@Andydude References about the formalization of the Hyperoperations

[jaydfox]

First of all, I believe I wrote most of the Hyperoperations article on Wikipedia. There was an existing article called "Hyper operator", but there were many people on the Talk sub-page that led me to believe that the page needed a lot of work, so I tried to do my best with the rewrite. I included every reference that I could find on the topic, and compiled what I think is a comprehensive list of references that doesn't focus on Tetration, but all hyperoperations in general.

Almost every article I've read about hyperoperations refers to them as a sequence of binary operations. This is the "specific sense" that the first part of the article is about. I don't remember ever reading about indexed families or the like, but it's the only definition that is broad and yet concise enough to describe the current usage of the word among those who use it in this area of research. Many people also use the word in the "general sense", and I wanted to be clear about what this more general idea was. If you can be even more clear about it, then be my guest. But I am concerned about Wikipedia's policies about original research. I do not believe that the definition of the "general sense" is original research because it is a natural translation of other people's terminology from 50 years ago into today's terminology. So don't think of it as: "Andrew wrote the first definition", think of it as: "this is the definition that everyone has been talking about with different words".

Speaking of Wikipedia, has anyone else noticed that the "linear" approximation for tetration that is listed there is the naive version on the interval [-1,0], which is only C1 continuous for base e? I'm pretty sure we've discussed a better linear approximation somewhere, which is linear on a base-specific unit interval, which is C1 continuous, not just C0 continuous.

I stumbled upon it way back in the day:
https://groups.google.com/d/msg/sci.math...GK4dbvdcMJ

I see the same issue with the wikepedia article for the superlogarithm:
http://en.wikipedia.org/wiki/Super-logar...roximation

In fact, that article explicitly calls out that it's C0 continuous, when a proper choice of interval would be C1 continuous. I found a post here on the forum where I discussed the C1 linear approximation of the slog:
http://math.eretrandre.org/tetrationforu...php?tid=98

Granted, these methods only work for real bases greater than eta*. However, I still think they're far more useful than the naive linear approximation.

* Real, because you need to be able to place e in a particular unit interval, based on iterated logarithms. Greater than eta, because, as I mentioned elsewhere, the linear approximation for bases less than eta is only valid between the primary fixed points, not between 0 and the lower fixed point.