It's not suitable for our aspiring beginners[0], but my recent
article suggests a programming exercise, not only easy (and
enjoyable), but also of historical interest:
http://queue.acm.org/detail.cfm?id=2724586
The basic problem is to write a "compiler-compiler" (the square root
of a compiler?). I provide some historical context and translations
to modern vocabulary, in an effort to make it easier to approach the
over 50-year-old original paper, but in a modern environment the task
should be well within the grasp of anyone who calls themselves a
programmer.
-Dave
(the former and current compiler writers among us should find the
task itself trivial, but —being less distracted by mechanics— are
more likely to appreciate the subtler pleasures of recreating[1] self-
reproducing systems)
[0] for beginners, I woud instead recommend the classic chestnut:
write a program which prints its own source. (and absolute beginners
are allowed and encouraged to find the "trivial" solutions, before
doing it properly)
[1] as per Schrödinger's aperiodic crystals: Val Schorre's program
concretely reproduced itself on IBM 1401s, but his paper allowed me
to transpose it (via a language invented roughly halfway between then
and now) to a program concretely reproducing itself on modern
machines — so, in my view, the paper is an abstract self-reproducer.
