5 Jun
2024
5 Jun
'24
1:17 p.m.
Hi
Out of curiosity, what would be considered the most direct descendent of Unix available
today? Yes, there are many descendants, but they've all gone down their own
evolutionary paths.
Is it FreeBSD or NetBSD? Something else? I don't think it would be Minix or Linux
because I remember when they came along, and it was well after various Unix versions were
around.
Does such a thing even exist anymore? I remember using AT&T Unix System V and various
BSD variants back in college in the 1980's. System V was the "new thing"
back then but was eventually sold and seems to have faded. Maybe it is only available
commercially, but it does not seem as prominent as it once was.
Any thoughts?
Thanks, Andrew Lynch
5 Jun
5 Jun
1:51 p.m.
On Wed, Jun 5, 2024 at 4:17 AM Andrew Lynch via TUHS <tuhs(a)tuhs.org> wrote:
Hi
Out of curiosity, what would be considered the most direct descendent of Unix available
today? Yes, there are many descendants, but they've all gone down their own
evolutionary paths.
Is it FreeBSD or NetBSD? Something else? I don't think it would be Minix or Linux
because I remember when they came along, and it was well after various Unix versions were
around.
Does such a thing even exist anymore? I remember using AT&T Unix System V and
various BSD variants back in college in the 1980's. System V was the "new
thing" back then but was eventually sold and seems to have faded. Maybe it is only
available commercially, but it does not seem as prominent as it once was.
Any thoughts?
What exactly do you mean by "most direct descendant of Unix"? Are you
specifically talking about Research Unix? Both USG (SysIII/SysV) and
BSD are actually more like side branches from Research Unix, and
neither is really a continuation of it. After V7, Research Unix
continued until V10, but was barely distributed outside Bell Labs and
had relatively little direct influence on anything else; these late
Research Unix versions did incorporate significant amounts of code
from the side branches that took over the mainstream (especially BSD,
although there may have been a bit of USG code incorporated as well).
I'd say the closest thing to "the most direct modern descendant of
Research Unix" would be Plan 9, which continued the development of the
networking and extensibility features of late Research Unix, but
significantly broke compatibility with Unix (sometimes in ways that
are IMO not really worth the incompatibility).
4:46 p.m.
On Wednesday, June 5, 2024 at 06:51:54 AM EDT, Andrew Warkentin
<andreww591(a)gmail.com> wrote:
On Wed, Jun 5, 2024 at 4:17 AM Andrew Lynch via TUHS <tuhs(a)tuhs.org> wrote:
Hi
Out of curiosity, what would be considered the most direct descendent of Unix available
today? Yes, there are many descendants, but they've all gone down their own
evolutionary paths.
Is it FreeBSD or NetBSD? Something else? I don't think it would be Minix or Linux
because I remember when they came along, and it was well after various Unix versions were
around.
Does such a thing even exist anymore? I remember using AT&T Unix System V and
various BSD variants back in college in the 1980's. System V was the "new
thing" back then but was eventually sold and seems to have faded. Maybe it is only
available commercially, but it does not seem as prominent as it once was.
Any thoughts?
What exactly do you mean by "most direct descendant of Unix"? Are you
specifically talking about Research Unix? Both USG (SysIII/SysV) and
BSD are actually more like side branches from Research Unix, and
neither is really a continuation of it. After V7, Research Unix
continued until V10, but was barely distributed outside Bell Labs and
had relatively little direct influence on anything else; these late
Research Unix versions did incorporate significant amounts of code
from the side branches that took over the mainstream (especially BSD,
although there may have been a bit of USG code incorporated as well).
I'd say the closest thing to "the most direct modern descendant of
Research Unix" would be Plan 9, which continued the development of the
networking and extensibility features of late Research Unix, but
significantly broke compatibility with Unix (sometimes in ways that
are IMO not really worth the incompatibility).
Hi
That's interesting. I've been pondering this question for a while and suspected
the answer is either "it doesn't exist" or "depends on who you
ask" but I hadn't considered Research Unix.
For a long time, I considered AT&T System V to be the primary Unix descendant but have
changed my mind and now not sure. The question is simple, but the answer seems quite
complicated.
Thanks, Andrew Lynch
8:34 p.m.
On Wednesday, June 5th, 2024 at 3:17 AM, Andrew Lynch via TUHS <tuhs(a)tuhs.org>
wrote:
Hi
Out of curiosity, what would be considered the most direct descendent of Unix available
today?
...
Thanks, Andrew Lynch
I don't think this question has one correct answer, rather, it really depends on your
definition of purity. My two cents:
From a purely source code perspective, much of System V and its kin can be traced back to
the Research implementations of various bits. Between V7 and V8, Research incorporated a
fair deal of BSD, but at a time prior to the majority of BSD being reimplemented as
unencumbered source code, so in many parts of the codebase, the actual source still very
much was descended from V7, just with some BSD "accent" incorporated. To me one
of the most notable userland divergences in the commercial stream is the init system, what
with commercial UNIX aligning more with what is seen in CB (and allegedly USG Program
Generic 3, but I have no direct proof, just speculation based on alleged manpages.)
In any case, if you did a huge diff of the source code between say V7 and SVR4, you would
likely find a fair deal of commonality, especially in userland.
Taking an alternate viewpoint, BSD, while entirely rewritten, strove for functional
compatibility with the bits that were being replaced, and in many ways BSD
"behaved" more like Research, in reality and in "spirit". Again using
the init system as an example, to this day the BSDs use an init system much closer to
Research init than USGs run-level system. BSD also shows up in many more UNIX
"places" than System V does. Indeed primarily System V distributions over time
incorporate aspects of BSD due to their proliferation elsewhere in the UNIX world, much
more than commercial backflow in the other direction.
Given this, my humble opinion (which again this sort of thing I believe is largely a
philosophical matter of opinion...) is that the BSD line captures the spirit of Research
UNIX much more than System V does, while System V retains much more of the source code
lineage of what most folks would consider a "pure" UNIX. Of course all of this
too is predicated on treating V7 (really 32V...) as that central point of divergence.
Good luck in your quest to find the answer to this question. I suspect it has no concrete
answer and rather is one of those more philosophical quandaries that makes UNIX something
worth pondering on this level. That all said, eventually I intend via my mandiff project
to determine which of the three "last" historic UNIX manuals (SVR4, 4.4BSD, V10)
has the highest parity with V7 literature, and similar work has been attempted via source
code (D. Spinellis git repo[1]), so if that sort of quantitative analysis is more your cup
of tea, then it may be possible to boil it down to ratios of "is and isn't
V7" in codebases...but that sort of thing doesn't paint the full picture. That
and the linked git repo doesn't incorporate System V for legal reasons...a bridge I
haven't had to cross yet as I'm between V6 and V7 on my own analysis presently.
One last disclaimer as I know this question can also stir up matters of pride, this is all
opinion, and I think only can be opinion at this point, but my opinion is also only based
on observations from afar. I wasn't a key player in this stuff, those folks'
thoughts carry much more weight than mine do, but I also suspect, like good parents, folks
with more heft to their involvement in things know the value in not playing favorites and
letting their issue stand on their own.
- Matt G.
P.S. Can you tell this is one of my favorite questions to ponder :)
[1] - https://github.com/dspinellis/unix-history-repo/branches/all?page=9
8:51 p.m.
On 6/5/24 12:34 PM, segaloco via TUHS wrote:
On Wednesday, June 5th, 2024 at 3:17 AM, Andrew Lynch
via TUHS <tuhs(a)tuhs.org> wrote:
When I saw this thread appear, I was of two minds about it, but this
lines up with where my thoughts were headed. I've done a lot of delving
into the v6/v7 environments over the last 10 years or so and it feels
much closer in kinship to BSD derivatives than to SysV... source code
lineages aside. Also, I get more mileage out of my BSD books and docs
than those treating SysV. I'd vote for *BSD as sticking closest to the
unix way, if there is still such a thing... I say this as I just typed
'kldload linux64' into freebsd's terminal so I could run sublime
alongside nvi... sometimes I wish I was a purist, but I'm way too fond
of experimentation :).
Will
Hi
Out of curiosity, what would be considered the most direct descendent of Unix available
today?
...
Thanks, Andrew Lynch
snip
Given this, my humble opinion (which again this sort of thing I believe is largely a
philosophical matter of opinion...) is that the BSD line captures the spirit of Research
UNIX much more than System V does, while System V retains much more of the source code
lineage of what most folks would consider a "pure" UNIX. Of course all of this
too is predicated on treating V7 (really 32V...) as that central point of divergence.
9:02 p.m.
You could argue that the most direct descendant is the one in which all
resources are presented and accessed via open/read/write/close.
If your kernel has separate system calls for reading directories, or
setting up network connections, or debugging processes, then you may not
be a direct descendant, at least philosophically (and, yes, I know about
ptrace ...)
But your kernel might be Plan 9, which at least to me, is the direct
descendant. :-)
On Wed, Jun 5, 2024 at 10:51 AM Will Senn <will.senn(a)gmail.com> wrote:
On 6/5/24 12:34 PM, segaloco via TUHS wrote:
is largely
a philosophical matter of opinion...) is that the BSD line
captures the spirit of Research UNIX much more than System V does, while
System V retains much more of the source code lineage of what most folks
would consider a "pure" UNIX. Of course all of this too is predicated on
treating V7 (really 32V...) as that central point of divergence.
When I saw this thread appear, I was of two minds about it, but this
lines up with where my thoughts were headed. I've done a lot of delving
into the v6/v7 environments over the last 10 years or so and it feels
much closer in kinship to BSD derivatives than to SysV... source code
lineages aside. Also, I get more mileage out of my BSD books and docs
than those treating SysV. I'd vote for *BSD as sticking closest to the
unix way, if there is still such a thing... I say this as I just typed
'kldload linux64' into freebsd's terminal so I could run sublime
alongside nvi... sometimes I wish I was a purist, but I'm way too fond
of experimentation :).
Will
On Wednesday, June 5th, 2024 at 3:17 AM, Andrew
Lynch via TUHS <
tuhs(a)tuhs.org> wrote:
> Hi
>
> Out of curiosity, what would be considered the most direct descendent
of Unix
available today?
...
Thanks, Andrew Lynch
snip
Given this, my humble opinion (which again this sort of thing I believe
6 Jun
6 Jun
2:07 a.m.
On Wed, Jun 5, 2024 at 12:02 PM ron minnich <rminnich(a)gmail.com> wrote:
You could argue that the most direct descendant is the one in which all resources are
presented and accessed via open/read/write/close.
If your kernel has separate system calls for reading directories, or setting up network
connections, or debugging processes, then you may not be a direct descendant, at least
philosophically (and, yes, I know about ptrace ...)
But your kernel might be Plan 9, which at least to me, is the direct descendant. :-)
Even Plan 9's model is more like "all I/O is a file" and not
"literally everything is a file", since regular process memory is
still anonymous and fork()/rfork() are still system calls. I've never
seen an OS that puts together the "all memory is a file" of Multics
and the "all I/O is a file" of Plan 9. I think the one I'm working on
<https://gitlab.com/uxrt/uxrt-toplevel> is probably the first. Its
public "system call" API (actually a jump table into a static shared
library; the real microkernel system calls will be considered a
private implementation detail) will just consist of
read()/write()/seek()-like calls plus a few support functions to go
with them; even things like open() and close() will be RPCs over a
permanently-open channel file, and process/thread creation and memory
allocation will be done through /proc (there will of course be a
library interface over this that implements regular Unix APIs).
5 Jun
5 Jun
9:22 p.m.
Following this line of thought, - and with the disclaimer that my own personal existence
begins roughly where what has been called “The Last True UNIX” [Seventh Edition] ends, I’d
say that, if ESR - who I know can be controversial - is correct, “BSD won in the
marketplace, but System V won the standards wars” or words to that effect.
With that in mind, and given that NetBSD was forked from 386BSD and in turn gave rise to
the other BSDs around today, it would be my candidate for “most direct descendant
available today,” particularly if we’re talking wide availability. (Whilst V1-6 and beyond
were of course only available to users of business and academic mainframes and
minicomputers, I’d argue that the other two contenders, Solaris and HP-UX, are
sufficiently rare in comparison to the availability even of the open source BSD’s that the
word “available” would be doing some rather heavy lifting if I were to include them.) The
BSDs (except macOS and whatever SCO’s cash cow is called this evening) are also open
source, of course, which is inline with the spirit of early Unix.
I’ve not done an audit - and am not qualified to - but I suspect the main objection to
this line of thinking is that despite the fact it still runs on VAX, it would not surprise
me in the least to find that (excluding comments, perhaps), not a single line of code
remains the same in NetBSD 10 (and indeed several versions prior) to the equivalent in V7
- and again, I’ve no idea how much of V1 remains in V7, nor (other than knowing it was
written in assembly) how closely early PDP-11 versions resembled PDP-7 versions. By then,
I suspect we really are getting into the Ship of Theseus problem - as the ancient Greeks
would have been familiar with the issue, by the time every single plank of Theseus’ Ship
has been replaced because the old ones have decayed, is it really the Ship of Theseus
anymore?
Plus of course, though it’s more a legal issue than a philosophical one, not only at least
one version of Mach-based macOS, but also one distribution of Linux - which is known not
to contain either Minix or UNIX code - have been certified as UNIX by The Open Group.
My 2c
Jeff
Sent from my iPhone
On 5 Jun 2024, at 18:51, Will Senn
<will.senn(a)gmail.com> wrote:
On 6/5/24 12:34 PM, segaloco via TUHS wrote:
When I saw this thread appear, I was of two minds about it, but this lines up with
where my thoughts were headed. I've done a lot of delving into the v6/v7 environments
over the last 10 years or so and it feels much closer in kinship to BSD derivatives than
to SysV... source code lineages aside. Also, I get more mileage out of my BSD books and
docs than those treating SysV. I'd vote for *BSD as sticking closest to the unix way,
if there is still such a thing... I say this as I just typed 'kldload linux64'
into freebsd's terminal so I could run sublime alongside nvi... sometimes I wish I
was a purist, but I'm way too fond of experimentation :).
Will
On
Wednesday, June 5th, 2024 at 3:17 AM, Andrew Lynch via TUHS <tuhs(a)tuhs.org> wrote:
Hi
Out of curiosity, what would be considered the most direct descendent of Unix available
today?
...
Thanks, Andrew Lynch
snip
Given this, my humble opinion (which again this sort of thing I believe is largely a
philosophical matter of opinion...) is that the BSD line captures the spirit of Research
UNIX much more than System V does, while System V retains much more of the source code
lineage of what most folks would consider a "pure" UNIX. Of course all of this
too is predicated on treating V7 (really 32V...) as that central point of divergence.
9:41 p.m.
On Wed, Jun 5, 2024, 11:23 AM Jeffrey Joshua Rollin <
jefftwopointzero(a)gmail.com> wrote:
Following this line of thought, - and with the
disclaimer that my own
personal existence begins roughly where what has been called “The Last True
UNIX” [Seventh Edition] ends, I’d say that, if ESR - who I know can be
controversial - is correct, “BSD won in the marketplace, but System V won
the standards wars” or words to that effect.
With that in mind, and given that NetBSD was forked from 386BSD and in
turn gave rise to the other BSDs around today,
That is not true. FreeBSD imported the 386BSD plus patchkit patches into
its CVS tree. It did not inport NetBSD's source, though NetBSD did import
the same sources into their CVS repo days (or maybe weeks) earlier. Much
of this early history, though, is not widely available as the early NetBSD
and FreeBSD CVS repos are not available in their original form due to the
AT&T lawsuit.
And then the redo of these groups of the 4.4BSD import, the 4.4BSD-lite and
lite 2 rebased both projects further muddy the waters since they now were
both based on approximately the same pure from CSRG sources, rendering the
earlier messiness perhaps moot. Or perhaps not, but not a point that has
universal agreement, even among those involved in doing the work.
It also gets muddy because of the original patchkit authors also spintered
to for both NetBSD and FreeBSD in a way that's most kindly described as
messy, so much spin was broadcast to characterize who was first or best.
The truth is that the split was messy and definitive statements around this
are troublesome at best.
Warner
it would be my candidate for “most direct descendant available today,”
particularly if we’re talking wide availability.
(Whilst V1-6 and beyond
were of course only available to users of business and academic mainframes
and minicomputers, I’d argue that the other two contenders, Solaris and
HP-UX, are sufficiently rare in comparison to the availability even of the
open source BSD’s that the word “available” would be doing some rather
heavy lifting if I were to include them.) The BSDs (except macOS and
whatever SCO’s cash cow is called this evening) are also open source, of
course, which is inline with the spirit of early Unix.
I’ve not done an audit - and am not qualified to - but I suspect the main
objection to this line of thinking is that despite the fact it still runs
on VAX, it would not surprise me in the least to find that (excluding
comments, perhaps), not a single line of code remains the same in NetBSD 10
(and indeed several versions prior) to the equivalent in V7 - and again,
I’ve no idea how much of V1 remains in V7, nor (other than knowing it was
written in assembly) how closely early PDP-11 versions resembled PDP-7
versions. By then, I suspect we really are getting into the Ship of Theseus
problem - as the ancient Greeks would have been familiar with the issue, by
the time every single plank of Theseus’ Ship has been replaced because the
old ones have decayed, is it really the Ship of Theseus anymore?
Plus of course, though it’s more a legal issue than a philosophical one,
not only at least one version of Mach-based macOS, but also one
distribution of Linux - which is known not to contain either Minix or UNIX
code - have been certified as UNIX by The Open Group.
My 2c
Jeff
Sent from my iPhone
On 5 Jun 2024, at 18:51, Will Senn
<will.senn(a)gmail.com> wrote:
On 6/5/24 12:34 PM, segaloco via TUHS wrote:
>> On Wednesday, June 5th, 2024 at 3:17 AM, Andrew Lynch via TUHS <
tuhs(a)tuhs.org> wrote:
>>
>> Hi
>>
>> Out of curiosity, what would be considered the most direct descendent
of
Unix available today?
>>
>> ...
>>
>> Thanks, Andrew Lynch
> snip
> Given this, my humble opinion (which again this sort of thing I believe
is
largely a philosophical matter of opinion...) is that the BSD line
captures the spirit of Research UNIX much more than System V does, while
System V retains much more of the source code lineage of what most folks
would consider a "pure" UNIX. Of course all of this too is predicated on
treating V7 (really 32V...) as that central point of divergence.
When I saw this thread appear, I was of two minds
about it, but this
lines up with where my thoughts were headed. I've done a
lot of delving
into the v6/v7 environments over the last 10 years or so and it feels much
closer in kinship to BSD derivatives than to SysV... source code lineages
aside. Also, I get more mileage out of my BSD books and docs than those
treating SysV. I'd vote for *BSD as sticking closest to the unix way, if
there is still such a thing... I say this as I just typed 'kldload linux64'
into freebsd's terminal so I could run sublime alongside nvi... sometimes I
wish I was a purist, but I'm way too fond of experimentation :).
Will
6 Jun
6 Jun
12:55 p.m.
Hi,
There's a chart of the connections between Unix versions at
https://en.wikipedia.org/wiki/List_of_Unix_systems, though I dislike the
lack of direction given there are some arcs with little incline.
It says it's based on https://www.levenez.com/unix/ where Éric notes his
chart is not limited to just source-code transfer.
--
Cheers, Ralph.
10:49 p.m.
Ralph Corderoy wrote in
<20240606095502.AD4EE210F4(a)orac.inputplus.co.uk>uk>:
|There's a chart of the connections between Unix versions at
|https://en.wikipedia.org/wiki/List_of_Unix_systems, though I dislike the
|lack of direction given there are some arcs with little incline.
|It says it's based on https://www.levenez.com/unix/ where Éric notes his
|chart is not limited to just source-code transfer.
I also admire that FreeBSD and NetBSD keep on maintaining the
bsd-family-tree (and in the original form, not that dots thing, or
how it was called). So that starts with
First Edition (V1)
|
Second Edition (V2)
|
Third Edition (V3)
|
Fourth Edition (V4)
|
Fifth Edition (V5)
|
Sixth Edition (V6) -----*
\ |
\ |
\ |
Seventh Edition (V7)----|----------------------*
\ | |
\ 1BSD |
32V | |
\ 2BSD---------------* |
\ / | |
\ / | |
\/ | |
3BSD | |
| | |
4.0BSD 2.79BSD |
| | |
4.1BSD --------------> 2.8BSD <-*
| |
4.1aBSD -----------\ |
| \ |
4.1bBSD \ |
| \ |
*------ 4.1cBSD --------------> 2.9BSD
/ | |
Eighth Edition | 2.9BSD-Seismo
| | |
+----<--- 4.2BSD 2.9.1BSD
...
and says
Multics 1965
UNIX Summer 1969
DEC PDP-7
First Edition 1971-11-03 [QCU]
DEC PDP-11/20, Assembler
Second Edition 1972-06-12 [QCU]
10 UNIX installations
Third Edition 1973-02-xx [QCU]
Pipes, 16 installations
Fourth Edition 1973-11-xx [QCU]
rewriting in C effected,
above 30 installations
Fifth Edition 1974-06-xx [QCU]
above 50 installations
Sixth Edition 1975-05-xx [QCU]
port to DEC Vax
Seventh Edition 1979-01-xx [QCU] 1979-01-10 [TUHS]
first portable UNIX
..
with a nice Bibliography with falsely underscored headline plus
URL: https://cgit.freebsd.org/src/tree/share/misc/bsd-family-tree
It also covers the system most of you are using (later).
--steffen
|
|Der Kragenbaer, The moon bear,
|der holt sich munter he cheerfully and one by one
|einen nach dem anderen runter wa.ks himself off
|(By Robert Gernhardt)
9 Jun
9 Jun
11 a.m.
Excellent responses here. Brings back so many great memories.
My 1 cent would be to ask the question:
Which of today's Unix variants (Linux, BSD, AIX, Cygwin, ...) is
closest to the philosophy of the Ken-Denis-Doug versions of V6 Unix?
All the variants I see today suffer from "complexification" - a John Mashey
term.
Documentation of commands today has grown 5 to 10 fold for each
command in /usr/bin. V7 had less than 64 well documented
system calls. Today's Linux, AIX, and others have how many?
I don't know.
The concept of producing a stream of text as the output of a program
that does simple jobs well has been replaced by "power-shell" thinking
of passing binary objects rather than text between program - a decidedly
non-portable idea.
Passing "objects" requires attaching to a dynamically linked
library (that will change or even disappear
with the next release of the OS or the object library).
With Research Unix, I could pipe the output of a Unix program
running on an Intel 486 to another program running on a Motorola
68000 or a Zilog Z80000 or an IBM AIX machine.
IPhones, iPads, and my Android tablet don't have a usable text editor. All
non-Unix text editors seem to struggle to offer a fixed width font. (Ever
try to make columns line up on an iPhone or Android tablet?)
Complexification
rears its ugly head.
I still use vi on both my Mac and PC (Cygwin). (I can't find a usable gvim
for Mac and Macvim is weird but doesn't seem to know what a mouse is.)
Unix brought automation to the forefront of possibilities. Using Unix,
anyone could do it - even that kid in Jurassic Park. Today, everything
is GUI and nothing can
be automated easily or, most of the time, not at all.
Unix is an ever shrinking oasis in a desert of non-automation and
complexity.
It is the loss of automation possibilities that frustrates me the most.
(Don't mind me, I'm just outgassing for no good reason.)
Ed
On Thu, Jun 6, 2024 at 3:06 PM Steffen Nurpmeso <steffen(a)sdaoden.eu> wrote:
Ralph Corderoy wrote in
<20240606095502.AD4EE210F4(a)orac.inputplus.co.uk>uk>:
|There's a chart of the connections between Unix versions at
|https://en.wikipedia.org/wiki/List_of_Unix_systems, though I dislike the
|lack of direction given there are some arcs with little incline.
|It says it's based on https://www.levenez.com/unix/ where Éric notes his
|chart is not limited to just source-code transfer.
I also admire that FreeBSD and NetBSD keep on maintaining the
bsd-family-tree (and in the original form, not that dots thing, or
how it was called). So that starts with
First Edition (V1)
|
Second Edition (V2)
|
Third Edition (V3)
|
Fourth Edition (V4)
|
Fifth Edition (V5)
|
Sixth Edition (V6) -----*
\ |
\ |
\ |
Seventh Edition (V7)----|----------------------*
\ | |
\ 1BSD |
32V | |
\ 2BSD---------------* |
\ / | |
\ / | |
\/ | |
3BSD | |
| | |
4.0BSD 2.79BSD |
| | |
4.1BSD --------------> 2.8BSD <-*
| |
4.1aBSD -----------\ |
| \ |
4.1bBSD \ |
| \ |
*------ 4.1cBSD --------------> 2.9BSD
/ | |
Eighth Edition | 2.9BSD-Seismo
| | |
+----<--- 4.2BSD 2.9.1BSD
...
and says
Multics 1965
UNIX Summer 1969
DEC PDP-7
First Edition 1971-11-03 [QCU]
DEC PDP-11/20, Assembler
Second Edition 1972-06-12 [QCU]
10 UNIX installations
Third Edition 1973-02-xx [QCU]
Pipes, 16 installations
Fourth Edition 1973-11-xx [QCU]
rewriting in C effected,
above 30 installations
Fifth Edition 1974-06-xx [QCU]
above 50 installations
Sixth Edition 1975-05-xx [QCU]
port to DEC Vax
Seventh Edition 1979-01-xx [QCU] 1979-01-10 [TUHS]
first portable UNIX
..
with a nice Bibliography with falsely underscored headline plus
URL: https://cgit.freebsd.org/src/tree/share/misc/bsd-family-tree
It also covers the system most of you are using (later).
--steffen
|
|Der Kragenbaer, The moon bear,
|der holt sich munter he cheerfully and one by one
|einen nach dem anderen runter wa.ks himself off
|(By Robert Gernhardt)
--
Advice is judged by results, not by intentions.
Cicero
30 Jun
30 Jun
2:05 p.m.
New subject: syscalls, records in pipe [was: Re: most direct Unix descendant]
On Sun, Jun 09, 2024 at 03:00:53AM -0500, Ed Bradford wrote:
Excellent responses here. Brings back so many great
memories.
My 1 cent would be to ask the question:
Which of today's Unix variants (Linux, BSD, AIX, Cygwin, ...) is
closest to the philosophy of the Ken-Denis-Doug versions of V6 Unix?
All the variants I see today suffer from "complexification" - a John Mashey
term.
Documentation of commands today has grown 5 to 10 fold for each
command in /usr/bin. V7 had less than 64 well documented
system calls. Today's Linux, AIX, and others have how many?
I don't know.
Well, here you are (on my more or less updated ParrotOS, a Debian
derivative) (watch for insider line break):
-$ man 2 syscalls | awk '/Sys.*Kern.*Not/,/On many plat/ {if ( $1
~ /.*\(2\)/ ) print $1;}' | sort | uniq | wc -l
468
The concept of producing a stream of text as the
output of a program
that does simple jobs well has been replaced by "power-shell" thinking
of passing binary objects rather than text between program - a decidedly
non-portable idea.
I guess an analogue of this could be made - for example, when I say
'dpkg -s scm' it prints:
-$ dpkg -s scm
Package: scm
Status: install ok installed
Priority: optional
Section: interpreters
Installed-Size: 2294
Maintainer: Debian Scheme Dream Team <debian-scheme(a)lists.debian.org>
Architecture: amd64
Version: 5f3-4
Depends: slib, libc6 (>= 2.34), libncurses6 (>= 6), libreadline8 (>=
6.0), libtinfo6 (>= 6), libx11-6
Suggests: r5rs-doc
Description: Scheme language interpreter
SCM conforms to the R5RS (Revised^5 Report on the Algorithmic
Language Scheme) and IEEE P1178 specifications, and is portable
across many architectures and operating systems. It additionally
includes a set of popular Common Lisp functions, POSIX and X Windows
integration, and the Hobbit scheme-to-C compiler.
Homepage: https://people.csail.mit.edu/jaffer/SCM.html
I believe the format is easy to grasp and I think there is an RFC for
this, but cannot quickly find the number.
But the cost - blowing and puffing up every and each of the sh tool.
[...]
Unix brought automation to the forefront of
possibilities. Using Unix,
anyone could do it - even that kid in Jurassic Park. Today, everything
is GUI and nothing can
be automated easily or, most of the time, not at all.
Unix is an ever shrinking oasis in a desert of non-automation and
complexity.
It is the loss of automation possibilities that frustrates me the most.
When I became fascinated with Unix while a student still, I could not
find anybody else drawn to the idea of automating stuff to be done,
like Unix enabled. OTOH, I never asked people much about what they
were thinking. My impression, however, was that Windows 3.1 had
already owned their minds and whatever Unix was, it did not concerned
anybody too much. Clients used Windows, future bosses wanted
Windows. Have 500 thousands lines in a file? Read it into Excel and
massage it there. Have fifty files, each 500000 lines worth? Oy,
repeat fifty times, manually. Find an intern slaving body and automate
with him, of him, on him, however you call it.
In a way, it was and still is better than all advances of modern so
called AI - after all, interns do really care about results.
--
Regards,
Tomasz Rola
--
** A C programmer asked whether computer had Buddha's nature. **
** As the answer, master did "rm -rif" on the programmer's home **
** directory. And then the C programmer became enlightened... **
** **
** Tomasz Rola mailto:tomasz_rola@bigfoot.com **
2:11 p.m.
New subject: syscalls, records in pipe [was: Re: most direct Unix descendant]
On Sun, Jun 30, 2024 at 01:05:43PM +0200, Tomasz Rola via TUHS wrote:
[...]
Well, here you are (on my more or less updated ParrotOS, a Debian
derivative) (watch for insider line break):
-$ man 2 syscalls | awk '/Sys.*Kern.*Not/,/On many plat/ {if ( $1
~ /.*\(2\)/ ) print $1;}' | sort | uniq | wc -l
468
Forgot to add - with the caveat that some calls are tagged as
deprecated, removed or specific to only one architecture. The number
should be treated as rough estimate.
--
Regards,
Tomasz Rola
--
** A C programmer asked whether computer had Buddha's nature. **
** As the answer, master did "rm -rif" on the programmer's home **
** directory. And then the C programmer became enlightened... **
** **
** Tomasz Rola mailto:tomasz_rola@bigfoot.com **
1
days inactive
26
days old
14 comments
11 participants
participants (11)
-
Andrew Lynch -
Andrew Warkentin -
Ed Bradford -
Jeffrey Joshua Rollin -
Ralph Corderoy -
ron minnich -
segaloco -
Steffen Nurpmeso -
Tomasz Rola -
Warner Losh -
Will Senn