12 Sep
2014
12 Sep
'14
3:54 a.m.
From: Dave Horsfall <dave(a)horsfall.org>
On the *nix systems to which I have access, bc(1) is a
standalone
program on FreeBSD and OSX, but pipes to dc(1) on OpenBSD. I cannot
check my Penguin box (Ubuntu)
Dude! Trying to answer questions about the origins of BC by looking at
systems this late is like trying to learn Latin by studying Italian! :-)
I looked at the Version 6 source, and it's a bunch of YACC code, but it pulls
the same technique of using a pipe to an instance of DC, viz:
pipe(p);
if (fork()==0) {
close(1);
dup(p[1]);
close(p[0]);
close(p[1]);
yyinit(argc, argv);
yyparse();
exit();
}
close(0);
dup(p[0]);
close(p[0]);
close(p[1]);
execl("/bin/dc", "dc", "-", 0);
There's likely even older versions than the V6 one, but that's the earliest
source I have online on my PC for easy access.
Noel
12 Sep
12 Sep
4:35 a.m.
New subject: Etymology of bc(1)
There's likely even older versions than the V6
one, but that's the earliest
source I have online on my PC for easy access.
There's nothing in the Unix v5 manual about bc and there's no code for
it either. If there's an older version I can't find it.
I have a version of Unix v6 that has a file called /usr/doc/bc that
describes bc at length and it mentions both Lorinda Cherry and Robert
Morris.
There's even a file called /usr/doc/bca which is titled "Appendix"
with even more info.
Bob's Calculator seems plausible enough.
Mark
On 9/11/14, Noel Chiappa <jnc(a)mercury.lcs.mit.edu> wrote:
> > From: Dave Horsfall <dave(a)horsfall.org>
>
> > On the *nix systems to which I have access, bc(1) is a standalone
> > program on FreeBSD and OSX, but pipes to dc(1) on OpenBSD. I cannot
> > check my Penguin box (Ubuntu)
>
> Dude! Trying to answer questions about the origins of BC by looking at
> systems this late is like trying to learn Latin by studying Italian! :-)
>
> I looked at the Version 6 source, and it's a bunch of YACC code, but it
> pulls
> the same technique of using a pipe to an instance of DC, viz:
>
> pipe(p);
> if (fork()==0) {
> close(1);
> dup(p[1]);
> close(p[0]);
> close(p[1]);
> yyinit(argc, argv);
> yyparse();
> exit();
> }
> close(0);
> dup(p[0]);
> close(p[0]);
> close(p[1]);
> execl("/bin/dc", "dc", "-", 0);
>
There's likely even older versions than the V6
one, but that's the earliest
source I have online on my PC for easy access.
>
> Noel
> _______________________________________________
> TUHS mailing list
> TUHS(a)minnie.tuhs.org
> https://minnie.tuhs.org/mailman/listinfo/tuhs
>
8 a.m.
New subject: Etymology of bc(1)
On Thu, 11 Sep 2014, Noel Chiappa wrote:
Dude! Trying to answer questions about the origins of
BC by looking at
systems this late is like trying to learn Latin by studying Italian! :-)
Perhaps so, but I was a bit surprised that FreeBSD (and I include the Mac
in that) totally rewrote it from a Posix draft, and dropped the pipe.
I no longer have my old manuals (they disappeared during a house move).
-- Dave
1:15 p.m.
New subject: Etymology of bc(1)
Dave Horsfall scripsit:
Perhaps so, but I was a bit surprised that FreeBSD
(and I include the Mac
in that) totally rewrote it from a Posix draft, and dropped the pipe.
So did GNU.
--
John Cowan http://www.ccil.org/~cowan cowan(a)ccil.org
Do I contradict myself?
Very well then, I contradict myself.
I am large, I contain multitudes.
--Walt Whitman, Leaves of Grass
1:36 p.m.
New subject: Etymology of bc(1)
On Fri, Sep 12, 2014 at 07:15:12AM -0400, John Cowan wrote:
Dave Horsfall scripsit:
FreeBSD head has OpenBSD's dc(1) and bc(1) , which does pipe to dc(1).
(I wrote both).
-Otto
Perhaps so, but I was a bit surprised that
FreeBSD (and I include the Mac
in that) totally rewrote it from a Posix draft, and dropped the pipe.
So did GNU. 
7:12 p.m.
New subject: Etymology of bc(1)
A gentle reminder that in that era, 4K words of memory cost in the 5
digits of $'s, and many Unix machines had far less than the 64K bytes
(yes, K) that was the maximum. And there was no paging. Piping was kind
of a poor man's paging to write programs that would work on small-memory
machines.
While there were benefits, it was also a big pain, especially when
debugging. And for the most part, we abandoned this style when we got
32-bit machines that had as much as a (gasp!) megabyte of memory...
Steve
> Dave Horsfall scripsit:
>
> > Perhaps so, but I was a bit surprised that FreeBSD (and I include the
> Mac
> > in that) totally rewrote it from a Posix draft, and dropped the pipe.
>
> So did GNU.
7:18 p.m.
New subject: Etymology of bc(1)
On Sep 12, 2014, at 11:12 AM, scj(a)yaccman.com wrote:
A gentle reminder that in that era, 4K words of memory
cost in the 5
digits of $'s, and many Unix machines had far less than the 64K bytes
(yes, K) that was the maximum. And there was no paging. Piping was kind
of a poor man's paging to write programs that would work on small-memory
machines.
While there were benefits, it was also a big pain, especially when
debugging. And for the most part, we abandoned this style when we got
32-bit machines that had as much as a (gasp!) megabyte of memory...
Sure beat the heck out of trying to get overlays right. I remember “fondly”
trying to get the overlay layout manager working to get the FORTRAN programs
I was writing at the time to fit into 64k (or was it 48k). Switching from the PDP-11
to the VAX was a godsend and all that moved into the kernel and got way simpler
from the user perspective…
Warner
9:02 p.m.
New subject: Etymology of bc(1)
Steve good point but ...
On Fri, Sep 12, 2014 at 1:12 PM, <scj(a)yaccman.com> wrote:
And for the most part, we abandoned this style when we
got
32-bit machines that had as much as a (gasp!) megabyte of memory...
Yes and no (know maybe) - as Pike said "cat -v" considered harmful.
The original small is beautiful idea that made UNIX so welcome, was lost
in the process.
I remember when we got the money to maxed out the memory on an 11/34 form
48K to 256K bytes (using aftermarket memory of course). And as you point
it, it is a lot easier when we use the computer to so a lot more of the
work (that why we have compilers, debuggers and tools in general). But
when I look at programs written by daugther and her peers, I realized that
just have the innate feel for what is needed to get the job done often
seems lost to them. Things just grow and use this framework or thank
subsystem. It so easy to grab, that I wonder if we are ahead or behind.
That said, I do not yearn longly for V6. It was a fun system and we did a
lot with it and I certainly learned a ton. But I am way more productive on
my Mac. But I think the program style and lessons that many of us have
from the those times is to try to get to the nut of the task at hand and
remember how you are getting there.
Let me be less abstract. I recently found an Open Source NIXIE tube
subsystem from a guy that used an Arduino. Its very cool and impressive.
Think about it, use a $4 computer to run each NIXIE tube, instead of lots
of logic to multiples the pins. But when I looked at his code, I shook my
head. Basically straight lined everything, brute force. I remember being
taught, if the routine is longer than a printer page or two, break it up.
I took his code, put a bunch of the things into tables, made a few
functions to use those tables and cut the size of the code even with the
tables, by almost a 3rd and its a lot easier to understand.
In fact, I wrote a small subsystem so I could debug the whole thing using a
UNIX pipe to represent each digit, as debugging on the Arduino is a really,
really poor.
But what did I do - I applied lessons from a great system, UNIX to create a
clean, easy to understand (and even debug).
Anyway - my thesis is that bigger is not always better.
Clem
9:36 p.m.
New subject: Etymology of bc(1)
On Fri, 12 Sep 2014, Otto Moerbeek wrote:
FreeBSD head has OpenBSD's dc(1) and bc(1) ,
which does pipe to dc(1).
Interesting; FreeBSD seems to be adopting a lot of OpenBSD stuff now, such
as the PF firewall etc.
(I wrote both).
Well done, sir! It's always nice to meet a developer.
-- Dave
10:03 p.m.
New subject: Etymology of bc(1)
On Sep 12, 2014, at 1:36 PM, Dave Horsfall <dave(a)horsfall.org> wrote:
On Fri, 12 Sep 2014, Otto Moerbeek wrote:
There’s a lot of flow between the BSDs these days. Much more than in the Bad Old Days
which spawned the 4 different distributions…
Warner
FreeBSD head has OpenBSD's dc(1) and bc(1) ,
which does pipe to dc(1).
Interesting; FreeBSD seems to be adopting a lot of OpenBSD stuff now, such
as the PF firewall etc.
9:32 p.m.
New subject: Etymology of bc(1)
On Fri, 12 Sep 2014, John Cowan wrote:
Which I'd pointed out in a later message (I think).
-- Dave
Perhaps so,
but I was a bit surprised that FreeBSD (and I include the
Mac in that) totally rewrote it from a Posix draft, and dropped the
pipe.
So did GNU. 
8:28 p.m.
New subject: Etymology of bc(1)
On 12 Sep 2014, at 02:54, Noel Chiappa <jnc(a)mercury.lcs.mit.edu> wrote:
Dude! Trying to answer questions about the origins of
BC by looking at
systems this late is like trying to learn Latin by studying Italian! :-)
[I'm not a latin scholar, so this is second-hand information.]
Curiously, I understand that studying Italian (or romance languages in general) turns out
to be a fairly good way of learning about the Latin, where by "Latin" they mean
the language people actually *spoke*, as opposed to the deformed version of it that some
tiny minority of people wrote.
In a thousand years, it will also turn out that the best way to learn about Unix will be
by reading the archives of some mailing list where people discussed their memories of the
early versions of it.
--tim
13 Sep
13 Sep
4:52 a.m.
New subject: Etymology of bc(1)
On Fri, 12 Sep 2014, Tim Bradshaw wrote:
Curiously, I understand that studying Italian (or
romance languages in
general) turns out to be a fairly good way of learning about the Latin,
where by "Latin" they mean the language people actually *spoke*, as
opposed to the deformed version of it that some tiny minority of people
wrote.
And as we stray even further from our original brief (Monty Python quote,
but OT stuff can be fun at times), English in turn is not Romance but
Germanic, and has been described as not so much as borrowing words from
other languages as chasing them down dark alleys and beating them
senseless.
In a thousand years, it will also turn out that the
best way to learn
about Unix will be by reading the archives of some mailing list where
people discussed their memories of the early versions of it.
Reckon they might also wonder why Americans don't use English? :-)
In the meantime, the next time NASA launches a probe then perhaps it could
carry a DVD of Warren's archives; there's both an LP and a CD out there,
after all.
-- Dave, who would also spell creat(2) with an "e"
3721
days inactive
3722
days old
12 comments
9 participants
participants (9)
-
Clem Cole -
Dave Horsfall -
jnc＠mercury.lcs.mit.edu -
John Cowan -
Mark Longridge -
Otto Moerbeek -
scj＠yaccman.com -
Tim Bradshaw -
Warner Losh