Rough ideas for pract. lessons

[code-hunger]

• show some tricks for quicker compile-run cycles in the CLI (use '&&' for joining commands and "↑" for accessing the previous command)
• let the students describe boolean conditions via propositional logic on the whiteboard before they make a C++ implementation
• let them describe formally their tasks (when the problem is in the form "how many numbers satisfy X" or "print the numbers which satisfy Y") using set-notation and predicate calculus
• let them find formulas for directly computing easier counting tasks (e.g. numbers divisible by N and M, less than 10³); probably on the whiteboard in parallel with the code-writing student.
• describe some tasks using only set notation and predicate calculus, no words, and let the students infer the meaning

[code-hunger]

• prepare students for a few things needed later in the course when teaching arrays
  • explain what is a constant, the difference between compile-time and runtime constants (const vs constexpr), how these can be used for safer code, what is sizeof.

[code-hunger]

• I lied about the impossibility for T var[n] (where n is non-constexpr) due to unextendability of the array. It is possible, and is available in C99, though its usage is discoraged.
  The problem's actually that the type system becomes kind of unstable (@todo think of a better word) at compile-time (E.g. sizeof will become non-constexpr).
• explain why you can't do stuff like array.size() or why there can't be such a thing as length(array), neither theoretically nor in practice (when working with arrays directly initialized in memory, be it the stack or the heap).
• explain how arrays in C/C++ are "pure", in the sense that we directly operate on the underlying data; talk about the complete lack of abstraction, that we're concrete as hell, what power this gives us and what errors are terribly easy to make and how to avoid them.