www.neuron.yale.edu

Is there a rule for deciding the nseg in relation to its diam and L?

There have been several attempts to come up with strategies for deciding
how big a compartment shoudl be. Most of them are pretty useless.

The only exception I know of is to make compartments shorter than
some fraction d_lambda of the AC length constant at some frequency
that is so high that most membrane current is capacitive, not ionic.
In practical terms, d_lambda = 0.1 for a frequency of 100 Hz is good
enough for most models.

This strategy, which Michael Hines and I developed, is called the
d_lambda strategy. It's real easy to use with NEURON because it
is built into the CellBuilder, and you can use it even if you don't use
the CellBuilder.

You can read about the d_lambda strategy by going to NEURON's
FAQ page
http://www.neuron.yale.edu/neuron/faq/general-questions
and looking for
What's a good strategy for specifying nseg?

You'll also want to read this one:
Why should I use an odd value for nseg?