Hacker News Reader

Comment by filoeleven | Hacker News Reader

One thing that stuck with me from a good Clojure book(1) is that using < is greatly preferable to using >. Coming from an OO background and reading left-to-right, this made sense to me: list the args from smallest to largest.

(< 1 2 3) ;; true

is a lot more intuitive to me than

(> 3 2 1) ;; true

even if one or more of those literals is replaced with an argument. It’s especially apparent with the two-arg use of the function. This is one of the few places where I still feel like infix notation has an advantage.

This isn’t even a nitpick about the article, just a remark about some cognitive dissonance seeing > used here.

(1) Maybe Elements of a Clojure? Not sure that’s it; it was a book referenced on HN which has a whole chapter on naming things.

View on HN

> (1) Maybe Elements of a Clojure? Not sure that’s it; it was a book referenced on HN which has a whole chapter on naming things.

That's correct.

Both factorizations feel the same to me, without context. Subsequent numbers are greater than the previous ones vs. subsequent numbers are lesser than the previous ones.

What should really matter is how you're "talking" in the code. If you've defined some relationship as y being greater than x, then that's how you should phrase the comparison as well: (> y x). (< x y) would be the same thing logically, but we don't read logically: it requires another mental step to understand that (> y x) and (< x y) are equivalent. It's avoiding those extra steps that gives code a really big readability boost.

The historical preference of < over > seems to be widespread in programming languages, here's a credible sounding explanation of the origins: https://stackoverflow.com/a/12146965

Regardless of the origins, < is just a more natural operation than >. I kind of mean that both figuratively and literally: the most "natural" numbers are the "natural numbers", which follow 1 < 2 < 3... and not 1 > 2 > 3...

I see what you mean. But if we consider endianness and reading direction arbitrary conventions does it still hold? Maybe it does, since in your latter example the reading direction is inverted in any case and endianness applies only intra-number.

Yeah, I was just talking about the mathematical less-than operator. How we choose to represent anything computationally or visually is orthogonal to it.

This is true in other languages as well, even Clojure's polar opposite, C++, where operator< is canonical tool used for sorting and other things.

Reply on news.ycombinator.com