Hacker News Reader

Comment by c0mptonFP | Hacker News Reader

You hit the nail on the head, mostly.

> then you read the room and go with the layer of abstraction needed.

Finding the right layer of abstraction is orthogonal to the write-speak axis. When speaking to my colleagues, I use technical jargon that no layman could understand. None of the topics are simple, or strongly abstracted. The issue of write vs. speak is more about the sentence structure, sentence length, and breadth of vocabulary.

But I generally agree that carefully crafted written language can capture and transport thoughts much, MUCH more effectively.

View on HN

Slightly off topic: What's with HN and the word "orthogonal"?

I'm not a native English speaker, but I read a lot in English and it seems like the word is extremely common on HN compared to anywhere else.

Isn't usually "unrelated" a more descriptive and even a more precise word in most HN discussions? (The parent comment here does seem to make a point using axes, so maybe it is more appropriate here?)

I see what you did there, but I will bite:

Orthogonal does not mean unrelated. Take two vectors in the plane. Them being orthogonal means that they have a 90 degree angle between them, so if you know the direction of one of them, the direction of the other one is severely restricted to two choices. So these vectors are very much RELATED. It's just that they are related in a way that makes them maximally different in a certain sense.

So if you want to say that two things are maximally different in a certain sense, you use orthogonal. If you want to say that one thing has no influence whatsoever on what the other thing is, and the other way around, you use unrelated.

For example, if you randomly choose a point in the plane, then its x and y coordinates will be unrelated, but not orthogonal. The vectors [x 0] and [0 y] are not unrelated, but certainly orthogonal.

Of course, this distinction is easily lost.

I understand that orthogonal and unrelated have different meanings. What I'm wondering is: Isn't "orthogonal" much more common on HN (18388 matches in search) than in other places?

I suspect that "orthogonal" is a word programmers fall in love with during some CS class and then overuse because it sounds sciency.

I think that orthogonal is a more visual word than unrelated. It invokes the image of axes pointing into different directions. I suspect many programmers just like this aspect of the word, while unrelated is somewhat bland, and also usually wrong.

Somewhat relatedly, I’ve never seen the word maximal used outside of HN and crypto rags.

They're both terms used in university level mathematics/CS courses, which folks in related industries are likely to have spent time in.

Unrelated means two things are not related in any sense. Orthogonal means two things are unrelated with respect to a specific property.

Unrelated is more general, and less precise. Orthogonal restricts the "unrelatedness" to the specific property being discussed. It's also a very visual and intuitive word.

It's not only HN btw.

Reply on news.ycombinator.com