Talk:Equivalence of Definitions of Real Exponential Function

From ProofWiki
Jump to navigation Jump to search

If we want to keep $\exp x := e^x$ as a definition, then we have to clarify exactly what we're asserting. As it is, $e^x$ is defined as $\exp x \ln e$, which doesn't allow you to define $\exp$ without being circular. Do we want to take the definition of integer and rational number indices and try to extend it to all real indices? Or scrap the def'n and just keep $e^x = \exp x \ln e$ as a justification of notation, rather than a definition? --GFauxPas 06:40, 8 February 2012 (EST)

They are now all interconnected except for the $e^x$ definition, which would be a nice, intellectually dishonest reason to scrap the $e^x$ definition. --GFauxPas 09:09, 8 February 2012 (EST)
Another suggestion: Keep the $\exp x = e^x$ as a definition for $x \in \Z, x \in \Q$, and prove the consistency for those. For $x \in \R \setminus \Q$, it will instead be a justification of notation. --GFauxPas 15:47, 9 February 2012 (EST)

There may be a theorem saying that a continuous function on $\Q$ (in the sense of the adapted $\epsilon$-$\delta$ definition) extends uniquely to an $\R$-continuous function. This will combine with the above for a proof of equivalence of definitions. That is, the $e^x$ notation well be denoting the unique extension of the function $q \to e^q$ on $\Q$. --Lord_Farin 17:19, 9 February 2012 (EST)

I took my last suggestion and implemented it. Please let us know if you find such a theorem. --GFauxPas 09:25, 10 February 2012 (EST)
A friend of mine immediately answered positively when I mentioned the theorem. It seems fairly straightforward; I haven't got a source, though. --Lord_Farin 18:03, 10 February 2012 (EST)

3 implies 5

Right now the 3 implies 5 section is a mixture of two different pages, but I'm not sure it's a good idea. What happened was that in the original proof (not mine):

$\ds \lim_{x \mathop \to 0} \frac {\exp x - 1} x$

was included, and I moved it to Derivative of Exponential at Zero. But there are three proofs up for Derivative of Exponent at Zero and only the second one works for this proof; proof 1 and proof three there would make 3 implies 5 here circular. What do I do? --GFauxPas 07:22, 17 February 2012 (EST) i

With the new structure for Derivative of Exponential at Zero up, I have rewritten the page; is this an acceptable solution? --Lord_Farin 07:38, 17 February 2012 (EST)
I think it's better this way. It seems more consistent with the approach I've been setting up; keeping the proofs "dissected" but telling the reader which proof and definition we're assuming. Thanks a lot. --GFauxPas 07:42, 17 February 2012 (EST)

Reordering Systematically

I'm working on re-ordering these proofs systematically. After a good amount of thought, this will require re-ordering the definitions. However, the pages on which these definitions reside are numbered. (E.g., the sequential definition is found on Definition:Exponential/Definition_3.) I could reorder the definitions in the theorem and still maintain the page names, but then we will have cases where, e.g., "Definition 1: Limit of Series" will point to a page called "Definition 4". This is inelegant for sure.

May I rename the definition subpages? (Note: I'd like to remove the arbitrary numbering system altogether.) Further, if I do so, will this cause issues with proofs and definitions that point to these subpages? Lastly, if this does cause issues, I will be happy to fix all links. However, I would require some way to find all the pages that point to these definition subpages. --Keith.U 22:57, 11 June 2016 (UTC)

We've had discussions about this in the past, and results were inconclusive.
a) Numbering them means we don't have to invent a name for them when a name is often difficult to come up with.
b) Naming things means we don't have put things in any particular order, and it makes it possible to put them in whatever order is considered more logical.
We had someone who got quite involved in this website who decided his own proofs were better than those that existed, so he renamed his proofs to be Proof 1 and renamed everyone else's proofs to 2 and 3 and so on, which irritated some of us.
But if pages are named rather than numbered, that would not matter. Might be worth removing the numbers altogether and just leave the names, but then, as you say, there will be a lot of work needed to rename all the pages which link to them. Not a particularly big job, or at least it shouldn't be.
Feel free to go ahead, but please: a) Never do a "move" without leaving a redirect (I believe you do have the authorisation to "move", I lose track of what basic users can do) and b) if you split up pages and move things around, please take note to invoke the SourceReview template (please take a look at this and see how it is used) because one of the aspects of this site which is important to accomplish, for various reasons, is a linear flow through the works cited in the Sources section, and this is easy to break with injudicious moving around of stuff. --prime mover (talk) 23:23, 11 June 2016 (UTC)
An interesting and laudable initiative. It might be good to try out the desired structure as subpages of your user page (User:Keith.U/Whatever) before tinkering in the main namespace (I've learnt this the hard way -- especially the risk of getting bogged down half-way and forgetting what was already done and what wasn't is substantial). As to finding the links, the menu on the left, under Tools, has the "What links here": Special:WhatLinksHere/Pagename. It's indispensible for refactoring jobs of any nontrivial size. — Lord_Farin (talk) 17:28, 12 June 2016 (UTC)