>Having a machine to interact with allows you to
test your code on the
spot
and if you are
writing in an interpreted language the error-checking the
interpreter provides is a godsend for the coder. Why anyone would code
without the interaction of the target machine is beyond me.
I write perfect code, like Mozart it flows out in its final form to the
paper, and then to the system.
I think the immediacy of interactive programming causes the programmer
to tend to exit the design loop early, before the design has actually
crystallized in the mind. Working with paper provides the slowdown
needed to allow this crystallization to occur.
And now that I think about it...
ISTR a paper by someone at Purdue during the 60s, a paper exploring
the value of interactive computing (timesharing), and it specifically
referred to hypothetical systems so fast that the benefits of being
able to immediately submit a design to test would be lost due to the
programmer jumping out of the design loop too quickly.
But, to each his own... as long as they're not working for me.
-dq