Get the latest tech news
Should programming languages be safe or powerful?
Should a programming language be powerful and let a programmer do a lot, or should it be safe and protect the programmer from bad mistakes? Contrary to what the title insinuates, these are not diametrically opposed attributes. Nevertheless, this is the mindset that underlies notions such as, “macros, manual memory management, etc. are power tools—they’re not supposed to be safe.” If safety and power are not necessarily opposed, why does this notion persist?
Haskell forces you to express side-effects in the type system, but this lets you know that calling a function with a signature like String → Int won’t do any IO or throw an exception. The evolution of syntax macros in Lisp, Scheme, and Racket provide an interesting real-world instance of how safety and power can start off as a trade-off, but with better language design, become complimentary. The problem with this is that these macros are unhygienic: if I introduce a new variable, as I did with tmp in my-or, that is just a bare symbol that can be inadvertently captured producing unexpected output:
Or read this on Hacker News
