From Newsgroup: comp.lang.fortran

xkcd: Day Counter
https://xkcd.com/3228/

Yes, floating point errors are the bane of my life ! Not so much now
with double precision but the single precision days were nightmares.

Explained at:
https://www.explainxkcd.com/wiki/index.php/3228:_Day_Counter

Whatever you do, do not select “Space Opera Mode” in the new mode drop down list. "Boat mode" made me a little queasy after a while. And
"modem mode" does not seem to do anything.

Lynn

--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

On 2026-04-04 03:40:15 +0000, Lynn McGuire said:

xkcd: Day Counter
https://xkcd.com/3228/

Yes, floating point errors are the bane of my life ! Not so much now
with double precision but the single precision days were nightmares.

Explained at:
https://www.explainxkcd.com/wiki/index.php/3228:_Day_Counter

Whatever you do, do not select “Space Opera Mode” in the new mode
drop down list. "Boat mode" made me a little queasy after a while.
And "modem mode" does not seem to do anything.

Lynn

Yep. It's a problem in Excel (still today!), which is relied on by
almost all businesses, thanks to the idiots at Microsoft not being able
to actually program. I discovered it by accident when I had some cross-checking calucations that said there was an error, when, after
quite a lot of time trying to find the error, turned out to in Excel
itself. :-\

Those of us with brains use Apple computers and the Numbers spreadsheet program does not have this issue.

--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

On Sat, 4 Apr 2026 16:46:58 +1300, Your Name wrote:

Those of us with brains use Apple computers and the Numbers
spreadsheet program does not have this issue.

Those of us with brains know how to use decimal or binary arithmetic
as appropriate to the problem at hand.

LibreOffice Calc:

= 0.1 + 0.1 + 0.1 - 0.3

evaluates to

0

Python, binary arithmetic:

0.1 + 0.1 + 0.1 - 0.3

evaluates to

5.551115123125783e-17

Python, decimal arithmetic:

>>> import decimal; D = decimal.Decimal
>>> D("0.1") + D("0.1") + D("0.1") - D("0.3")

evaluates to

Decimal('0.0')

--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

On 4/3/26 11:40 PM, Lynn McGuire wrote:

xkcd: Day Counter
   https://xkcd.com/3228/

Yes, floating point errors are the bane of my life !  Not so much now
with double precision but the single precision days were nightmares.

Explained at:
   https://www.explainxkcd.com/wiki/index.php/3228:_Day_Counter

Whatever you do, do not select “Space Opera Mode” in the new mode drop down list.

But if you do select Space Opera Mode, the box to change modes is active
as it scrolls by. And the rollover text still works too.

Tony



"Boat mode" made me a little queasy after a while.  And

"modem mode" does not seem to do anything.

Lynn

--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

On 4/4/2026 10:40 AM, Tony Nance wrote:

On 4/3/26 11:40 PM, Lynn McGuire wrote:

xkcd: Day Counter
    https://xkcd.com/3228/

Yes, floating point errors are the bane of my life !  Not so much now
with double precision but the single precision days were nightmares.

Explained at:
    https://www.explainxkcd.com/wiki/index.php/3228:_Day_Counter

Whatever you do, do not select “Space Opera Mode” in the new mode drop >> down list.

But if you do select Space Opera Mode, the box to change modes is active
as it scrolls by.

And if you let it get too "far away," the scrollwheel on your mouse can
bring it back. Not a problem at all.
--
Mark Jackson - https://mark-jackson.online/
I think it's a mistake — a very tempting mistake to make —
to take stock by looking at what we still have rather than
what we have already lost. - M. Gessen
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

LibreOffice Draw and LibreOffice Writer are buggy.
(S. HTTP://Gloucester.Insomnia247.NL/ fuer Kontaktdaten!)
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

On Sat, 4 Apr 2026 16:46:58 +1300, Your Name <YourName@YourISP.com>
wrote:

On 2026-04-04 03:40:15 +0000, Lynn McGuire said:

xkcd: Day Counter
https://xkcd.com/3228/

Yes, floating point errors are the bane of my life ! Not so much now
with double precision but the single precision days were nightmares.

Explained at:
https://www.explainxkcd.com/wiki/index.php/3228:_Day_Counter

Whatever you do, do not select “Space Opera Mode�? in the new mode
drop down list. "Boat mode" made me a little queasy after a while.
And "modem mode" does not seem to do anything.

Lynn

Yep. It's a problem in Excel (still today!), which is relied on by
almost all businesses, thanks to the idiots at Microsoft not being able
to actually program. I discovered it by accident when I had some >cross-checking calucations that said there was an error, when, after
quite a lot of time trying to find the error, turned out to in Excel
itself. :-\

In OOo, use of round(<value>;2) (or equivalent function and desired
precision) can help a lot.

Those of us with brains use Apple computers and the Numbers spreadsheet >program does not have this issue.

--
"Here lies the Tuscan poet Aretino,
Who evil spoke of everyone but God,
Giving as his excuse, 'I never knew him.'"
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

On 4/3/2026 10:40 PM, Lynn McGuire wrote:

xkcd: Day Counter
   https://xkcd.com/3228/

Yes, floating point errors are the bane of my life !  Not so much now
with double precision but the single precision days were nightmares.

Explained at:
   https://www.explainxkcd.com/wiki/index.php/3228:_Day_Counter

Whatever you do, do not select “Space Opera Mode” in the new mode drop down list.  "Boat mode" made me a little queasy after a while.  And
"modem mode" does not seem to do anything.

Lynn

When I selected modem mode it made the appropriate noises and the
picture slowly displayed as it downloaded.
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

On Sat, 4 Apr 2026 16:19:43 -0000 (UTC), Nioclás Pól Caileán de
Ghloucester wrote:

LibreOffice Draw and LibreOffice Writer are buggy.

Have you tried submitting bug reports?
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

In comp.lang.c++ Lawrence D’Oliveiro <ldo@nz.invalid> wrote: |----------------------------------------|
|"Have you tried submitting bug reports?"| |----------------------------------------|

Dear Mister D’Oliveiro,

I do not submit a bug report for LibreOffice. I almost never use
LibreOffice Draw. Opening the same file in different versions of
LibreOffice Writer results in different layouts.

I do not expect a programmer to remove these bugs. I can cope with
them.

Thanks for asking. Have a good week!
(S. HTTP://Gloucester.Insomnia247.NL/ fuer Kontaktdaten!)
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

Lynn McGuire wrote this screed in ALL-CAPS:

xkcd: Day Counter
https://xkcd.com/3228/

Yes, floating point errors are the bane of my life ! Not so much now
with double precision but the single precision days were nightmares.

Well these days we have long double.

Explained at:
https://www.explainxkcd.com/wiki/index.php/3228:_Day_Counter

Whatever you do, do not select “Space Opera Mode” in the new mode drop down list. "Boat mode" made me a little queasy after a while. And
"modem mode" does not seem to do anything.

Took me a little while to figure out where the mode dropdown was
when in Origami mode.
--
"One day I woke up and discovered that I was in love with tripe."
-- Tom Anderson
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

On Sun, 5 Apr 2026 12:50:55 -0000 (UTC), Nioclás Pól Caileán de
Ghloucester wrote:

I do not expect a programmer to remove these bugs. I can cope with
them.

Your posting came across as a complaint. Which is why I wondered why
you were bothering to mention them at all.
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

Chris Ahlstrom <OFeem1987@teleworm.us> writes:

Lynn McGuire wrote this screed in ALL-CAPS:

xkcd: Day Counter
https://xkcd.com/3228/

Yes, floating point errors are the bane of my life ! Not so much now
with double precision but the single precision days were nightmares.

Well these days we have long double.

C has had long double as a standard type since the original 1989
ANSI C standard. (K&R1, 1978, didn't have long double.)

But wider floating-point types don't eliminate rounding errors.

[...]
--
Keith Thompson (The_Other_Keith) Keith.S.Thompson+u@gmail.com
void Void(void) { Void(); } /* The recursive call of the void */
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

In comp.lang.c++ Lawrence D’Oliveiro <ldo@nz.invalid> wrote: |-------------------------------------------|
|"Your posting came across as a complaint." | |-------------------------------------------|

A posting by me about LibreOffice is a complaint. It is a complaint
without deserving improvements to gratis software to which I do not
contribute. So I do not burden the LibreOffice donaters with a
demand.

|-------------------------------------------|
|"Which is why I wondered why |
|you were bothering to mention them at all."| |-------------------------------------------|

A good trait of LibreOffice Calc is mentioned, so I reacted via a
remark about problematic parts of LibreOffice.
(S. HTTP://Gloucester.Insomnia247.NL/ fuer Kontaktdaten!)
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

On Sun, 5 Apr 2026 23:06:32 -0000 (UTC), Nioclás Pól Caileán de
Ghloucester wrote:

A good trait of LibreOffice Calc is mentioned, so I reacted via a
remark about problematic parts of LibreOffice.

I wonder why you felt the need to?
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

Keith Thompson <Keith.S.Thompson+u@gmail.com> schrieb:

Chris Ahlstrom <OFeem1987@teleworm.us> writes:

Lynn McGuire wrote this screed in ALL-CAPS:

xkcd: Day Counter
https://xkcd.com/3228/

Yes, floating point errors are the bane of my life ! Not so much now
with double precision but the single precision days were nightmares.

Well these days we have long double.

C has had long double as a standard type since the original 1989
ANSI C standard. (K&R1, 1978, didn't have long double.)

But wider floating-point types don't eliminate rounding errors.

Nor does "long double" have to be any more accurate thean
"double". C's type system leaves something to be desired
in that respect.

Fortran showed how it's done with its kind numbers and
SELECTED_REAL_KIND function - you ask for a certain minimum
precision, and the compiler either gives it to you, or your program
fails at compile time.
--
This USENET posting was made without artificial intelligence,
artificial impertinence, artificial arrogance, artificial stupidity,
artificial flavorings or artificial colorants.
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

Thomas Koenig <tkoenig@netcologne.de> writes:

Keith Thompson <Keith.S.Thompson+u@gmail.com> schrieb:

Chris Ahlstrom <OFeem1987@teleworm.us> writes:

Lynn McGuire wrote this screed in ALL-CAPS:

xkcd: Day Counter
https://xkcd.com/3228/

Yes, floating point errors are the bane of my life ! Not so much now >>>> with double precision but the single precision days were nightmares.

Well these days we have long double.

C has had long double as a standard type since the original 1989
ANSI C standard. (K&R1, 1978, didn't have long double.)

But wider floating-point types don't eliminate rounding errors.

Nor does "long double" have to be any more accurate thean
"double". C's type system leaves something to be desired
in that respect.

Fortran showed how it's done with its kind numbers and
SELECTED_REAL_KIND function - you ask for a certain minimum
precision, and the compiler either gives it to you, or your program
fails at compile time.

You can do the same thing in C or C++ with a compile-time test.

#include <float.h>

#if LDBL_DIG < 18
#error "long double doesn't have enough precision"
#endif
--
Keith Thompson (The_Other_Keith) Keith.S.Thompson+u@gmail.com
void Void(void) { Void(); } /* The recursive call of the void */
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

Your Name <YourName@YourISP.com> wrote:

On 2026-04-04 03:40:15 +0000, Lynn McGuire said:

xkcd: Day Counter
https://xkcd.com/3228/

Yes, floating point errors are the bane of my life ! Not so much now
with double precision but the single precision days were nightmares.

Explained at:
https://www.explainxkcd.com/wiki/index.php/3228:_Day_Counter

Whatever you do, do not select “Space Opera Mode” in the new mode >> drop down list. "Boat mode" made me a little queasy after a while.
And "modem mode" does not seem to do anything.

Lynn

Yep. It's a problem in Excel (still today!), which is relied on by
almost all businesses, thanks to the idiots at Microsoft not being able
to actually program. I discovered it by accident when I had some cross-checking calucations that said there was an error, when, after
quite a lot of time trying to find the error, turned out to in Excel
itself. :-\

Those of us with brains use Apple computers and the Numbers spreadsheet program does not have this issue.

I encountered the following problem with Excel: if you use Excel’s
Automation API (based on the Microsoft COM framework), you can access everything inside an Excel workbook programmatically. Also any charts that
have been added to a spreadsheet. However, if you‘re trying to iterate through the series of the graph and try to access the formatting of the
series (are there lines between the points of the series, if so which
color, thickness, etc.), you‘ll get an error when you’re accessing a series that is not shown in the legend of the chart because the legend is too
small. If you increase the size of the chart to an ridiculous amount (so
that you can be sure that the legend will be big enough to show every
series of the chart), you can access everything without a problem.

If a program contains such incredible errors, I would keep my fingers from
it. Although I haven’t tried to do the same thing with Open/LibreOffice‘s automation API, I‘d vouch that such an error would not occur.

And yes, us professionals prefer Apple over MS, even though Visual Studio
has some features that I sorely miss in other IDEs.

Regards
Stuart

--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

Lynn McGuire <lynnmcguire5@gmail.com> wrote at 03:40 this Saturday (GMT):

xkcd: Day Counter
https://xkcd.com/3228/

Yes, floating point errors are the bane of my life ! Not so much now
with double precision but the single precision days were nightmares.

Explained at:
https://www.explainxkcd.com/wiki/index.php/3228:_Day_Counter

Whatever you do, do not select “Space Opera Mode” in the new mode drop down list. "Boat mode" made me a little queasy after a while. And
"modem mode" does not seem to do anything.

Lynn

I think modem mode actually slows down the transfer rate, cute idea but
it freezes my browser. Boat Mode is pretty nostalgic tho, reminds me of WaveHello :)
--
user <candycane> is generated from /dev/urandom
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

On Mon, 6 Apr 2026 16:10:08 +0200, Stuart Redmann <DerTopper@web.de>
wrote:

Your Name <YourName@YourISP.com> wrote:

On 2026-04-04 03:40:15 +0000, Lynn McGuire said:

xkcd: Day Counter
https://xkcd.com/3228/

Yes, floating point errors are the bane of my life ! Not so much now
with double precision but the single precision days were nightmares.

Explained at:
https://www.explainxkcd.com/wiki/index.php/3228:_Day_Counter

Whatever you do, do not select �??Space Opera Mode�?? in the new mode
drop down list. "Boat mode" made me a little queasy after a while.
And "modem mode" does not seem to do anything.

Lynn

Yep. It's a problem in Excel (still today!), which is relied on by
almost all businesses, thanks to the idiots at Microsoft not being able
to actually program. I discovered it by accident when I had some
cross-checking calucations that said there was an error, when, after
quite a lot of time trying to find the error, turned out to in Excel
itself. :-\

Those of us with brains use Apple computers and the Numbers spreadsheet
program does not have this issue.

I encountered the following problem with Excel: if you use Excel�s
Automation API (based on the Microsoft COM framework), you can access >everything inside an Excel workbook programmatically. Also any charts that >have been added to a spreadsheet. However, if you�re trying to iterate >through the series of the graph and try to access the formatting of the >series (are there lines between the points of the series, if so which
color, thickness, etc.), you�ll get an error when you�re accessing a series >that is not shown in the legend of the chart because the legend is too
small. If you increase the size of the chart to an ridiculous amount (so
that you can be sure that the legend will be big enough to show every
series of the chart), you can access everything without a problem.

If a program contains such incredible errors, I would keep my fingers from >it. Although I haven�t tried to do the same thing with Open/LibreOffice�s >automation API, I�d vouch that such an error would not occur.

Theoretically speaking, would that allow creating a program that can
convert an ODT file containing bits of spreadsheet (copied and pasted
from an ODS file) to a text file that actually includes the spreasheet
bits instead of ignoring them?

And yes, us professionals prefer Apple over MS, even though Visual Studio
has some features that I sorely miss in other IDEs.

--
"Here lies the Tuscan poet Aretino,
Who evil spoke of everyone but God,
Giving as his excuse, 'I never knew him.'"
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

Thomas Koenig <tkoenig@NetCologne.De> hat geschrieben: |-------------------------------------------------------------------|
|"Nor does "long double" have to be any more accurate thean | |"double". C's type system leaves something to be desired |
|in that respect. |
| |
|Fortran showed how it's done with its kind numbers and | |SELECTED_REAL_KIND function - you ask for a certain minimum | |precision, and the compiler either gives it to you, or your program|
|fails at compile time." | |-------------------------------------------------------------------|

I recommend using Ada instead.
(S. HTTP://Gloucester.Insomnia247.NL/ fuer Kontaktdaten!)
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

Keith Thompson <Keith.S.Thompson+u@gmail.com> schrieb:

Thomas Koenig <tkoenig@netcologne.de> writes:

Keith Thompson <Keith.S.Thompson+u@gmail.com> schrieb:

Chris Ahlstrom <OFeem1987@teleworm.us> writes:

Lynn McGuire wrote this screed in ALL-CAPS:

xkcd: Day Counter
https://xkcd.com/3228/

Yes, floating point errors are the bane of my life ! Not so much now >>>>> with double precision but the single precision days were nightmares.

Well these days we have long double.

C has had long double as a standard type since the original 1989
ANSI C standard. (K&R1, 1978, didn't have long double.)

But wider floating-point types don't eliminate rounding errors.

Nor does "long double" have to be any more accurate thean
"double". C's type system leaves something to be desired
in that respect.

Fortran showed how it's done with its kind numbers and
SELECTED_REAL_KIND function - you ask for a certain minimum
precision, and the compiler either gives it to you, or your program
fails at compile time.

You can do the same thing in C or C++ with a compile-time test.

#include <float.h>

#if LDBL_DIG < 18
#error "long double doesn't have enough precision"
#endif

Not quite the same. SELECTED_REAL_KIND gives you the
_minimum_ real type for your requirements.
--
This USENET posting was made without artificial intelligence,
artificial impertinence, artificial arrogance, artificial stupidity,
artificial flavorings or artificial colorants.
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

On Mon, 6 Apr 2026 15:54:55 -0000 (UTC), Nioclás Pól Caileán de
Ghloucester wrote:

I recommend using Ada instead.

Unless Ada offers the option for decimal arithmetic, that’s not likely
to help.
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

Thomas Koenig <tkoenig@netcologne.de> writes:

Keith Thompson <Keith.S.Thompson+u@gmail.com> schrieb:

Thomas Koenig <tkoenig@netcologne.de> writes:

[...]

Fortran showed how it's done with its kind numbers and
SELECTED_REAL_KIND function - you ask for a certain minimum
precision, and the compiler either gives it to you, or your program
fails at compile time.

You can do the same thing in C or C++ with a compile-time test.

#include <float.h>

#if LDBL_DIG < 18
#error "long double doesn't have enough precision"
#endif

Not quite the same. SELECTED_REAL_KIND gives you the
_minimum_ real type for your requirements.

You can do that in C or C++ too, though not as cleanly. With a
series of #if directives, you can select the minimal floating-point
type (out of float, double, long double) that satisfies some
requirement. But I don't recall ever seeing code that does this.

This is rapidly going off-topic for most of the newsgroups this is
posted to, so I won't go into more detail.
--
Keith Thompson (The_Other_Keith) Keith.S.Thompson+u@gmail.com
void Void(void) { Void(); } /* The recursive call of the void */
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

On 4/5/2026 6:50 AM, Nioclás Pól Caileán de Ghloucester wrote:

In comp.lang.c++ Lawrence D’Oliveiro <ldo@nz.invalid> wrote: |----------------------------------------|
|"Have you tried submitting bug reports?"| |----------------------------------------|

Dear Mister D’Oliveiro,

I do not submit a bug report for LibreOffice. I almost never use
LibreOffice Draw. Opening the same file in different versions of
LibreOffice Writer results in different layouts.

I do not expect a programmer to remove these bugs. I can cope with
them.

Thanks for asking. Have a good week!
(S. HTTP://Gloucester.Insomnia247.NL/ fuer Kontaktdaten!)

I believe you would be doing the maintainers (and other users) of
LibreOffice a favor by calling attention to bugs, especially if you can provide a clear description of the bug and a small test case. A better
product is a win for everyone; it might even make your life easier some day.

Louis
--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

Le 06/04/2026 à 22:40, Lawrence D’Oliveiro a écrit :

On Mon, 6 Apr 2026 15:54:55 -0000 (UTC), Nioclás Pól Caileán de Ghloucester wrote:

I recommend using Ada instead.

Unless Ada offers the option for decimal arithmetic, that’s not likely
to help.??

Ada HAS decimal arithmetic, with specifiable accuracy, as well as
floating point arithmetic, with specifiable accuracy.
In addition, the guarantee of accuracy extends to the whole mathematical library (with the Numerics annex)

--- Synchronet 3.21f-Linux NewsLink 1.2

From Newsgroup: comp.lang.fortran

In comp.lang.c++ Lawrence D’Oliveiro <ldo@nz.invalid> wrote: |-----------------------------------------------------|
|"> LibreOffice Draw and LibreOffice Writer are buggy.|
| |
|Have you tried submitting bug reports?" | |-----------------------------------------------------|

I do use LibreOffice Writer's feature to submit crashes' reports.
(S. HTTP://Gloucester.Insomnia247.NL/ fuer Kontaktdaten!)
--- Synchronet 3.21f-Linux NewsLink 1.2