On 10/21/2023 7:51 PM, bill wrote:
> On 10/21/2023 7:44 PM, Arne Vajhøj wrote:
>> On 10/21/2023 7:22 PM, bill wrote:
>>> On 10/21/2023 5:24 PM, Johnny Billquist wrote:
>>>> On 2023-10-21 17:39, bill wrote:
>>>>> On 10/21/2023 10:55 AM, Arne Vajhøj wrote:
>>>>>> Are you telling me that you have never seen i, j, ...
>>>>>> used as variable in a for loop in C?
>>>>>>
>>>>>
>>>>> Probably, but if someone is going to use a single letter variable
>>>>> for a loop index more likely would be a, b, x, y, z.  But I'm sure
>>>>> an i would show up more for it being the first letter of the word
>>>>> index than any influence from Fortran.  Can't say I have ever seen
>>>>> j used.
>>>>
>>>> i and j as loop variables probably beats all other letters combined.
>>>>
>>>> If you haven't seen that, then I really wonder what code you have
>>>> looked at.
>>>>
>>>
>>> My early days doing production programming (In COBOL, Fortran, Pascal, C
>>> and a lot of different assemblers) was all done in an environment with
>>> very strict programming standards.  Probably explains a lot about my
>>> style and why I don't have problems with things like over-runs and
>>> bounds violations.  :-)
>>
>> C coding standards often allow or even encourage i and j.
>>
>> A few random examples:
>>
>> Linux kernel coding style:
>>
>> https://www.kernel.org/doc/html/v4.10/process/coding-style.html
>>
>> <quote>
>> LOCAL variable names should be short, and to the point. If you have
>> some random integer loop counter, it should probably be called i.
>> Calling it loop_counter is non-productive, if there is no chance of it
>> being mis-understood. Similarly, tmp can be just about any type of
>> variable that is used to hold a temporary value.
>> </quote>
>>
>> FreeBSD style guide:
>>
>> https://man.freebsd.org/cgi/man.cgi?style(9)
>>
>> <quote>
>> A for loop may declare and initialize its counting variable.
>>
>>             for (int    i = 0; i < 15; i++) {
>> </quote>
>>
>> Gnome coding style:
>>
>> https://developer.gnome.org/documentation/guidelines/programming/coding-style.html
>>
>> <quote>
>> for (i = 0; i < num_elements; i++) {
>> </quote>
>>
>
> To each his own.  I was taught with a much stricter standard
> and I think  it has done me well over my career.  I will admit
> that I have used i and variants like i1 and i2 for loops when
> writing something quick and dirty that I will only be using
> myself, but even then I know deep inside that I am doing
> something basically wrong.  :-)

Stricter?

Is DoD Ada guidelines strict enough for you?

https://apps.dtic.mil/sti/tr/pdf/ADA241772.pdf

page 84 has an example with a for loop using I.

https://www.adaic.org/resources/add_content/docs/95style/95style.pdf\

<quote>
To iterate over all elements of an array:
for I in Array_Name'Range loop
</quote>

Arne

On 10/21/2023 8:04 PM, Arne Vajhøj wrote:
> On 10/21/2023 7:51 PM, bill wrote:
>> On 10/21/2023 7:44 PM, Arne Vajhøj wrote:
>>> On 10/21/2023 7:22 PM, bill wrote:
>>>> On 10/21/2023 5:24 PM, Johnny Billquist wrote:
>>>>> On 2023-10-21 17:39, bill wrote:
>>>>>> On 10/21/2023 10:55 AM, Arne Vajhøj wrote:
>>>>>>> Are you telling me that you have never seen i, j, ...
>>>>>>> used as variable in a for loop in C?
>>>>>>>
>>>>>>
>>>>>> Probably, but if someone is going to use a single letter variable
>>>>>> for a loop index more likely would be a, b, x, y, z.  But I'm sure
>>>>>> an i would show up more for it being the first letter of the word
>>>>>> index than any influence from Fortran.  Can't say I have ever seen
>>>>>> j used.
>>>>>
>>>>> i and j as loop variables probably beats all other letters combined.
>>>>>
>>>>> If you haven't seen that, then I really wonder what code you have
>>>>> looked at.
>>>>>
>>>>
>>>> My early days doing production programming (In COBOL, Fortran,
>>>> Pascal, C
>>>> and a lot of different assemblers) was all done in an environment with
>>>> very strict programming standards.  Probably explains a lot about my
>>>> style and why I don't have problems with things like over-runs and
>>>> bounds violations.  :-)
>>>
>>> C coding standards often allow or even encourage i and j.
>>>
>>> A few random examples:
>>>
>>> Linux kernel coding style:
>>>
>>> https://www.kernel.org/doc/html/v4.10/process/coding-style.html
>>>
>>> <quote>
>>> LOCAL variable names should be short, and to the point. If you have
>>> some random integer loop counter, it should probably be called i.
>>> Calling it loop_counter is non-productive, if there is no chance of
>>> it being mis-understood. Similarly, tmp can be just about any type of
>>> variable that is used to hold a temporary value.
>>> </quote>
>>>
>>> FreeBSD style guide:
>>>
>>> https://man.freebsd.org/cgi/man.cgi?style(9)
>>>
>>> <quote>
>>> A for loop may declare and initialize its counting variable.
>>>
>>>             for (int    i = 0; i < 15; i++) {
>>> </quote>
>>>
>>> Gnome coding style:
>>>
>>> https://developer.gnome.org/documentation/guidelines/programming/coding-style.html
>>>
>>> <quote>
>>> for (i = 0; i < num_elements; i++) {
>>> </quote>
>>>
>>
>> To each his own.  I was taught with a much stricter standard
>> and I think  it has done me well over my career.  I will admit
>> that I have used i and variants like i1 and i2 for loops when
>> writing something quick and dirty that I will only be using
>> myself, but even then I know deep inside that I am doing
>> something basically wrong.  :-)
>
> Stricter?
>
> Is DoD Ada guidelines strict enough for you?
>
> https://apps.dtic.mil/sti/tr/pdf/ADA241772.pdf
>
> page 84 has an example with a for loop using I.
>
> https://www.adaic.org/resources/add_content/docs/95style/95style.pdf\
>
> <quote>
> To iterate over all elements of an array:
> for I in Array_Name'Range loop
> </quote>

OK. Well, my early programming days were also with DOD and
later with DOD Contractors. And they didn't believe in that
kind of programming. I see the number 95 in there. I assume
that's a year. Looks like someone got really lazy after I
left. It was 1988 when I left that world and went into academia.
I'll take the old way over the new way any day.

bill

On 10/21/2023 8:17 PM, bill wrote:
> On 10/21/2023 8:04 PM, Arne Vajhøj wrote:
>> On 10/21/2023 7:51 PM, bill wrote:
>>> On 10/21/2023 7:44 PM, Arne Vajhøj wrote:
>>>> On 10/21/2023 7:22 PM, bill wrote:
>>>>> On 10/21/2023 5:24 PM, Johnny Billquist wrote:
>>>>>> On 2023-10-21 17:39, bill wrote:
>>>>>>> On 10/21/2023 10:55 AM, Arne Vajhøj wrote:
>>>>>>>> Are you telling me that you have never seen i, j, ...
>>>>>>>> used as variable in a for loop in C?
>>>>>>>>
>>>>>>>
>>>>>>> Probably, but if someone is going to use a single letter variable
>>>>>>> for a loop index more likely would be a, b, x, y, z.  But I'm sure
>>>>>>> an i would show up more for it being the first letter of the word
>>>>>>> index than any influence from Fortran.  Can't say I have ever seen
>>>>>>> j used.
>>>>>>
>>>>>> i and j as loop variables probably beats all other letters combined.
>>>>>>
>>>>>> If you haven't seen that, then I really wonder what code you have
>>>>>> looked at.
>>>>>>
>>>>>
>>>>> My early days doing production programming (In COBOL, Fortran,
>>>>> Pascal, C
>>>>> and a lot of different assemblers) was all done in an environment with
>>>>> very strict programming standards.  Probably explains a lot about my
>>>>> style and why I don't have problems with things like over-runs and
>>>>> bounds violations.  :-)
>>>>
>>>> C coding standards often allow or even encourage i and j.
>>>>
>>>> A few random examples:
>>>>
>>>> Linux kernel coding style:
>>>>
>>>> https://www.kernel.org/doc/html/v4.10/process/coding-style.html
>>>>
>>>> <quote>
>>>> LOCAL variable names should be short, and to the point. If you have
>>>> some random integer loop counter, it should probably be called i.
>>>> Calling it loop_counter is non-productive, if there is no chance of
>>>> it being mis-understood. Similarly, tmp can be just about any type
>>>> of variable that is used to hold a temporary value.
>>>> </quote>
>>>>
>>>> FreeBSD style guide:
>>>>
>>>> https://man.freebsd.org/cgi/man.cgi?style(9)
>>>>
>>>> <quote>
>>>> A for loop may declare and initialize its counting variable.
>>>>
>>>>             for (int    i = 0; i < 15; i++) {
>>>> </quote>
>>>>
>>>> Gnome coding style:
>>>>
>>>> https://developer.gnome.org/documentation/guidelines/programming/coding-style.html
>>>>
>>>> <quote>
>>>> for (i = 0; i < num_elements; i++) {
>>>> </quote>
>>>>
>>>
>>> To each his own.  I was taught with a much stricter standard
>>> and I think  it has done me well over my career.  I will admit
>>> that I have used i and variants like i1 and i2 for loops when
>>> writing something quick and dirty that I will only be using
>>> myself, but even then I know deep inside that I am doing
>>> something basically wrong.  :-)
>>
>> Stricter?
>>
>> Is DoD Ada guidelines strict enough for you?
>>
>> https://apps.dtic.mil/sti/tr/pdf/ADA241772.pdf
>>
>> page 84 has an example with a for loop using I.
>>
>> https://www.adaic.org/resources/add_content/docs/95style/95style.pdf\
>>
>> <quote>
>> To iterate over all elements of an array:
>> for I in Array_Name'Range loop
>> </quote>
>
> OK.  Well, my early programming days were also with DOD and
> later with DOD Contractors.  And they didn't believe in that
> kind of programming.  I see the number 95 in there. I assume
> that's a year.  Looks like someone got really lazy after I
> left.  It was 1988 when I left that world and went into academia.
> I'll take the old way over the new way any day.

The last link is indeed from 1995.

The first link is from 1989.

Arne

On Saturday, October 21, 2023 at 7:59:03 AM UTC-7, Arne Vajhøj wrote:
> On 10/21/2023 1:45 AM, gah4 wrote:
> > Many Unix programs work differently when stdout is a terminal instead
> > of a file. That often makes sense. Also, many people are used to them
> > working that way, and know what to do about it.

> If being a terminal activates some fancy color highlightning
> then fine.

> I still don't like changing behavior.
>
> It is not that hard to add /NOLIST and /NOMAP but it
> just irritates me.
Well for one example, more does its "more" thing only on a terminal,
otherwise it works more like cat.

Convenient when you used more, and then want to redirect to a file.
You don't have to remove the more, just add the redirect.

I suspect interactive editors will also check for a terminal.

Running vi from batch doesn't seem very useful.

On Saturday, October 21, 2023 at 7:55:17 AM UTC-7, Arne Vajhøj wrote:
> On 10/21/2023 8:45 AM, bill wrote:
> > On 10/21/2023 1:45 AM, gah4 wrote:

(snip)

> >> even though it used to be true in C, isn't any more, it does

> > When did C ever have "default typing" like Fortran?

> I don't think C ever allowed undeclared variables
> giving them a type.

> C does default return type of functions to int. It is
> horrible, but it does.

K&R C variables (and functions, as you note) default to int.

int is still the default type if you declare a variable without a type.

auto i;
static j;

declare int variables with the appropriate attribute.
(auto is the default, but you can still declare it.)

I believe short and long were originally modifiers, such
that short is short for short int, and long is short for long int.

And void wasn't added until the ANSI standard. Before that, it was
usual for functions to be int, and you weren't required to actually
return any value.

On 10/21/2023 7:31 PM, bill wrote:
> On 10/21/2023 1:48 PM, Arne Vajhøj wrote:
>> On 10/21/2023 12:12 PM, Arne Vajhøj wrote:
>>> On 10/21/2023 11:39 AM, bill wrote:
>>>> On 10/21/2023 10:55 AM, Arne Vajhøj wrote:
>>>>> On 10/21/2023 8:45 AM, bill wrote:
>>>>>> On 10/21/2023 1:45 AM, gah4 wrote:
>>>>>>> seem that many like to use i, j, k, l, m, and n for the
>>>>>>> first letter of integer variables in C and other languages,
>>>>>>> following the Fortran convention of over 60 years ago.
>>>>>>
>>>>>> I have also never seen this. Except for entries in the
>>>>>> obfuscated C contest the norm for C program, in my experience,
>>>>>> has always been to use meaningful names for variables and not
>>>>>> some old convention from a different language.
>>>>>
>>>>> Hmmm.
>>>>>
>>>>> Are you telling me that you have never seen i, j, ...
>>>>> used as variable in a for loop in C?
>>>>
>>>> Probably, but if someone is going to use a single letter variable
>>>> for a loop index more likely would be a, b, x, y, z. But I'm sure
>>>> an i would show up more for it being the first letter of the word
>>>> index than any influence from Fortran. Can't say I have ever seen
>>>> j used.
>>>
>>> My impression is that:
>>>
>>> for(i = 0, ...
>>> for(int i = 0, ...
>>>
>>> outnumber:
>>>
>>> for(a = 0, ...
>>> for(int a = 0, ...
>>>
>>> by a factor 1000.
>>>
>>> :-)
>>
>> The gentlemen K & R also use i and j in the C book.
>>
>> And the gentleman Stroustrup also use i in the C++ book.
>
> What can I say. The places where I worked frowned on (read
> that as forbid) single letter variables for any use at any
> time. They also frowned on the concept of IMPLICIT. Define
> everything and explain it in comments and the documentation.
>
> bill
>
>

As they were signing the paychecks, they were entitled to an opinion. But other
opinions are valid.

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

On Saturday, October 21, 2023 at 8:40:02 AM UTC-7, bill wrote:
> On 10/21/2023 10:55 AM, Arne Vajhøj wrote:

(snip)

> > Are you telling me that you have never seen i, j, ...
> > used as variable in a for loop in C?

> Probably, but if someone is going to use a single letter variable
> for a loop index more likely would be a, b, x, y, z. But I'm sure
> an i would show up more for it being the first letter of the word
> index than any influence from Fortran. Can't say I have ever seen
> j used.

Mathematics notation for summation commonly uses i and j
as the index variables.

Since I didn't learn algebra and calculus until after Fortran,
I can't say which got it first, but I suspect it was math.

I do have a Fortran program with 26 one letter integer variables, though.

> I have also never seen this. [...]

> [...] Can't say I have ever seen j used.

> [...] I was taught with a much stricter standard [...]

If the only programs which you have ever seen were written by you,
then a qualifier like "I have never seen" does not mean what a typical
reader expects it to mean. To avoid wasting so much time of so many
people, you might consider using instead a qualifier like, say, "I never
do". (It's still foolish, but might promote fewer senseless arguments.)

For a good time, try commands like the following in the main
directory of almost any C project (written my neurotypical people):

search /statistics [...]*.c " i,", "i;"
search /statistics [...]*.c " j,", "j;"

On 2023-10-22 01:22, bill wrote:
> On 10/21/2023 5:24 PM, Johnny Billquist wrote:
>> On 2023-10-21 17:39, bill wrote:
>>> On 10/21/2023 10:55 AM, Arne Vajhøj wrote:
>>>> Are you telling me that you have never seen i, j, ...
>>>> used as variable in a for loop in C?
>>>>
>>>
>>> Probably, but if someone is going to use a single letter variable
>>> for a loop index more likely would be a, b, x, y, z.  But I'm sure
>>> an i would show up more for it being the first letter of the word
>>> index than any influence from Fortran.  Can't say I have ever seen
>>> j used.
>>
>> i and j as loop variables probably beats all other letters combined.
>>
>> If you haven't seen that, then I really wonder what code you have
>> looked at.
>>
>
> My early days doing production programming (In COBOL, Fortran, Pascal, C
> and a lot of different assemblers) was all done in an environment with
> very strict programming standards.  Probably explains a lot about my
> style and why I don't have problems with things like over-runs and
> bounds violations.  :-)

It's unlikely that avoiding single letter variable names for small,
simple loops, helped you avoid over-runs and bounds violations.

It's like using a register for a simple loop in assembler. Sure, you can
do the loops with a value somewhere in memory as well, but it's just
making for much more code, and less readable code at that.

I've worked in many places, and have faced many different coding styles
and requirements. Some of them really complex and elaborate. But I've
not encountered a single one that didn't recognize that for some
situations, and single letter variable name for small loops just makes
more sense to leave at that.

Johnny

On 2023-10-22 02:17, bill wrote:
> On 10/21/2023 8:04 PM, Arne Vajhøj wrote:
>>
>> Stricter?
>>
>> Is DoD Ada guidelines strict enough for you?
>>
>> https://apps.dtic.mil/sti/tr/pdf/ADA241772.pdf
>>
>> page 84 has an example with a for loop using I.
>>
>> https://www.adaic.org/resources/add_content/docs/95style/95style.pdf\
>>
>> <quote>
>> To iterate over all elements of an array:
>> for I in Array_Name'Range loop
>> </quote>
>
> OK.  Well, my early programming days were also with DOD and
> later with DOD Contractors.  And they didn't believe in that
> kind of programming.  I see the number 95 in there. I assume
> that's a year.  Looks like someone got really lazy after I
> left.  It was 1988 when I left that world and went into academia.
> I'll take the old way over the new way any day.

Yes, Ada95 came out in 95. Before that you had Ada83. I believe the
guidelines on this one didn't change.
(I actually have done paid work in both Ada83 and Ada95.)

Johnny

On 2023-10-22 03:25, gah4 wrote:
> On Saturday, October 21, 2023 at 7:55:17 AM UTC-7, Arne Vajhøj wrote:
>> On 10/21/2023 8:45 AM, bill wrote:
>>> On 10/21/2023 1:45 AM, gah4 wrote:
>
> (snip)
>
>>>> even though it used to be true in C, isn't any more, it does
>
>>> When did C ever have "default typing" like Fortran?
>
>> I don't think C ever allowed undeclared variables
>> giving them a type.
>
>> C does default return type of functions to int. It is
>> horrible, but it does.
>
> K&R C variables (and functions, as you note) default to int.
>
> int is still the default type if you declare a variable without a type.
>
> auto i;
> static j;
>
> declare int variables with the appropriate attribute.
> (auto is the default, but you can still declare it.)

Sure. But for variables, they have to be declared. Even if it's just
"auto i;". You cannot use a variable that have not been declared in any way.

> And void wasn't added until the ANSI standard. Before that, it was
> usual for functions to be int, and you weren't required to actually
> return any value.

Well. They still return a value. But you've always been allowed to just
ignore the returned value of a function in C. Compilers these days
usually issue at least a warning if a non-void function don't explicitly
return a value, but I think it's still technically not disallowed (but I
could be wrong on that one...).

Johnny

On 10/21/2023 9:13 PM, gah4 wrote:
> On Saturday, October 21, 2023 at 7:59:03 AM UTC-7, Arne Vajhøj wrote:
>> On 10/21/2023 1:45 AM, gah4 wrote:
>>> Many Unix programs work differently when stdout is a terminal instead
>>> of a file. That often makes sense. Also, many people are used to them
>>> working that way, and know what to do about it.
>
>> If being a terminal activates some fancy color highlightning
>> then fine.
>
>> I still don't like changing behavior.
>>
>> It is not that hard to add /NOLIST and /NOMAP but it
>> just irritates me.
>
> Well for one example, more does its "more" thing only on a terminal,
> otherwise it works more like cat.
>
> Convenient when you used more, and then want to redirect to a file.
> You don't have to remove the more, just add the redirect.
>
> I suspect interactive editors will also check for a terminal.
>
> Running vi from batch doesn't seem very useful.

Actually, it works quite well (Vim warns about not having a terminal
but then, Vim isn't vi :-).

sed is better for the task but some people know vi better than sed.

bill

On 10/21/2023 10:59 PM, Dave Froble wrote:
> On 10/21/2023 7:31 PM, bill wrote:
>> On 10/21/2023 1:48 PM, Arne Vajhøj wrote:
>>> On 10/21/2023 12:12 PM, Arne Vajhøj wrote:
>>>> On 10/21/2023 11:39 AM, bill wrote:
>>>>> On 10/21/2023 10:55 AM, Arne Vajhøj wrote:
>>>>>> On 10/21/2023 8:45 AM, bill wrote:
>>>>>>> On 10/21/2023 1:45 AM, gah4 wrote:
>>>>>>>> seem that many like to use i, j, k, l, m, and n for the
>>>>>>>> first letter of integer variables in C and other languages,
>>>>>>>> following the Fortran convention of over 60 years ago.
>>>>>>>
>>>>>>> I have also never seen this.  Except for entries in the
>>>>>>> obfuscated C contest the norm for C program, in my experience,
>>>>>>> has always been to use meaningful names for variables and not
>>>>>>> some old convention from a different language.
>>>>>>
>>>>>> Hmmm.
>>>>>>
>>>>>> Are you telling me that you have never seen i, j, ...
>>>>>> used as variable in a for loop in C?
>>>>>
>>>>> Probably, but if someone is going to use a single letter variable
>>>>> for a loop index more likely would be a, b, x, y, z.  But I'm sure
>>>>> an i would show up more for it being the first letter of the word
>>>>> index than any influence from Fortran.  Can't say I have ever seen
>>>>> j used.
>>>>
>>>> My impression is that:
>>>>
>>>> for(i = 0, ...
>>>> for(int i = 0, ...
>>>>
>>>> outnumber:
>>>>
>>>> for(a = 0, ...
>>>> for(int a = 0, ...
>>>>
>>>> by a factor 1000.
>>>>
>>>> :-)
>>>
>>> The gentlemen K & R also use i and j in the C book.
>>>
>>> And the gentleman Stroustrup also use i in the C++ book.
>>
>> What can I say.  The places where I worked frowned on (read
>> that as forbid) single letter variables for any use at any
>> time.  They also frowned on the concept of IMPLICIT.  Define
>> everything and explain it in comments and the documentation.
>>
>> bill
>>
>>
>
> As they were signing the paychecks, they were entitled to an opinion.
> But other opinions are valid.
>

Valid does not necessarily mean better or even good.

As an interesting note, out of curiosity, I grabbed a Fortran book
I had sitting here within easy reach. It's from 1983. There are
at least 30 example programs written (and copyrighted!) by the author.
While he uses I and J as indexes in a couple of them more often he uses
INDEX and other times he uses words like IREC, LOC, ICOUNT. The one he
uses the most was INDEX. So it appears there was a lot of variation.
I still hold to the opinion that single letter variables with no
immediate apparent function are a bad idea. The compiler can understand
it, but humans may become confused. :-)

bill

On Sunday, October 22, 2023 at 10:12:45 AM UTC-4, bill wrote:
> On 10/21/2023 10:59 PM, Dave Froble wrote:
> > On 10/21/2023 7:31 PM, bill wrote:
> >> On 10/21/2023 1:48 PM, Arne Vajhøj wrote:
> >>> On 10/21/2023 12:12 PM, Arne Vajhøj wrote:
> >>>> On 10/21/2023 11:39 AM, bill wrote:
> >>>>> On 10/21/2023 10:55 AM, Arne Vajhøj wrote:
> >>>>>> On 10/21/2023 8:45 AM, bill wrote:
> >>>>>>> On 10/21/2023 1:45 AM, gah4 wrote:
> >>>>>>>> seem that many like to use i, j, k, l, m, and n for the
> >>>>>>>> first letter of integer variables in C and other languages,
> >>>>>>>> following the Fortran convention of over 60 years ago.
> >>>>>>>
> >>>>>>> I have also never seen this. Except for entries in the
> >>>>>>> obfuscated C contest the norm for C program, in my experience,
> >>>>>>> has always been to use meaningful names for variables and not
> >>>>>>> some old convention from a different language.
> >>>>>>
> >>>>>> Hmmm.
> >>>>>>
> >>>>>> Are you telling me that you have never seen i, j, ...
> >>>>>> used as variable in a for loop in C?
> >>>>>
> >>>>> Probably, but if someone is going to use a single letter variable
> >>>>> for a loop index more likely would be a, b, x, y, z. But I'm sure
> >>>>> an i would show up more for it being the first letter of the word
> >>>>> index than any influence from Fortran. Can't say I have ever seen
> >>>>> j used.
> >>>>
> >>>> My impression is that:
> >>>>
> >>>> for(i = 0, ...
> >>>> for(int i = 0, ...
> >>>>
> >>>> outnumber:
> >>>>
> >>>> for(a = 0, ...
> >>>> for(int a = 0, ...
> >>>>
> >>>> by a factor 1000.
> >>>>
> >>>> :-)
> >>>
> >>> The gentlemen K & R also use i and j in the C book.
> >>>
> >>> And the gentleman Stroustrup also use i in the C++ book.
> >>
> >> What can I say. The places where I worked frowned on (read
> >> that as forbid) single letter variables for any use at any
> >> time. They also frowned on the concept of IMPLICIT. Define
> >> everything and explain it in comments and the documentation.
> >>
> >> bill
> >>
> >>
> >
> > As they were signing the paychecks, they were entitled to an opinion.
> > But other opinions are valid.
> >
> Valid does not necessarily mean better or even good.
>
> As an interesting note, out of curiosity, I grabbed a Fortran book
> I had sitting here within easy reach. It's from 1983. There are
> at least 30 example programs written (and copyrighted!) by the author.
> While he uses I and J as indexes in a couple of them more often he uses
> INDEX and other times he uses words like IREC, LOC, ICOUNT. The one he
> uses the most was INDEX. So it appears there was a lot of variation.
> I still hold to the opinion that single letter variables with no
> immediate apparent function are a bad idea. The compiler can understand
> it, but humans may become confused. :-)
>
>
> bill

You are correct. When I was teaching FORTRAN, it was common to use the initial letter to determine the variable type. Standard practice was to use single letters for "temporary" DO loop counters with more descriptive variables elsewhere. For example, within READ or WRITE statements: Write (2,6) a(i), I=1,10

Note that at that time, the IMPLICIT NONE was not yet common.

Dan

On 10/22/2023 10:12 AM, bill wrote:

> Valid does not necessarily mean better or even good.

Nor does it mean bad, either. That cuts both ways.

> As an interesting note, out of curiosity, I grabbed a Fortran book
> I had sitting here within easy reach. It's from 1983. There are
> at least 30 example programs written (and copyrighted!) by the author.
> While he uses I and J as indexes in a couple of them more often he uses
> INDEX and other times he uses words like IREC, LOC, ICOUNT. The one he
> uses the most was INDEX. So it appears there was a lot of variation.
> I still hold to the opinion that single letter variables with no
> immediate apparent function are a bad idea. The compiler can understand
> it, but humans may become confused. :-)
>
>
> bill
>

Ok, trivial, but then again, trivial sometimes occurs in programming.

For Z% = 1% to 10%
Print Z%
Next Z%

Ok, what is not apparent about Z% ?

And:

For This_is_the_Loop_Index_variable% = 1% to 10%
Print This_is_the_Loop-Index_variable%
Next This_is_the_Loop_Index_variable%

Sure seems less understandable after trying to read the variable name. As a
human I'd become so upset that I might become confused. The compiler won't
care, unless the multiple references to the variable get mis-spelled, and there
is much more possibility of doing so, I did just writing this. Can you find it?
Got to wonder how many times Z% might be mis-spelled?

KISS would seem to apply here ....

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

> For This_is_the_Loop_Index_variable% = 1% to 10%
> Print This_is_the_Loop-Index_variable%
> Next This_is_the_Loop_Index_variable%
>
> Sure seems less understandable after trying to read the variable name. As a
> human I'd become so upset that I might become confused. The compiler won't
> care, unless the multiple references to the variable get mis-spelled, and there
> is much more possibility of doing so, I did just writing this. Can you find it?

Yes: _ vs -

Dan

On 10/22/2023 2:19 PM, Dave Froble wrote:
> On 10/22/2023 10:12 AM, bill wrote:
>
>> Valid does not necessarily mean better or even good.
>
> Nor does it mean bad, either.  That cuts both ways.
>
>> As an interesting note, out of curiosity, I grabbed a Fortran book
>> I had sitting here within easy reach.  It's from 1983.  There are
>> at least 30 example programs written (and copyrighted!) by the author.
>> While he uses I and J as indexes in a couple of them more often he uses
>> INDEX and other times he uses words like IREC, LOC, ICOUNT.  The one he
>> uses the most was INDEX.  So it appears there was a lot of variation.
>> I still hold to the opinion that single letter variables with no
>> immediate apparent function are a bad idea.  The compiler can understand
>> it, but humans may become confused.  :-)
>>
>>
>> bill
>>
>
> Ok, trivial, but then again, trivial sometimes occurs in programming.
>
> For Z% = 1% to 10%
>     Print Z%
> Next Z%
>
> Ok, what is not apparent about Z% ?

We are back to the absurd. That's what I like about c.o.v.

>
> And:
>
> For This_is_the_Loop_Index_variable% = 1% to 10%
>     Print This_is_the_Loop-Index_variable%
> Next This_is_the_Loop_Index_variable%
>
> Sure seems less understandable after trying to read the variable name.

Because it is absurd. As is the program it's supposed to be.
The variable should be something simple but understandable.
Like INDEX or COUNT or even NUMBER.
> As a human I'd become so upset that I might become confused.  The

Because it's absurd.

> compiler won't care, unless the multiple references to the variable get
> mis-spelled, and there is much more possibility of doing so, I did just
> writing this.  Can you find it?

Saw it immediately. But then, it's only three lines. Compiler
would have flagged it and programmer would have fixed it if this
had been a real program longer than three lines. But then, only
an idiot would have used it.

> Got to wonder how many times Z% might
> be mis-spelled?

Wonder how one would recognize what Z% was and what it was used for in
the middle of a ten thousand line program?

>
> KISS would seem to apply here ....
>

I agree and a longer than one letter meaningfully named variable
meets that criteria.

bill

And with this, I withdraw. No one here is going to change the way
they program based on my opinion and I am certainly not going to
change the way I do it. To each his own.

On Sun, 2023-10-22 at 14:19 -0400, Dave Froble wrote:
>  This_is_the_Loop-Index_variable%

I see it. Programmer needs remedial lessons.
--
Tactical Nuclear Kittens

On 10/22/2023 3:46 PM, Single Stage to Orbit wrote:
> On Sun, 2023-10-22 at 14:19 -0400, Dave Froble wrote:
>> This_is_the_Loop-Index_variable%
>
> I see it. Programmer needs remedial lessons.
>

Doncha think that I might have done that on purpose? I did, to illustrate how
mis-spellings could occur when going overboard on long variable names.
Actually, I originally spelled it correctly, then went back and changed it, just
to show what might be more likely.

To look at a past application I wrote, just for kicks, I reviewed my RMS_LOCKS
utility. It scans the lock database, building linked lists of everything, then
processes the data to find interlocks, or to just list all the locks.

Lots and lots of declared variables and structures ...
List counters had reasonable names, usually about 4 characters ...
And yes, some small tight loops ...

One size does not fit all, use what's reasonable, what's easy to read.

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

On Sun, 2023-10-22 at 20:55 -0400, Dave Froble wrote:
> On 10/22/2023 3:46 PM, Single Stage to Orbit wrote:
> > On Sun, 2023-10-22 at 14:19 -0400, Dave Froble wrote:
> > >  This_is_the_Loop-Index_variable%
> >
> > I see it. Programmer needs remedial lessons.
>
> Doncha think that I might have done that on purpose?  I did, to
> illustrate how mis-spellings could occur when going overboard on long
> variable names. Actually, I originally spelled it correctly, then
> went back and changed it, just to show what might be more likely.

Yes, that's true and I'm occasionally guilty of doing just that. Last
paper-bag issue was using the wrong function as there were two with
similiar names.

> To look at a past application I wrote, just for kicks, I reviewed my
> RMS_LOCKS utility.  It scans the lock database, building linked lists
> of everything, then processes the data to find interlocks, or to just
> list all the locks.
>
> Lots and lots of declared variables and structures ...
> List counters had reasonable names, usually about 4 characters ...
> And yes, some small tight loops ...

To unroll or not to unroll, that is the question.

> One size does not fit all, use what's reasonable, what's easy to
> read.

Again, true enough.
--
Tactical Nuclear Kittens

On 10/22/2023 2:55 PM, bill wrote:
> And with this, I withdraw.  No one here is going to change the way
> they program based on my opinion and I am certainly not going to
> change the way I do it.  To each his own.

I don't think many if any of the participants in this subthread
wanted you to change your style.

They were puzzled over you claim:

# > seem that many like to use i, j, k, l, m, and n for the
# > first letter of integer variables in C and other languages,
# > following the Fortran convention of over 60 years ago.
# # I have also never seen this. Except for entries in the
# obfuscated C contest the norm for C program, in my experience,
# has always been to use meaningful names for variables and not
# some old convention from a different language.

If you have said something like "I know it is common to use
i, j, ... for loop counters, but I don't like that practice",
then that would just be your preference.

But saying that you have not seen i, j, ... for loop
counters is weird as they actually are common.

Arne

On 2023-10-22, bill <bill.gunshannon@gmail.com> wrote:
> On 10/21/2023 9:13 PM, gah4 wrote:
>>
>> I suspect interactive editors will also check for a terminal.
>>
>> Running vi from batch doesn't seem very useful.
>
> Actually, it works quite well (Vim warns about not having a terminal
> but then, Vim isn't vi :-).
>
> sed is better for the task but some people know vi better than sed.
>

vi seems to work everywhere. :-)

I strongly prefer emacs, but I have vi running on my Android phone for
small amounts of editing while away from a proper keyboard and a PC.

For various values of "works", it actually works just fine provided you
have a proper soft keyboard installed on your phone. (I use Hacker's Keyboard).

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

On 2023-10-22, Johnny Billquist <bqt@softjar.se> wrote:
>
> Well. They still return a value. But you've always been allowed to just
> ignore the returned value of a function in C. Compilers these days
> usually issue at least a warning if a non-void function don't explicitly
> return a value, but I think it's still technically not disallowed (but I
> could be wrong on that one...).
>

That's what -Werror is for. :-)

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

On 10/21/2023 7:51 PM, bill wrote:
> On 10/21/2023 7:44 PM, Arne Vajhøj wrote:
>> On 10/21/2023 7:22 PM, bill wrote:
>>> On 10/21/2023 5:24 PM, Johnny Billquist wrote:
>>>> On 2023-10-21 17:39, bill wrote:
>>>>> On 10/21/2023 10:55 AM, Arne Vajhøj wrote:
>>>>>> Are you telling me that you have never seen i, j, ...
>>>>>> used as variable in a for loop in C?
>>>>>>
>>>>>
>>>>> Probably, but if someone is going to use a single letter variable
>>>>> for a loop index more likely would be a, b, x, y, z.  But I'm sure
>>>>> an i would show up more for it being the first letter of the word
>>>>> index than any influence from Fortran.  Can't say I have ever seen
>>>>> j used.
>>>>
>>>> i and j as loop variables probably beats all other letters combined.
>>>>
>>>> If you haven't seen that, then I really wonder what code you have
>>>> looked at.
>>>>
>>>
>>> My early days doing production programming (In COBOL, Fortran, Pascal, C
>>> and a lot of different assemblers) was all done in an environment with
>>> very strict programming standards.  Probably explains a lot about my
>>> style and why I don't have problems with things like over-runs and
>>> bounds violations.  :-)
>>
>> C coding standards often allow or even encourage i and j.
>>
>> A few random examples:
>>
>> Linux kernel coding style:
>>
>> https://www.kernel.org/doc/html/v4.10/process/coding-style.html
>>
>> <quote>
>> LOCAL variable names should be short, and to the point. If you have
>> some random integer loop counter, it should probably be called i.
>> Calling it loop_counter is non-productive, if there is no chance of it
>> being mis-understood. Similarly, tmp can be just about any type of
>> variable that is used to hold a temporary value.
>> </quote>
>>
>> FreeBSD style guide:
>>
>> https://man.freebsd.org/cgi/man.cgi?style(9)
>>
>> <quote>
>> A for loop may declare and initialize its counting variable.
>>
>>             for (int    i = 0; i < 15; i++) {
>> </quote>
>>
>> Gnome coding style:
>>
>> https://developer.gnome.org/documentation/guidelines/programming/coding-style.html
>>
>> <quote>
>> for (i = 0; i < num_elements; i++) {
>> </quote>
>>
>
> To each his own.  I was taught with a much stricter standard
> and I think  it has done me well over my career.  I will admit
> that I have used i and variants like i1 and i2 for loops when
> writing something quick and dirty that I will only be using
> myself, but even then I know deep inside that I am doing
> something basically wrong.  :-)

I found this.

https://i.chzbgr.com/full/9833112832/hB64CB122/others-why-do-always-use-j-variabes-loops-programmers-disney-s-law

And immediately thought of this thread.

:-)

Arne

On Sunday, October 22, 2023 at 3:01:51 AM UTC-7, Johnny Billquist wrote:
> On 2023-10-22 02:17, bill wrote:
> >
> > OK. Well, my early programming days were also with DOD and
> > later with DOD Contractors. And they didn't believe in that
> > kind of programming. I see the number 95 in there. I assume
> > that's a year. Looks like someone got really lazy after I
> > left. It was 1988 when I left that world and went into academia.
> > I'll take the old way over the new way any day.
> Yes, Ada95 came out in 95. Before that you had Ada83. I believe the
> guidelines on this one didn't change.
> (I actually have done paid work in both Ada83 and Ada95.)
>
> Johnny

I learned Ada 83 at my university's CS program in the 1990s, but I haven't (yet?) done paid work in it. This was at CSU Pomona (Cal Poly, Pomona). They switched over to teaching Java a few years after a graduated and that seems to be the standard teaching language at many universities these days.

I also taught myself Ada 95 with GNAT and have been hopeful, on and off, that a "safe" language like Ada would take off. It now appears that the time has come, and that the new safe language is called Rust. The fact that AdaCore are now promoting their expertise in safety-critical software toolchains for Rust as well as for Ada, C, and C++ means the writing must be on the wall.

The newest and nicest version of Ada is Ada 2012. Ada's original claim to fame by the mid-1990s was that it had generics and exceptions before C++ copied them, it had tasking with safe message passing between tasks, and your programs would either fail to compile or throw an exception, as opposed to dumping core like other languages. It simply wasn't an option for most programmers because available compilers were too expensive, buggy, generated slow code, or some combination.

Ada 95 added objects, but with an awkward and confusing syntax. I just found a rather lengthy section of the Ada 2005 rationale explaining in detail how irritating it was to use objects in Ada 95 and how much cleaner it is to write "Y.Op(...)" as opposed to having to call the virtual method and pass it the object as the first parameter, e.g. "P.Op(Y, ...)".

https://www.adaic.org/resources/add_content/standards/05rat/html/Rat-1-3-1.html

Ada 95 greatly improved tasking, although I haven't bothered to learn the model properly since no one uses Ada. It also has a lot of support for multiple memory pools. You can write unsafe code in Ada, but it's very verbose, even by Ada standards, to declare types for data you're going to put on the heap instead of the stack or static storage, and that you're going to free yourself.

Personally, I think recent versions of C++ are the best solution for writing reliable software when a JVM-based language, or some other memory-safe language isn't an option. Getting back to coding standards, I was looking at MISRA and the actual standards appear to not be freely available, but I found something even better: JPL/NASA coding standards for C and C++.

https://nasa.github.io/fprime/UsersGuide/dev/code-style.html

Not a word about single-letter loop variable names being problematic. Lots of words about keeping your functions and code simple and straightforward. I especially like "The Power of 10" rules from Gerard J. Holzmann of the NASA/JPL Laboratory for Reliable Software:

https://en.wikipedia.org/wiki/The_Power_of_10:_Rules_for_Developing_Safety-Critical_Code

In practice, using C/C++ with many restrictions on "cleverness" and turning Clang's warning level all the way up, plus using static analysis tools, is what people are doing because the industry has collectively decided not to use Ada 2005/2012, and that's the next best thing. My one big reservation about Rust is that it really has too much cleverness in it for writing code that has to be rigorously reviewed for correctness. It's a clear improvement over C/C++ in memory-safety, but also a moving target.

Regards,
Jake Hamby