30km off?
In the second example you came across another bug 
as the definition of the SRS 100000432 is not correct in revision 70!
It should have "planar degree" as its Linear UoM and not "meter" this is why it produces the same results as SRS 0 as it is not detected that the radians must be converted to meter first.
The fix is already made and scheduled for the next revision.
In the third test you are using the default SRS 0 as you already found out which does a plain euclidian distance calculation.
Cheers
Gerrit