The true beauty of Logo comes in the realization that you can add your own words to it, and that all these specialist words you devise are themselves first-class citizens of the language, of syntax and status equal to its generic builtins. Hence:
TO CIRCLE
REPEAT 360 [FD 1 RT 1]
END
TO FLOWER
REPEAT 20 [CIRCLE RT 18]
END
and so on ad infinitum, until you arrive at a complete custom vocabulary that concisely and precisely expresses the particular concepts and behaviors of interest and importance to you. In doing so, you move from thinking algorithmically (which is glorified spaghetti) to thinking
compositionally, which is the key to
scalability – managing complexity as your needs and ambitions grow.
Whereas Algol-y languages treat user-defined vocabulary as second-class citizens, beneath their own privileged built-ins. Which is a ridiculous of status when you consider which is actually important to the user: precise, powerful, tailored words that describe their particular problem, or the crude primitive undifferentiated building blocks that the language dumps out of the box?
The beauty of bottom-up programming (as any Lisp fule kno:) is that endlessly tests your own understanding of the problem domain: to define effective, productive words you must have some idea of what you’re talking about; you can’t help but learn the foundations of the problem space as you go. There’s a basic humility to this approach; there’s nowhere to hide laziness or ignorance.
Whereas in top-down programming it’s much too easy for highly-educated highly-paid absolute know-nothings to bullshit eternally, constructing great theatrical class architectures; grand mechanical castles in the sky that look all very difficult and impressive to observers while never saying anything relevant or useful.
That key switch from algorithmic to compositional thinking is not a natural conceptual leap for self-learners – it takes a carefully directed prod at a particular point in the learning curve to jump those rails – but it opens up worlds. #PlatosCave