Jump to content

Talk:Trigonometric integral

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Pictures, please!

[edit]

This article needs pics. dima (talk) 00:31, 5 February 2008 (UTC)[reply]

Is the integral sinus entire function?

[edit]

Can anybody suggest the reference to confirm that the integral sinus is entire function? dima (talk) 04:27, 17 May 2008 (UTC)[reply]

This artcile need a reformation

[edit]

It's using complex formulas and graphs to show something that most people don't undertand. It need to be made simpler for the everage joe like me. Less graphs more words!--209.80.246.30 (talk) 13:26, 9 April 2009 (UTC)[reply]

[edit]

Given function

f(x)=1, when |x|<1,
f(x)=1/2, when |x|=1,
f(x)=0, when |x|>1.
Need to write this function through Fourier integral.
Solution is for this example:
In particular case if x=0 (|x|<1), then
and we put 0 into x place and we get And so we have:
As far as I understand about Fourier integral, this integral means:
But problem is, that I check it through Free Pascal program "Version 1.0.12 2011/04/23; Compiler Version 2.4.4; Debugger GDB 7.2" with this code:
  var
  a:longint;
  c:real;
  begin
        c:=0;
        a:=0;
          for a:=1 to 100000
          do
            c:=c+sin(a)/a;
  writeln(c);
  readln();
  end.
so I get result 1.07080565212341. Even not close to — Preceding unsigned comment added by Versatranitsonlywaytofly (talkcontribs) 22:52, 22 December 2011 (UTC) BTW, you can use it like benchmark changing number in line "for a:=1 to 100000" to bigger than 100000. With number 1000000000 I got 1.07079632630307 and it take for CPU 52 seconds to compute result. You can use "a:integer" instead "a:longint", but then smaller number you will be able to choose. With number 100000000 it tooks only 5 seconds and result is 1.07079633477997.[reply]
It apears just simple mistake, I thought impossible that it mean that it mean, because but it not, it equal to ~1. But interesting coincidence, that result is it something must to do with Fourier series and coefficient . So real code is:
  var
  b:real;
  a:longint;
  begin
        b:=0;
        a:=0;
          for a:=1 to 1000000000
          do
           b:=b+0.00001*sin(a/100000)/(a/100000);
  writeln(b);
  readln();
  end.
And result is 1.57088654523321 after 63 seconds.

Gutting of "Efficient Evaluation" Section

[edit]

I propose to greatly reduce the somewhat sloppy listing of numbers in the "Efficient evaluation" section of the article. The compiled picture of multitudes of decimal numbers with 10 digits looks ridiculous and needs to be changed. My suggestion would be to reduce this section to the first sentence. Miles Cranmer 15:10, 12 December 2015 (UTC) — Preceding unsigned comment added by Mdtmc (talkcontribs)

I undid this excision. While you may not particularly care about how to efficiently evaluate Si(x) or Ci(x), there is a remarkable lack of information online about this, so this contribution on Wikipedia is particularly useful. The formulae come from a paper by Rowe et al (2015) in the journal Astronomy and Computing (see Appendix B of the paper). This is probably not a place that many people looking for this information will stumble across, so having the information public on Wikipedia serves a useful service.Rmjarvis (talk) 17:20, 20 October 2016 (UTC)[reply]

I agree that this evaluation section is useful, but it might be hidden behind a [show] drop down table to avoid cluttering. Also: for Si(x) the approximations are valid for also, by making use of the property: . This property might be useful to state somewhere in the article. 84.236.7.224 (talk) 19:58, 22 October 2016 (UTC)[reply]

Ci/Si demo

[edit]

Hello,

I made an interactive plot of the trig integrals: https://anematode.github.io/grapheme/demos/trig_integrals.html. Can I add it to the External Links section? — Preceding unsigned comment added by Ovinus Real (talkcontribs) 03:42, 12 July 2020 (UTC)[reply]

Probably not a good idea. See WP:ELNO #11. –Deacon Vorbis (carbon • videos) 04:26, 12 July 2020 (UTC)[reply]