GetNewSamplePoints Method, relation 'chordHeight' to curve parameter (start/end)

[adadnet]

the class reference guide states that the above:

Returns a list of points on the curve. so far so good. it continuous:

This method's behavior is similar to GetSamplePoints except that this method forces the generation of new samples. The samples returned by GetSamplePoints could have a much smaller error than is specified by chordHeight. This method returns sample points with a maximum error as close to chordHeight as possible.

this is where i'm running into difficulties: what exactly is an 'error' with regards to sample points? the results that i'm getting from different chordHeight runs are tricky in that in some cases they seem perfectly fine but in others appear unpredictable; notably when the extent of the sample 'ignores' the set boundaries along the curve and 'jumps' to, say, the startparameter. however, i believe to recall instances where start- and endparameter of a curve NOT necessarily are synonymous with start- and endpoint of that curve, i guess depending on how that curve was created maybe - something that i have been conciously trying to avoid in this case.

anyway, help appreciated, thanks

felix

[adadnet]

PS well, if someone is interested, here's how i try make sense of it:

start- and endPOINT of a curve may be the same but any problem this may cause can be addressed by using its .evaluatepoint() method with its .startPARAMETER and .endPARAMETER value respectively as input parameter.

the chordHeight is somewhat of a density of the returned sample points along the curve, possibly in relation to the curve's vertices; this density increases with input values approaching towards zero+. i think an actual input value of zero increases the risk of a crash and so do negative values, whereas values above 1 don't actually effect further change.

comments still welcome of course