Hacker News Reader

Comment by catnaroek | Hacker News Reader

> List comprehension should be no stranger to a Python or (and?) Haskell user. It’s a really compact syntax that deals with Cartesian product [emphasis mine] of lists.

Ugh, no. I fail to see where the Cartesian product is in::

    [(x,y,z) | x <- [1..5], y <- [x..10], z <- [x..y]]

---

@dennisvennink

What you feel doesn't matter. It's not a zip.

View on HN

I think you're being downvoted for being pedantic (and rude, in your edit), but I agree. This is a pretty good example of software engineers adopting mathematical terms they don't quite grasp in order to add an air of rigour to what they write. The statement is unhelpful for anyone that doesn't know what a Cartesian product is and irritatingly inaccurate for anyone who does.

I'm not sure of the original intent, but you could approach this as the Cartesian product of [1..5] * [1..10] * [1..10] which is then filtered such that the 2nd value is not smaller than the first, and the 3rd value is inclusively between the two

A subset of a Cartesian product is not a Cartesian product. inb4, spare me the wisecrack “It's the Cartesian product of itself and no other factors”.

> It’s a really compact syntax that deals with Cartesian product of lists.

Couldn't taking the subset of a Cartesian product be considered to be dealing with one?

Exactly. Indeed, from what I recall from relational database theory, joins are usually treated as cartesian products (outer joins) which are then filtered by the join condition.

An outer join a Cartesian product is not.

Even if you stick to ordinary (inner) joins, they are pullbacks, not products.

https://en.wikipedia.org/wiki/Pullback

D'oh, sorry, I meant "cross join"

Indeed. Feels more like a zip (or convolution) to me.

Reply on news.ycombinator.com