I'd say that 2 is indeed quite common in Computer science. In graphs, for example, each edge has 2 ends. In an abstract syntax tree, a binary operator has 2 children. All raster images are based on 2 dimensions. A "for each" loop often involves 2 variables: the iterator object and an index.

In all such cases it makes sense that a language/ construct/library special-cases 2 elements without generalizing for an arbitrary number of elements.

By the way, 3 is much less common, except in some specific domains, like 3D vectors, UNIX permissions (user, group, other), etc.

The argument against two would be: if you can do two, why can you not do three?

Wherever I need "Two" I usually find that what I actually need is a one and a many; there's not a master and a replica, there's one master and many replicas. There's not two places that the data is when doing a raid restripe, there's many places the data probably is and one place it should be.

I seem to recall that there are certain languages (human languages) that have cases specifically for talking about pairs of objects, situated semantically between singular and plural. I think one of these is an archaic form of Greek, maybe ancient or Hellenistic.

It has always "bothered" me that AWS allowed for up to two API keys per IAM user.

How about a sign bit?

https://en.wikipedia.org/wiki/Unum_(number_format)

No.

But I have the decency to feel bad and leave an apologetic comment.

A ping-pong buffer certainly seems like generalising to a ping-pong-pang-peng-pung-p(infinity)ng buffer would be overkill.

i fairly regularly want pairwise, yep.

Two is quite relevant for relationships