RE: atoms, molecules

[Gavin Thomas Nicol <gtn@ebt.com>]

> I think that's one of the motivation of Simon. If you have to take leap
> second into account, then it would be better if you can create it from
> atoms, instead of pushing your requirement into the part 2 spec.

Right. I think the point is that the first step toward managing 
complexity is to decompose the complex problem into simpler
sub-problems. 

I think time has proven that in general, fine-grained, composable 
solutions are more easily understood, and more flexible than 
monolithic solutions.

The trick, of course, is getting the atoms right so that when
they are mixed together, the result isn't explosive!