XPost: comp.os.linux.misc, comp.os.linux.hardawe

To get a taste of creating a program that implements government
procedures, I challenge all to write a C function (or the equivalent
in some other language) to calculate the federal withholding tax.

The procedures are spelled out here:

https://www.irs.gov/pub/irs-pdf/p15t.pdf

Don't forget that we are dealing with MONEY, i.e. BIG BUCKS.
Therefore, the use of decimal floating point according to
IEEE 754-2008 is required:

https://en.wikipedia.org/wiki/IEEE_754-2008_revision


So get going!

If you get stuck just contact know-it-all rblowman. He is
a personal associate of Elon Musk.

Ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha!



--
Systemd: solving all the problems that you never knew you had.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Wed, 02 Apr 2025 11:16:42 +0000, Farley Flud <ff@linux.rocks> wrote in <pan$778d4$398e04fe$a51094fc$8dc6afee@linux.rocks>:

To get a taste of creating a program that implements government
procedures, I challenge all to write a C function (or the equivalent in
some other language) to calculate the federal withholding tax.

Why bother?

The last of your "programming challenges" I bothered with was using libgmp
for computing subfactorials. When I published my answer, not only
did you not acknowledge it, but you never published _your_ answer.

Supposition: I think you used my program for your homework.

Conclusion: You're a blow hard and probably a plagiarist.

--
-v System76 Thelio Mega v1.1 x86_64 NVIDIA RTX 3090 Ti
OS: Linux 6.14.0 Release: Mint 22.1 Mem: 258G
"Three out of five people aren't the other two."

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 2 Apr 2025 11:28:44 GMT, vallor wrote:

Why bother?

See first sentence.

Then enroll in a remedial reading program.

The last of your "programming challenges" I bothered with was using libgmp for computing subfactorials. When I published my answer, not only
did you not acknowledge it, but you never published _your_ answer.

Oh but I did publish:

https://www.rocksolidbbs.com/computers/article-flat.php?id=18113&group=comp.os.linux.advocacy#18113

My solution, however, being the ultimate, was so much better
than yours that you became extremely flustered and blotted
it out of your mind.





--
Systemd: solving all the problems that you never knew you had.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: comp.os.linux.misc

On 4/2/25 7:16 AM, Farley Flud wrote:

To get a taste of creating a program that implements government
procedures, I challenge all to write a C function (or the equivalent
in some other language) to calculate the federal withholding tax.

The procedures are spelled out here:

https://www.irs.gov/pub/irs-pdf/p15t.pdf

Don't forget that we are dealing with MONEY, i.e. BIG BUCKS.
Therefore, the use of decimal floating point according to
IEEE 754-2008 is required:

https://en.wikipedia.org/wiki/IEEE_754-2008_revision

So get going!

If you get stuck just contact know-it-all rblowman. He is
a personal associate of Elon Musk.

Ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha!

HA HA HA ! That's what so few seem to GET !
Huge, old, govt bureaucracy with a mass of
rules tweaked and re-tweaked to please various
pols and pressure groups. You need a good CPA
or dedicated student at minimum to kind-of
understand it.

PROBABLY similar issues in the UK/EU.

Oh, and it's always 'Kind Of' - even the
agencies themselves don't know how all the
rules work out anymore. SSA - awful. IRS,
way way beyond awful. AM hoping Trump nukes
the existing IRS. A new, plain, deal is now
BADLY needed. The bureaucracy was bad enough,
but all the 'fixes' created by pols and their
favorite donors ..... !!!

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Wed, 02 Apr 2025 11:16:42 +0000, Farley Flud wrote:

So get going!

I know that all of you astute and enthusiastic GNU/Linux advocates
[snicker, snicker] are just dying for a solution.

This project was only a hobby of mine that stemmed from wanting
to predict my paycheck deductions. The following code is from
2023 but can easily be updated to 2025 by changing the constants.
It is quite accurate in that all returned values agree exactly
with the published manual look-up tables.

Note that it uses the GNU/Linux libdfp decimal FP libraries:

https://github.com/libdfp/libdfp

===========================================
Fed W/H C Function
===========================================

// Fed Tax Witholding Function 2023

#define _GNU_SOURCE
#define __STDC_WANT_DEC_FP__
#include <fenv.h>
#include <math.h>
#include <stdio.h>
#include <stdlib.h>
extern _Decimal64 __bid_floatsidd (int i);

// subroutine fed_tax; pass period in per_type = 260, 52, 26, 24, 12, 4, 2
// pre2020, w4_2c = 1 if true, 0 if false

_Decimal64 fed_tax(_Decimal64 gross, int pre2020, int fstat, int fedallow, _Decimal64 fed_add, int per_type, int w4_2c, _Decimal64 w4_3, _Decimal64 w4_4a, _Decimal64 w4_4b, _Decimal64 w4_4c)
{

// These tables should be put in an include file
// Left Tables
// Going down, table-left-1st-column-1... table-left-2nd-column-1...

static _Decimal64 tbll11[8] = { 0.0DD, 14800.0DD, 36800.0DD, 104250.0DD, 205550.0DD, 379000.0DD, 477300.0DD, 708550.0DD};

static _Decimal64 tbll12[8] = { 0.0DD, 0.0DD, 2200.0DD, 10294.0DD, 32580.0DD, 74208.0DD, 105664.0DD, 188601.50DD };

static _Decimal64 tbll13[8] = { 0.0DD, 0.10DD, 0.12DD, 0.22DD, 0.24DD, 0.32DD, 0.35DD, 0.37DD };

static _Decimal64 tbll21[8] = { 0.0DD, 5250.0DD, 16250.0DD, 49975.0DD, 100625.0DD, 187350.0DD, 236500.0DD, 583375.0DD };

static _Decimal64 tbll22[8] = { 0.0DD, 0.0DD, 1100.0DD, 5147.0DD, 16260.0DD, 37104.0DD, 52832.0DD, 174238.25DD };

static _Decimal64 tbll23[8] = { 0.0DD, 0.10DD, 0.12DD, 0.22DD, 0.24DD, 0.32DD, 0.35DD, 0.37DD };

static _Decimal64 tbll31[8] = { 0.0DD, 12200.0DD, 27900.0DD, 72050.0DD, 107550.0DD, 194300.0DD, 243450.0DD, 590300.0DD };

static _Decimal64 tbll32[8] = { 0.0DD, 0.0DD, 1570.0DD, 6868.0DD, 14678.0DD, 35498.0DD, 51226.0DD, 172623.0DD };

static _Decimal64 tbll33[8] = { 0.0DD, 0.10DD, 0.12DD, 0.22DD, 0.24DD, 0.32DD, 0.35DD, 0.37DD };


// Right Tables

static _Decimal64 tblr11[8] = { 0.0DD, 13850.0DD, 24850.0DD, 58575.0DD, 109225.0DD, 195550.0DD, 245100.0DD, 360725.0DD };

static _Decimal64 tblr12[8] = { 0.0DD, 0.0DD, 1100.0DD, 5147.0DD, 16290.0DD, 37104.0DD, 52832.0DD, 93300.75DD};

static _Decimal64 tblr13[8] = { 0.0DD, 0.10DD, 0.12DD, 0.22DD, 0.24DD, 0.32DD, 0.35DD, 0.37DD };

static _Decimal64 tblr21[8] = { 0.0DD, 6925.0DD, 12425.0DD, 29288.0DD, 54613.0DD, 97975.0DD, 122550.0DD, 295988.0DD };

static _Decimal64 tblr22[8] = { 0.0DD, 0.0DD, 550.0DD, 2573.50DD, 8145.0DD, 18552.0DD, 26416.0DD, 87119.13DD };

static _Decimal64 tblr23[8] = { 0.0DD, 0.10DD, 0.12DD, 0.22DD, 0.24DD, 0.32DD, 0.35DD, 0.37DD };

static _Decimal64 tblr31[8] = { 0.0DD, 10400.0DD, 18250.0DD, 40325.0DD, 58075.0DD, 101450.0DD, 126025.0DD, 299450.0DD };

static _Decimal64 tblr32[8] = { 0.0DD, 0.0DD, 785.0DD, 3434.0DD, 7339.0DD, 17749.0DD, 25613.0DD, 86311.75DD };

static _Decimal64 tblr33[8] = { 0.0DD, 0.10DD, 0.12DD, 0.22DD, 0.24DD, 0.32DD, 0.35DD, 0.37DD };

static _Decimal64 stdallow=4300.0DD;
_Decimal64 *p1, *p2, *p3, adj_ann_gross, ann_gross, tentwh, fedwh, tmp1, tmp2; int index;

// set rounding mode to round half up fe_dec_setround(FE_DEC_TONEARESTFROMZERO);

//
// Step 1 -- determine adj_ann_gross
//

ann_gross = gross * __bid_floatsidd(per_type);

if (pre2020) {
adj_ann_gross = ann_gross - __bid_floatsidd(fedallow) * stdallow;
if (adj_ann_gross <= 0.0DD) { adj_ann_gross = 0.0DD; }
}
else { // not pre2020
tmp1 = ann_gross + w4_4a;
if (w4_2c) {
tmp2=0.0DD; }
else {
if (fstat==1) {
tmp2=12900.0DD;
}
else { tmp2=8600.0DD; }
}
tmp2=tmp2+w4_4b;
adj_ann_gross = tmp1-tmp2;
if (adj_ann_gross <= 0.0DD) { adj_ann_gross = 0.0DD; }
}

//
//Step 2 -- determine withholding
//

if (pre2020) {
// use left singe/married tables
switch(fstat)
{
case 0:
p1=tbll21; p2=tbll22; p3=tbll23; break;
case 1:
p1=tbll11; p2=tbll12; p3=tbll13; break;
case 2:
p1=tbll21; p2=tbll22; p3=tbll23; break;
}
}
else if ((!pre2020) && (!w4_2c)) {
//use right tables with fstat
switch(fstat)
{
case 0:
p1=tbll21; p2=tbll22; p3=tbll23; break;
case 1:
p1=tbll11; p2=tbll12; p3=tbll13; break;
case 2:
p1=tbll31; p2=tbll32; p3=tbll33; break;
}
}
else if ((!pre2020) && (w4_2c)) {
//use right tables with fstat
switch(fstat)
{
case 0:
p1=tblr21; p2=tblr22; p3=tblr23; break;
case 1:
p1=tblr11; p2=tblr12; p3=tblr13; break;
case 2:
p1=tblr31; p2=tblr32; p3=tblr33; break;
}
}
else {
fprintf(stdout, "\n%s\n", "Error on Table Select");
return 0.0DD;
}

index=0;
while(index<7)
{
if((adj_ann_gross>p1[index]) && (adj_ann_gross<=p1[index+1]))
break;
else
index++;
} tentwh=(p2[index]+p3[index]*(adj_ann_gross-p1[index]))/__bid_floatsidd(per_type);

tmp1=0.0DD;
if (!pre2020)
{ tmp1 = w4_3/__bid_floatsidd(per_type); }

fedwh = tentwh - tmp1;
if (fedwh<=0.0DD) { fedwh = 0.0DD; }

if (!pre2020)
{ fedwh = fedwh + w4_4c; } //defaults to 0
else { fedwh = fedwh + fed_add;
}
return quantized64(fedwh, 1.00DD);

} // end fed_tax





--
Systemd: solving all the problems that you never knew you had.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Wed, 02 Apr 2025 12:03:25 +0000, Farley Flud <ff@linux.rocks> wrote in <pan$b828b$8d4c48d3$557b76ca$8d04fc1c@linux.rocks>:

On 2 Apr 2025 11:28:44 GMT, vallor wrote:

Why bother?

See first sentence.

Then enroll in a remedial reading program.

The last of your "programming challenges" I bothered with was using
libgmp for computing subfactorials. When I published my answer, not
only did you not acknowledge it, but you never published _your_ answer.

Oh but I did publish:

https://www.rocksolidbbs.com/computers/article-flat.php?

id=18113&group=comp.os.linux.advocacy#18113

My solution, however, being the ultimate, was so much better than yours
that you became extremely flustered and blotted it out of your mind.

Riiiiight.

Mine formatted the numbers with commas to make them more readable.

So, nyah!

https://github.com/vallor/subfactorial

--
-v System76 Thelio Mega v1.1 x86_64 NVIDIA RTX 3090 Ti
OS: Linux 6.14.0 Release: Mint 22.1 Mem: 258G
"Some people like learning Eskimo, but I can't get Inuit."

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

XPost: comp.os.linux.misc

On 4/2/25 13:19, c186282 wrote:

On 4/2/25 7:16 AM, Farley Flud wrote:

To get a taste of creating a program that implements government
procedures, I challenge all to write a C function (or the equivalent
in some other language) to calculate the federal withholding tax.

The procedures are spelled out here:

https://www.irs.gov/pub/irs-pdf/p15t.pdf

Don't forget that we are dealing with MONEY, i.e. BIG BUCKS.
Therefore, the use of decimal floating point according to
IEEE 754-2008 is required:

https://en.wikipedia.org/wiki/IEEE_754-2008_revision


So get going!

If you get stuck just contact know-it-all rblowman.  He is
a personal associate of Elon Musk.

Ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha!

  HA HA HA ! That's what so few seem to GET !
  Huge, old, govt bureaucracy with a mass of
  rules tweaked and re-tweaked to please various
  pols and pressure groups. You need a good CPA
  or dedicated student at minimum to kind-of
  understand it.

  PROBABLY similar issues in the UK/EU.

  Oh, and it's always 'Kind Of' - even the
  agencies themselves don't know how all the
  rules work out anymore. SSA - awful.

A lot of the times the rules were made up by software developers, and
then became the official definition.

You sit in front of a half-arsed spec see some ambiguity, flip a coin,
and code up your arbitrary interpretation.

The real amusing thing is if you go back to the same company a decade
later, ask them to explain the rules, and you find an impressive, but
bogus, rationale has been constructed to justify why the code works the
way it does. Why your arbitrary decision was right.

But yes, people do underestimate the huge amount of complexity in old
systems. That complexity may be unnecessary, but it takes a huge amount
of time to understand it and simplify it. Plus, you don't get credit for
the skill and effort.

Develop a fancy mathematical algorithm, and you are a genius, even if it
is orders of magnitude easier to do. Hence, career wise, I quickly made
it my mission to stay away from old code bases.






IRS,

  way way beyond awful. AM hoping Trump nukes
  the existing IRS. A new, plain, deal is now
  BADLY needed. The bureaucracy was bad enough,
  but all the 'fixes' created by pols and their
  favorite donors ..... !!!

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 2 Apr 2025 13:52:35 GMT, vallor wrote:

Mine formatted the numbers with commas to make them more readable.

Oh yeah. That'll be very helpful.

The currently accepted groupings of 1000 (10^3), or decimal prefixes,
only go up to 10^30, which is called "quetta:"

https://en.wikipedia.org/wiki/Metric_prefix

The subfactorials pass 10^30 very quickly. For a 10,000 digit number
all those commas would only create confusion as the human brain cannot
grasp groupings beyond the giga or tera range.





--
Systemd: solving all the problems that you never knew you had.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)