# Fractals/Iterations in the complex plane/wake

How to find the angles of external rays that land on the p/q root point on the boundary of Mandelbrot set's main cardioid ?

# irreducible fraction

First check of p/q is irreducible

/*
gcc i.c -Wall
./a.out
*/
#include <stdio.h>

/*
https://stackoverflow.com/questions/19738919/gcd-function-for-c
The GCD function uses Euclid's Algorithm.
It computes A mod B, then swaps A and B with an XOR swap.
*/
int gcd(int a, int b)
{
int temp;
while (b != 0)
{
temp = a % b;

a = b;
b = temp;
}
return a;
}

int main (){

// internal angle = n/m  in turns
int n;  // numerator
int d;  // denominator

int dMax = 17;

for (d = 2; d <= dMax; ++d )
for (n = 1; n < d; ++n )
if (gcd(n,d)==1 ){

printf("n/d = %d/%d\n", n,d);	// irreducible fraction
}

return 0;
}


Output:

n/d = 1/2
n/d = 1/3
n/d = 2/3
n/d = 1/4
n/d = 3/4
n/d = 1/5
n/d = 2/5
n/d = 3/5
n/d = 4/5
n/d = 1/6
n/d = 5/6
n/d = 1/7
n/d = 2/7
n/d = 3/7
n/d = 4/7
n/d = 5/7
n/d = 6/7
n/d = 1/8
n/d = 3/8
n/d = 5/8
n/d = 7/8
n/d = 1/9
n/d = 2/9
n/d = 4/9
n/d = 5/9
n/d = 7/9
n/d = 8/9
n/d = 1/10
n/d = 3/10
n/d = 7/10
n/d = 9/10
n/d = 1/11
n/d = 2/11
n/d = 3/11
n/d = 4/11
n/d = 5/11
n/d = 6/11
n/d = 7/11
n/d = 8/11
n/d = 9/11
n/d = 10/11
n/d = 1/12
n/d = 5/12
n/d = 7/12
n/d = 11/12
n/d = 1/13
n/d = 2/13
n/d = 3/13
n/d = 4/13
n/d = 5/13
n/d = 6/13
n/d = 7/13
n/d = 8/13
n/d = 9/13
n/d = 10/13
n/d = 11/13
n/d = 12/13
n/d = 1/14
n/d = 3/14
n/d = 5/14
n/d = 9/14
n/d = 11/14
n/d = 13/14
n/d = 1/15
n/d = 2/15
n/d = 4/15
n/d = 7/15
n/d = 8/15
n/d = 11/15
n/d = 13/15
n/d = 14/15
n/d = 1/16
n/d = 3/16
n/d = 5/16
n/d = 7/16
n/d = 9/16
n/d = 11/16
n/d = 13/16
n/d = 15/16
n/d = 1/17
n/d = 2/17
n/d = 3/17
n/d = 4/17
n/d = 5/17
n/d = 6/17
n/d = 7/17
n/d = 8/17
n/d = 9/17
n/d = 10/17
n/d = 11/17
n/d = 12/17
n/d = 13/17
n/d = 14/17
n/d = 15/17
n/d = 16/17


# Wake and limb

Wake is the region of parameter plane enclosed by two external rays landing on the same root point on the boundary of main cardioid (period 1 hyperbolic component).

p/q-limb is a part of Mandelbrot set contained inside p/q-wake

Limb:

• start from from the root point
• end with tip point

External angles of p/q-wake:

• have period q under doubling map. It is the same as period of it's landing point ( c = root point ) and parent hyperbolic component
• length of periodic part of binary expansion is q
• preperiod under doubling map is zero

Points of of p/q-wake:

• Roots are landing points of parameter rays with periodic angles
• Misiurewicz points have preperiodic external angles

The size of the p/q limb is ${\frac {1}{2^{q}-1}}$

The p/q bulb can be recognized by locating the smallest spoke ( branch) in the antenna and determining its relative location to the main spoke.

# Combinatorial algorithm = Devaney's method

Devaney's method for finding external angles of primary buds

Steps :

Input : rational rotation angle

Outpout : external angle ( decimal or binary fraction )

## The code

### C++

Here is C++ code from the program Mandel by Wolf Jung :

// mndcombi.cpp  by Wolf Jung (C) 2007-2015, part of Mandel 5.13,
qulonglong mndAngle::wake(int k, int r, qulonglong &n)
{  if (k <= 0 || k >= r || r > 64) return 0LL;
qulonglong d = 1LL;
int j, s = 0;
n = 1LL;

for (j = 1; j < r; j++)
{  s -= k; if (s < 0) s += r; if (!s) return 0LL;
if (s > r - k) n += d << j;
}
//
d <<= (r - 1); d--; d <<= 1; d++; //2^r - 1 for r <= 64
return d;
}


### C GMP and MPFR

/*

------- Git -----------------
cd existing_folder
git init
git commit -m ""
git push -u origin master
-------------------------------

?? http://stackoverflow.com/questions/2380415/how-to-cut-a-mpz-t-into-two-parts-using-gmp-lib-on-c

to compile from console:
gcc w.c -lgmp -lmpfr -Wall

to run from console :

./a.out

tested on Ubuntu 14.04 LTS

Using MPFR-3.1.2-p3 with GMP-5.1.3 with precision = 200 bits
internal angle = 34/89
first external angle :
period = denominator of internal angle = 89
external angle as a decimal fraction = 179622968672387565806504265/618970019642690137449562111 = 179622968672387565806504265 /( 2^89 - 1)
External Angle as a floating point decimal number =  2.9019655713870868535821260055542440298749779423213948304299730531995503353103626302473331181359966368582651105245850405837027542373052381532777325121338632071561064451614697645709384232759475708007812e-1
external angle as a binary rational (string) : 1001010010010100101001001010010010100101001001010010100100101001001010010100100101001001/11111111111111111111111111111111111111111111111111111111111111111111111111111111111111111
external angle as a binary floating number in exponential form =0.10010100100101001010010010100100101001010010010100101001001010010010100101001001010010010100101001001010010100100101001001010010100100101001010010010100100101001010010010100100101001010010010100101001*2^-1
external angle as a binary floating number in periodic form =0.(01001010010010100101001001010010010100101001001010010100100101001001010010100100101001001)

.(01001010010010100101001001010010010100101001001010010100100101001001010010100100101001001)

*/

#include <stdlib.h> // malloc
#include <stdio.h>
#include <gmp.h>  // for rational numbers
#include <mpfr.h> // for floating point mumbers

// rotation map
//the number  n  is always increased by n0 modulo d

// input :  op = n/d ( rational number ) and n0 ( integer)
//  n = (n + n0 ) % d
// d = d
// output = rop = n/d
void mpq_rotation(mpq_t rop, const mpq_t op, const mpz_t n0)
{

mpz_t n; // numerator
mpz_t d; // denominator
mpz_inits( n, d, NULL);

//
mpq_get_num (n, op); //
mpq_get_den (d, op);

// n = (n + n0 ) % d
mpz_mod( n, n, d);

// output
mpq_set_num(rop, n);
mpq_set_den(rop, d);

mpz_clears( n, d, NULL);

}

void mpq_wake(mpq_t rop, mpq_t op)
{

// arbitrary precision variables from GMP library
mpz_t  n0 ; // numerator of q
mpz_t  nc;
mpz_t  n;
mpz_t  d ; // denominator of q
mpz_t  m; // 2^i

mpz_t  num ; // numerator of rop
mpz_t  den ; // denominator of rop
long long int i;
unsigned long int base = 2;
unsigned long int id;
int cmp;

mpz_inits(n, n0,nc,d,num,den,m, NULL);

mpq_get_num(n0,op);
mpq_get_den(d,op);
id = mpz_get_ui(d);
//  if (n <= 0 || n >= d ) error !!!! bad input
mpz_sub(nc, d, n0); // nc = d - n0
mpz_set(n, n0);
mpz_set_ui(num, 0);

// rop
// num = numerator(rop)

// denominator = den(rop) = (2^i) -1
mpz_ui_pow_ui(den, base, id) ;  // den = base^id
mpz_sub_ui(den, den, 1);   // den = den-1

// numerator
for (i=0; i<id ; i++){

mpz_set_ui(m, 0);
cmp = mpz_cmp(n,nc);// Compare op1 and op2. Return a positive value if op1 > op2, zero if op1 = op2, or a negative value if op1 < op2.
if ( cmp>0 ) {
mpz_ui_pow_ui(m, 2, id-i-1); // m = 2^(id-i   )
mpz_add(num, num, m); // num = num + m
if (mpz_cmp(num, den) >0) mpz_mod( num, num, den); // num = num % d ; if num==d gives 0
//gmp_printf("s = 1");

}
// else gmp_printf("s = 0");
//gmp_printf (" i = %ld internal angle = %Zd / %Zd ea = %Zd / %Zd ; m = %Zd \n", i, n, d, num, den, m);

// n = (n + n0 ) % d = rotation
if (mpz_cmp(n, d)>0) mpz_mod( n, n, d);
//

//
}

// rop = external angle
mpq_set_num(rop,num);
mpq_set_den(rop,den);
mpq_canonicalize (rop); // It is the responsibility of the user to canonicalize the assigned variable before any arithmetic operations are performed on that variable.

// clear memory
mpz_clears(n, n0, nc, d, num,den, m, NULL);

}

/*

http://stackoverflow.com/questions/9895216/remove-character-from-string-in-c

"The idea is to keep a separate read and write pointers (pr for reading and pw for writing),
always advance the reading pointer, and advance the writing pointer only when it's not pointing to a given character."

modified

remove first length2rmv chars and after that take only length2stay chars from input string
output = input string
*/
void extract_str(char* str, unsigned int length2rmv, unsigned long int length2stay) {
// separate read and write pointers
char *pr = str; // read pointer
char *pw = str; // write pointer
int i =0; // index

while (*pr) {
if (i>length2rmv-1 && i <length2rmv+length2stay)
pw += 1; // advance the writing pointer only when
*pw = *pr;
i +=1;
}
*pw = '\0';
}

int main ()
{

// notation :
//number type : s = string ; q = rational ; z = integer, f = floating point
// base : b = binary ; d = decimal

char *sqdInternalAngle = "13/34";
mpq_t qdInternalAngle;   // internal angle = rational number q = n/d
mpz_t den;

mpq_t  qdExternalAngle;   // rational number q = n/d
char  *sqbExternalAngle;
mpfr_t  fdExternalAngle ;  //
char  *sfbExternalAngle; //

mp_exp_t exponent ; // holds the exponent for the result string
mpz_t zdEANumerator;
mpfr_t EANumerator;
mpfr_prec_t p = 200; // in bits , should be > denominator of internal angle

mpfr_set_default_prec (p); // but previously initialized variables are unaffected.
//mpfr_set_default_prec (precision);

// init variables
//mpf_init(fdExternalAngle);
mpq_inits (qdExternalAngle, qdInternalAngle, NULL); //

// set variables
mpq_set_str(qdInternalAngle, sqdInternalAngle, 10); // string is an internal angle
mpq_canonicalize (qdInternalAngle); // It is the responsibility of the user to canonicalize the assigned variable before any arithmetic operations are performed on that variable.
mpq_get_den(den,qdInternalAngle);

if ( p < uiIADenominator) printf("increase precision !!!!\n");
mpfr_printf("Using MPFR-%s with GMP-%s with precision = %u bits \n", mpfr_version, gmp_version, (unsigned int) p);

//
mpq_wake(qdExternalAngle, qdInternalAngle); // internal -> external

mpq_get_num(zdEANumerator  ,qdExternalAngle);
// conversions
mpfr_set_z (EANumerator,   zdEANumerator,   GMP_RNDN);

sqbExternalAngle = mpq_get_str (NULL, 2, qdExternalAngle); // rational number = fraction : from decimal to binary

sfbExternalAngle = (char*)malloc((sizeof(char) * uiIADenominator*2*4) + 3);
// mpfr_get_str (char *str, mpfr_exp_t *expptr, int b, size_t n, mpfr_t op, mpfr_rnd_t rnd)
if (sfbExternalAngle==NULL ) {printf("sfbExternalAngle error \n"); return 1;}
mpfr_get_str(sfbExternalAngle, &exponent, 2,200, fdExternalAngle, GMP_RNDN);

// print
gmp_printf ("internal angle = %Qd\n", qdInternalAngle); //
printf("first external angle : \n");
gmp_printf ("period = denominator of internal angle = %Zd\n", den); //

gmp_printf ("external angle as a decimal fraction = %Qd = %Zd /( 2^%Zd - 1) \n", qdExternalAngle, zdEANumerator, den); //
printf ("External Angle as a floating point decimal number =  ");
mpfr_out_str (stdout, 10, p, fdExternalAngle, MPFR_RNDD); putchar ('\n');
gmp_printf ("external angle as a binary rational (string) : %s \n", sqbExternalAngle); //

printf ("external angle as a binary floating number in exponential form =0.%s*%d^%ld\n", sfbExternalAngle, 2, exponent);
printf ("external angle as a binary floating number in periodic form =0.(%s)\n", sfbExternalAngle);

// clear memory
//mpf_clear(fdExternalAngle);
mpq_clears(qdExternalAngle, qdInternalAngle, NULL);
free(sfbExternalAngle);

return 0;
}


## Examples

One can check the results with

### 1/2

The 1/2-wake of the main cardioid is bounded by the parameter rays with the angles:

• 1/3 = p01 = 0.(01)
• 2/3 = p10 = 0.(10)

The 1/2-wake of the main cardioid is bounded by the parameter rays with the angles 1/3  or  p01  and 2/3  or  p10 .

The angle  1/3  or  p01 has  preperiod = 0  and  period = 2.
The conjugate angle is  2/3  or  p10 .
The kneading sequence is  A*  and the internal address is  1-2 .
The corresponding parameter rays land at the root of a satellite component of period 2. It bifurcates from period 1.


Important points

• root point between period 1 and 2 = c = -0.75 = -3/4 = birurcation point for internal angle 1/2. Landing point of 2 external rays 1/3 and 2/3
• center of period 2 components c = -1
• tip of main antenna c = -2 = $M_{1,1}$ . It is landing point of externa ray for angle $0.01={\frac {1}{2}}=0.5$

### 1/3

Orbit of rational angle 3/7 ( and position in subintervals):

 1 / 3  = 0
2 / 3  = 0
0 / 3  = 1


so intinerary = 001

first external angle  = 001 = 1 / 7


The 1/3-wake of the main cardioid is bounded by the parameter rays with the angles

• 1/7 = p001 = 0.(001)
• 2/7 = p010 = 0.(010)

note that

• 1/7 = 0.(142857)= 0.1428571428571428571428571428571428571428571428571428571428571428571428571428571428571428571428571428571428571428...

but decimal expansion is not important here. Only decimal ratio and binary floating point is important

### 1/4

The 1/4-wake of the main cardioid is bounded by the parameter rays with the angles

• 1/15 or p0001 or $0.(0001)$
• 2/15 or p0010 or $0.(0010)$

### n/5

There are 4 period 5 wakes:

• 1/5
• 2/5
• 3/5
• 4/5

The 1/5-wake of the main cardioid is bounded by the parameter rays with the angles :

• 1/31 = p00001 = 0.(00001)
• 2/31 = p00010 = 0.(00010)

The 4/5-wake of the main cardioid is bounded by the parameter rays with the angles

• 29/31 = p11101 = 0.(11101)
• 30/31 = p11110 = 0.(11110)

### 3/7

Divide interval ( circle):

$I={\big (}0,1{\big ]}$ into 2 subintervals ( lower partition) :

$I_{0}={\big (}{\tfrac {0}{7}},{\tfrac {4}{7}}{\big ]}$ $I_{1}={\big (}{\tfrac {4}{7}},{\tfrac {7}{7}}{\big ]}$ Orbit of rational angle 3/7 ( and position in subintervals):

 3 / 7  = 0
6 / 7  = 1
2 / 7  = 0
5 / 7  = 1
1 / 7  = 0
4 / 7  = 0
0 / 7  = 1


So itinerary is :

 $0101001$ One can convert it to number :

$0101001\to 0.(0101001)_{2}={\frac {0101001}{1111111}}_{2}={\frac {41}{127}}_{10}$

The 3/7-wake of the main cardioid is bounded by the parameter rays with the angles

• 41/127 = p0101001 = 0.(0101001)
• 42/127 = p0101010 = 0.(0101010)

root point :

 c = -0.606356884415893  +0.412399740175787 i


Orbit of 41/127 under doubling map modulo 1 computed with this program ( exponent = 7 and mpz_init_set_ui(n, 41); :

41/127
82/127
37/127
74/127
21/127
42/127
84/127


### 5/11

ghci
GHCi, version 8.10.7: https://www.haskell.org/ghc/  :? for help
Prelude> :l bh.hs
[1 of 1] Compiling Main             ( bh.hs, interpreted )
*Main> :main 5 11
internal angle p/q = 5 / 11
internal angle in lowest terms =
5 % 11
rays of the bulb:
(01010101001) = 681 % 2047
(01010101010) = 682 % 2047


The 5/11-wake of the main cardioid is bounded by the parameter rays with the angles:

• 681/2047 = p01010101001 = 0.(01010101001)
• 682/2047 = p01010101010 = 0.(01010101010)

Center of period 11: c = -0.697838195122425 +0.279304134101366 i root point ( bond): c = -0.690059870015044 +0.276026482784614 i

Angled internal address: $1\xrightarrow {5/11} 11$

### n/17

The 1/17-wake of the main cardioid is bounded by the parameter rays with the angles

• 1/131071 = p00000000000000001 = 0.(00000000000000001)
• 2/131071 = p00000000000000010 = .(00000000000000010)

### 1/25

uiIADenominator = 25
Using MPFR-3.1.5 with GMP-6.1.1 with precision = 200 bits
internal angle = 1/25
first external angle :
period = denominator of internal angle = 25
external angle as a decimal fraction = 1/33554431 = 1 /( 2^25 - 1)
External Angle as a floating point decimal number =  2.9802323275873758669905622896719661257256902970579355078320375103410059138021557873907862143745145987726127630324761942815747600646073471636075786857847163330961076713939572483194617724677755177253857e-8
external angle as a binary rational (string) : 1/1111111111111111111111111
external angle as a binary floating number in exponential form =0.10000000000000000000000001000000000000000000000000100000000000000000000000010000000000000000000000001000000000000000000000000100000000000000000000000010000000000000000000000001000000000000000000000001*2^-24
external angle as a binary floating number in periodic form =0.(0000000000000000000000001)



So 1/25-wake of the main cardioid is bounded by the parameter rays with the angles :

• 0.0000000298 = 1/33554431 = 1 /( 2^25 - 1) = 0.(0000000000000000000000001)
• 0,0000000596 = 2/33554431 = 2 /( 2^25 - 1) = 0.(0000000000000000000000010)

One can check it with Mandel

The angle  1/33554431  or  p0000000000000000000000001
has  preperiod = 0  and  period = 25.
The conjugate angle is  2/33554431  or  p0000000000000000000000010 .
The kneading sequence is  AAAAAAAAAAAAAAAAAAAAAAAA*  and
the internal address is  1-25 .
The corresponding parameter rays are landing at the root of a satellite component of period 25.
It is bifurcating from period 1.
Do you want to draw the rays and to shift c
to the corresponding center?


The center is :

 c = 0.265278321904606  +0.003712059989878 i    period = 25


### 1/31

The 1/31-wake of the main cardioid

• is bounded by the parameter rays with the angles:
• 1/2147483647 = p0000000000000000000000000000001 = 0.(0000000000000000000000000000001)
• 2/2147483647 = p0000000000000000000000000000010 = 0.(0000000000000000000000000000010)
• root point : c = 0.260025517721190 +0.002060296266000 i
• center c = 0.260025517721190 +0.002060296266000 i
• principal Misiurewicz point c = 0.259995759918769 +0.001610271381965*i
• has preperiod = 31 , period = 1
• is a landing point for 31 external rays
• 2147483649/4611686016279904256 = 0000000000000000000000000000001p0000000000000000000000000000010 = .0000000000000000000000000000001(0000000000000000000000000000010)
• the biggest baby Mandelbrot set has the kneading sequence AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAB* corresponds to the internal address 1-31-32 . The period is 32. The smallest angles are 3/4294967295 = 0.(00000000000000000000000000000011) and 4/4294967295 = 0.(00000000000000000000000000000100)

On the dynamical plane :

• The angle

### 13/34

The 13/34-wake of the main cardioid is bounded by the parameter rays with the angles

• 4985538889/17179869183 = p0100101001001010010100100101001001 = 0.(0100101001001010010100100101001001)
• 4985538890/17179869183 = p0100101001001010010100100101001010 = 0.(0100101001001010010100100101001010)

 s = 0 i = 0 internal angle = 13 / 34 ea = 0 / 17179869183 ; m = 0
s = 1 i = 1 internal angle = 26 / 34 ea = 4294967296 / 17179869183 ; m = 4294967296
s = 0 i = 2 internal angle = 5 / 34 ea = 4294967296 / 17179869183 ; m = 0
s = 0 i = 3 internal angle = 18 / 34 ea = 4294967296 / 17179869183 ; m = 0
s = 1 i = 4 internal angle = 31 / 34 ea = 4831838208 / 17179869183 ; m = 536870912
s = 0 i = 5 internal angle = 10 / 34 ea = 4831838208 / 17179869183 ; m = 0
s = 1 i = 6 internal angle = 23 / 34 ea = 4966055936 / 17179869183 ; m = 134217728
s = 0 i = 7 internal angle = 2 / 34 ea = 4966055936 / 17179869183 ; m = 0
s = 0 i = 8 internal angle = 15 / 34 ea = 4966055936 / 17179869183 ; m = 0
s = 1 i = 9 internal angle = 28 / 34 ea = 4982833152 / 17179869183 ; m = 16777216
s = 0 i = 10 internal angle = 7 / 34 ea = 4982833152 / 17179869183 ; m = 0
s = 0 i = 11 internal angle = 20 / 34 ea = 4982833152 / 17179869183 ; m = 0
s = 1 i = 12 internal angle = 33 / 34 ea = 4984930304 / 17179869183 ; m = 2097152
s = 0 i = 13 internal angle = 12 / 34 ea = 4984930304 / 17179869183 ; m = 0
s = 1 i = 14 internal angle = 25 / 34 ea = 4985454592 / 17179869183 ; m = 524288
s = 0 i = 15 internal angle = 4 / 34 ea = 4985454592 / 17179869183 ; m = 0
s = 0 i = 16 internal angle = 17 / 34 ea = 4985454592 / 17179869183 ; m = 0
s = 1 i = 17 internal angle = 30 / 34 ea = 4985520128 / 17179869183 ; m = 65536
s = 0 i = 18 internal angle = 9 / 34 ea = 4985520128 / 17179869183 ; m = 0
s = 1 i = 19 internal angle = 22 / 34 ea = 4985536512 / 17179869183 ; m = 16384
s = 0 i = 20 internal angle = 1 / 34 ea = 4985536512 / 17179869183 ; m = 0
s = 0 i = 21 internal angle = 14 / 34 ea = 4985536512 / 17179869183 ; m = 0
s = 1 i = 22 internal angle = 27 / 34 ea = 4985538560 / 17179869183 ; m = 2048
s = 0 i = 23 internal angle = 6 / 34 ea = 4985538560 / 17179869183 ; m = 0
s = 0 i = 24 internal angle = 19 / 34 ea = 4985538560 / 17179869183 ; m = 0
s = 1 i = 25 internal angle = 32 / 34 ea = 4985538816 / 17179869183 ; m = 256
s = 0 i = 26 internal angle = 11 / 34 ea = 4985538816 / 17179869183 ; m = 0
s = 1 i = 27 internal angle = 24 / 34 ea = 4985538880 / 17179869183 ; m = 64
s = 0 i = 28 internal angle = 3 / 34 ea = 4985538880 / 17179869183 ; m = 0
s = 0 i = 29 internal angle = 16 / 34 ea = 4985538880 / 17179869183 ; m = 0
s = 1 i = 30 internal angle = 29 / 34 ea = 4985538888 / 17179869183 ; m = 8
s = 0 i = 31 internal angle = 8 / 34 ea = 4985538888 / 17179869183 ; m = 0
s = 0 i = 32 internal angle = 21 / 34 ea = 4985538888 / 17179869183 ; m = 0
s = 1 i = 33 internal angle = 34 / 34 ea = 4985538889 / 17179869183 ; m = 1
internal angle = 13/34
period = denominator of internal angle = 34
external angle as a decimal fraction = 4985538889/17179869183 = 4985538889 /( 2^34 - 1)
external angle as a binary rational (string) : 100101001001010010100100101001001/1111111111111111111111111111111111
external angle as a binary floating number in exponential form =0.1001010010010100101001001010010010100101001001010010100100101001*2^-1
external angle as a binary floating number in periodic form =0.(0100101001001010010100100101001)


### 34/89

Using GMP-5.1.3 with precision = 256 bits
internal angle = 34/89
period = denominator of internal angle = 89
external angle as a decimal fraction = 179622968672387565806504265/618970019642690137449562111
external angle as a binary rational (string) : 1001010010010100101001001010010010100101001001010010100100101001001010010100100101001001/11111111111111111111111111111111111111111111111111111111111111111111111111111111111111111
external angle as a binary floating number in exponential form =0.10010100100101001010010010100100101001010010010100101001001010010010100101001001010010010100101001001010010100100101001001010010100100101001010010010100100101001010010010100100101001010010010100101001001010010010100101001001010010100100101001001010010100101*2^-1
external angle as a binary floating number in periodic form =0.(01001010010010100101001001010010010100101001001010010100100101001001010010100100101001001)



### 1/128

uiIADenominator = 128
Using MPFR-4.0.2 with GMP-6.2.0 with precision = 200 bits
internal angle = 1/128
first external angle :
period = denominator of internal angle = 128
external angle as a decimal fraction = 1/340282366920938463463374607431768211455 = 1 /( 2^128 - 1)
External Angle as a floating point decimal number =  2.9387358770557187699218413430556141945553000604853132483972656175588435482079339324933425313850237034701685918031624270579715075034722882265605472939461496635969950989468319466936530037770580747746862e-39
external angle as a binary rational (string) : 1/11111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111
external angle as a binary floating number in exponential form =0.10000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000100000000000000000000000000000000000000000000000000000000000000000000000*2^-127
external angle as a binary floating number in periodic form =0.(00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000001)


### 89/268


Using GMP-5.1.3 with precision = 320 bits
internal angle = 89/268
period = denominator of internal angle = 268
external angle as a decimal fraction = 67754913930863876636420964942226524366713408170066250043659752013773168429311121/474284397516047136454946754595585670566993857190463750305618264096412179005177855
external angle as a binary rational (string) : 0010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010001
/1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111
external angle as a binary floating number in exponential form =0.10010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010001
001001001001001001001001001001001001001001001001001001*2^-2
external angle as a binary floating number in periodic form =
0.(0010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010010001)



G Pastor gave an example of external rays for which the resolution of the IEEE 754 is not sufficient :

$\theta _{268}^{-}=0.((001)^{88}0001)_{2}={\frac {67754913930863876636420964942226524366713408170066250043659752013773168429311121}{474284397516047136454946754595585670566993857190463750305618264096412179005177855}}$

$\theta _{268}^{+}=0.((001)^{88}0010)_{2}={\frac {67754913930863876636420964942226524366713408170066250043659752013773168429311122}{474284397516047136454946754595585670566993857190463750305618264096412179005177855}}$