User:IssaRice/Tao's notation for limits: Difference between revisions

From Machinelearning
No edit summary
No edit summary
 
(3 intermediate revisions by the same user not shown)
Line 12: Line 12:


Rather than thinking of this as a result per se, I think it's better to think of this as ''alternative notations''. Why might one prefer one notation over the other? I think the strength of Tao's notation is that it works for anonymous functions/expressions. To be able to use the function restriction notation <math>f|_E</math>, we must have named our function beforehand. To give an example, we can write something like <math>\lim_{x\to 0;\, x\in (0,\infty)} |x|/x = 1</math>, but this is difficult to write in the other notation; we would have to say something like, "Let <math>f:\mathbf R\setminus \{0\} \to \mathbf R</math> be defined by <math>f(x) := |x|/x</math>. Then we have <math>\lim_{x\to 0} f|_{(0,\infty)}(x) = 1</math>."
Rather than thinking of this as a result per se, I think it's better to think of this as ''alternative notations''. Why might one prefer one notation over the other? I think the strength of Tao's notation is that it works for anonymous functions/expressions. To be able to use the function restriction notation <math>f|_E</math>, we must have named our function beforehand. To give an example, we can write something like <math>\lim_{x\to 0;\, x\in (0,\infty)} |x|/x = 1</math>, but this is difficult to write in the other notation; we would have to say something like, "Let <math>f:\mathbf R\setminus \{0\} \to \mathbf R</math> be defined by <math>f(x) := |x|/x</math>. Then we have <math>\lim_{x\to 0} f|_{(0,\infty)}(x) = 1</math>."
An amusing fact/potential advantage of the function restriction notation is that one can write it in a ''point-free notation''. Namely, we can write <math>\lim_{x_0} f|_E</math>. Actually, even with Tao's notation we could have written <math>\lim_{x_0;\, E} f</math>.


I think usually one would write the above like <math>\lim_{x\to 0+} |x|/x = 1</math>. So then one gets to keep anonymous functions, but at the expense of adding more complexity to the notation (one now needs to assign meaning to "<math>0+</math>").
I think usually one would write the above like <math>\lim_{x\to 0+} |x|/x = 1</math>. So then one gets to keep anonymous functions, but at the expense of adding more complexity to the notation (one now needs to assign meaning to "<math>0+</math>").
Line 23: Line 25:
| Tao's notation || <math>\lim_{x\to x_0;\, x\in E} f(x)</math> || Works with anonymous functions. || Somewhat verbose: the "<math>x\in E</math>" part must be written out each time.
| Tao's notation || <math>\lim_{x\to x_0;\, x\in E} f(x)</math> || Works with anonymous functions. || Somewhat verbose: the "<math>x\in E</math>" part must be written out each time.
|-
|-
| Function restriction notation || <math>\lim_{x\to x_0} f|_E(x)</math> || || Doesn't work with anonymous functions.
| Function restriction notation || <math>\lim_{x\to x_0} f|_E(x)</math> || Consistent with the standard notation. || Doesn't work with anonymous functions.
|-
|-
| Standard notation || <math>\lim_{x\to x_0} f(x)</math> || Works with anonymous functions. ||
| Standard notation || <math>\lim_{x\to x_0} f(x)</math> || Works with anonymous functions. || Set used in the limit does not appear in the notation, so one must assume some default set (such as the entire domain of the function) or introduce new notation for e.g. one-sided limits as e.g. <math>\lim_{x\to x_0+} f(x)</math> and <math>\lim_{x\to x_0-} f(x)</math>.
|}
|}

Latest revision as of 03:23, 1 December 2018

Tao's notation for a limit is limxx0;xEf(x).

Can we write this in a more standard way? basically, if we give one additional definition, we have the very appealing formula limxx0;xEf(x)=limxx0f|E(x).

The additional definition is this: if f:XR, then we define limxx0f(x):=limxx0;xXf(x). In other words, by default we assume that the limit is taken over the entire domain of the function.

Now, given f:XR and some EX, we have f|E:ER. Thus, limxx0f|E(x)=limxx0;xEf|E(x).

By exercise 9.4.6, limxx0;xEf(x)=limxx0;xEf|E(x)

Combining these two equalities, we have limxx0;xEf(x)=limxx0f|E(x) as promised.

Rather than thinking of this as a result per se, I think it's better to think of this as alternative notations. Why might one prefer one notation over the other? I think the strength of Tao's notation is that it works for anonymous functions/expressions. To be able to use the function restriction notation f|E, we must have named our function beforehand. To give an example, we can write something like limx0;x(0,)|x|/x=1, but this is difficult to write in the other notation; we would have to say something like, "Let f:R{0}R be defined by f(x):=|x|/x. Then we have limx0f|(0,)(x)=1."

An amusing fact/potential advantage of the function restriction notation is that one can write it in a point-free notation. Namely, we can write limx0f|E. Actually, even with Tao's notation we could have written limx0;Ef.

I think usually one would write the above like limx0+|x|/x=1. So then one gets to keep anonymous functions, but at the expense of adding more complexity to the notation (one now needs to assign meaning to "0+").

One should compare this to the function notation (f) vs Leibniz notation (ddxf(x)) for derivatives.

Notation Symbols Strengths Weaknesses
Tao's notation limxx0;xEf(x) Works with anonymous functions. Somewhat verbose: the "xE" part must be written out each time.
Function restriction notation limxx0f|E(x) Consistent with the standard notation. Doesn't work with anonymous functions.
Standard notation limxx0f(x) Works with anonymous functions. Set used in the limit does not appear in the notation, so one must assume some default set (such as the entire domain of the function) or introduce new notation for e.g. one-sided limits as e.g. limxx0+f(x) and limxx0f(x).