In the README, make consistent evaluation models a design decision

[kloimhardt]

I propose to add the following sentence to the README:

Using Noj, any file can be executed in two different ways, namely with and without an explicit (load-file ...) in the command-line. It is a design decision of Noj that both ways share the same evaluation model. Especially the number of evaluated expressions and results of (print ) seen in the terminal on a first execution are expected to be the same in both cases.

I propose this because time and again there are discussions about changing the execution model of Clay (e.g. evading whole chunks of time-consuming expressions, continue after un-catched exceptions). We have to expect that this discussion will pop up again and again as people tend to forget to think about keeping the Noj beginner experience clean. I think it is important for the future of Clay to nail down once and for all the decision that such ideas (at the least) must and will never affect a simple first execution of a Clojure file.

Any change in the evaluation model can be done via an option to (clay/make! ...)

It think the Noj beginner experience must go far beyond SciCloj. It is meant as a generic Clojure experience, a carte de visite for the whole Clojure community to the outside world. Changing by default the execution model of Clay jeopardizes all the SciCloj efforts to be accepted WITHIN the Clojure community for SciCloj as a Clojure ambassador.