[TUHS] tabs vs spaces - entab, detab

George Michaelson ggm at algebras.org
Fri Mar 5 11:09:16 AEST 2021


I use them incoherently, depending on context, time, and adherence to
something approximating to xNF for some X, normal-form. Other peoples
rules.

What I found interesting, is the divergence in strategy in managing.

strategy a: I shall show these as if they were 8 space indents. You
maybe can change this. Its columns on the screen.
strategy b: I shall, in a fit of lunacy, actually change these to
align to the 8 space indent column implications. your source is now
different
strategy c: You ran some tool like tr over these, and now, Its your
fault not mine. Good luck.

I usually wound up in strategy c but some editors of the day did
strategy b. I harbour a belief, this amongst other things, is why
patch -l was invented.

I know its heresy to contradict wiser people, but I think the number
of circumstances where space/tab actually did affect baud- or data-
rate was minimal. It was an effect more apparent and believed than
real. Sure, if you used an ASR33 the head positioning was better. That
experience set against being in a Vt52 or ADM5 world, by the time you
were in a vdu display of 80x24 I think it was gone.  I never
consciously thought about it and I used both. It wasn't very material
for the circumstances but I wasn't trying to cram source into a 2k
memory to put into a rocket or a telephone switch or anything. It's a
bit like steam-engine enthusiasts arguing about boiler tube design:
most shunters don't care.

People mostly hated me mucking around with their code. I can't blame
them. It would be horrid to live in a really nicely laid out world in
either tab or space-land and have a hooligan come and scribble over
your art with a texta.  Aside from simply coding bad answers to
problems, I do note that sending people who demanded KNF non-KNF
formatted code usually led to rapid rejection.

The US is a place where people still use fixed-width fonts and
specific spacing to file paper on the law courts, although I note
Sydney Powell didn't bother spell checking before she sent in the
incredibly weird conforming style of text. I should complain: the IETF
idnit checker is feral.

-G

On Fri, Mar 5, 2021 at 10:55 AM Larry McVoy <lm at mcvoy.com> wrote:
>
> On Thu, Mar 04, 2021 at 07:44:12PM -0500, John Cowan wrote:
> > On Thu, Mar 4, 2021 at 2:10 PM emanuel stiebler <emu at e-bbes.com> wrote:
> >
> > Does it help, if we differentiate with the type of text ?
> > >
> > > Assembler : Tabs = 8 spaces
> > > (c, c++, pascal, java, etc.) : tabs = 4 spaces
> > >
> >
> > The Lisp community long ago standardized on 2-space indentation.
>
> I used to be a 4 spaces are tabs guy but Sun beat that out of me.
> Tabs are tabs and they are for a reason, though that reason is pretty
> dead.  The reason was pretty printing listings, anything but tabs got
> all screwed up.  But it has probably been a decade or more since I've
> pretty printed anything, maybe two.  Old habits...
>
> I developed my own use of 4 spaces, those are "continuation lines"
>
>         if (some_stupidly_long_expression_that_goes_on_forever >=
>             this_never_happens_but_it_does_happen_when_deeply_nested) {
>                 statement;
>                 statement2;
>                 etc;
>         }
>
> But I'm weird, I hate
>
>         if (expr)
>                 statement;
>
> I do
>
>         if (expr) statement;
>
> Curly braces are for more than one statement or I do do
>
>         if (expr) {
>                 statement;
>         } else {
>                 statement2;
>         }
>
> I also like perl so I do
>
> #define unless(x) if (!(x))
>
> because I think this reads better:
>
>         FILE    *f;
>
>         unless (f = fopen(argv[1], "r")) {
>                 perror(argv[1]);
>                 exit(1);
>         }


More information about the TUHS mailing list