In the beginning was the code, and
the code was with the computer. But the code stayed with the computer. And if
we changed the computer like for an upgrade or something then we had to rewrite all the code.
And everyone agreed that this was wasteful. What we needed was a platform on which we can develop software. Capable of
handling different hardware. But also not too expensive. Perhaps
collectively managed and developed by a team of volunteers around the world?
I wonder if such a thing is possible. The creation and emergence of the operating
system Unix was one of those special moments in technology history. In this video,
the rise and fragmentation of Unix. ## Beginnings In 1965, a few scientists at Bell Labs joined
with peers from MIT and General Electric on a project called "Multiplexed Information
and Computing Service", or Multics. The idea was to create this general-purpose
utility for sharing time on a computer system. Computers were super expensive,
so time-sharing operating systems were developed for multiple users to
efficiently share computer resources. Over time these time-sharing systems
matured into a communications tool connecting multiple users on the time share,
with user public profiles and everything. General Electric then had a business selling
time shares for their computer systems - and offered the Multics team the use of a GE 645
mainframe computer. Simulated with a 635 computer. Multics experimented with some
interesting concepts. For instance, the idea of having arbitrary file names and
directory structures - a virtual memory system. Virtual memory is where secondary
storage like from a hard drive can be used like as if it were part of
the main memory - the RAM. With this, the computer can handle more data than its RAM
would otherwise physically allow it to handle. It was a massive improvement over existing file
systems of the day and is still used today. The Multics team tried to bring together ideas
like these - which previously were floating around but not implemented -
into a single commercial product. ## The End of Multics Looking back at it, they were
probably trying to boil the ocean. Progress turned into a slog - too much
money spent on too few people following too vague of a plan. Frustrated
with the absence of a workable product, Bell Labs formally pulled
out of the Multics project in 1969. Soon thereafter GE later decided
to exit the computer industry entirely - selling the division to Honeywell. Despite this, a few scientists at the Bell Labs
Computing Science Research Center - Ken Thompson, Dennis Ritchie, Rudd Canaday, Doug McIlroy, and
J.F. Ossanna - continued working on the project. However, the end of Multics did
mean losing the GE 635 computer that they had been working with
up until then. Ken Thompson had written up this little space game for
that computer called "Space Travel". Space Travel was an ambitious game that simulated
the movement of the planets in the solar system and Pluto. The player can guide a ship through
them and try to land on the planets ... and Pluto. A nice and mildly addicting game, but it
cost $50-75 to play each time - per the cost of the computer timeshare. So when Thompson
found a graphics-capable PDP-7 minicomputer that another department wasn't using, he
decided in 1969 to rewrite the game for it. This turned out to be an ambitious project -
more so since Thompson had to re-implement from scratch things like a debugging subsystem
and a floating point arithmetic package. The work took some time and had been
quite tedious - with the GE computer OS outputting paper tapes that
had to be carried to the PDP-7. ## A File System After finishing that, Thompson thought
to then try implementing some ideas that he had been banding around
with Dennis Ritchie and Canaday. They had been talking about a new type of file
system for the GE 635 computer - sketching ideas for keeping files out of each other's
hair. Now quite familiar with the PDP-7, he hacked out an implementation of
this file system in a day or two. From there, Thompson and others added
a series of simple utilities - copying, printing, deleting, and editing
files. As well as a simple command interpreter - a shell - which was
a program that ran other programs. Over time, the concept of the "file"
coalesced. A file was an interface through which you can perform certain
data operations like reading or writing. But the File System itself doesn't
care what was actually in the file. In doing so, the concept abstracted away the
differences between various computer hardwares. Now anyone on any device can make changes to a
file - and this became one of Unix’s killer apps. In the summer of 1969, Ken Thompson's wife
took their newborn son to see her parents, leaving Thompson with a lot of time. By the end of that summer, the whole thing
had been rewritten into something separate from the original GECOS operating
system this whole journey started on. They wouldn't have called it an
operating system back then. Back then it was just seen as this convenient
platform for developing software on. Some time next year 1970, team member Brian
Kernighan suggested a new name for this system - "Unics" - as a "treacherous pun" on Multics.
At some point, the spelling switched to "Unix". ## The Usefulness of Unix Unix on the PDP-7 offered a decent
environment on which you can do programming. But the PDP-7 was an outdated minicomputer - first introduced back in 1964. Not to
mention, it wasn't even theirs. So in 1970 they asked the
company for a DEC PDP-11, which had then been recently introduced. Now,
Thompson, Ritchie and the team have long been asking for money for a new computer but
those requests were always rejected. This time, however, the request was
granted. In part because the PDP-11 was a cheaper computer at just $65,000. And
also in part the request now had a compelling pitch attached to it - to apply the file system
for the use of creating and editing text files. Now with an actual use case, the
request was eventually approved. Unix was rewritten for the PDP-11's
low level assembly language, now with a text editor and a typesetting
markup language called "roff". With this, the Unix operating system was offered
to the Bell Labs Patent Department. The Patent Department chose to use Unix over the
competing commercial product because Thompson and Ritchie quickly endowed "roff"
with the ability to do line-numbered pages. What began with just three typists in the Patent
Department became a popular homegrown product across the whole Bell Labs organization with
multiple versions and its own support group. ## Unix Spreads Unix had only ever been intended
as an internal tool for Bell Labs. However, it quickly escaped the laboratory
and started spreading widely throughout the computing communities. Its unexpected
popularity can be attributed to a few things. First, Unix was born on relatively
humble hardware. Back then, the users of the hardware were
not the same as those buying it. Programmers of the day had to make do
with whatever the budget could afford. A standard Unix computer back then
- usually a PDP-11/40 - cost about $50-150,000 in 1977 dollars depending
on the memory configuration. Considering a graduate student cost about
$10,000 back then it was a lot, but far cheaper than mainframes which might
have cost a half million dollars at the minimum. Second, the Unix source code
was written in an attractive, higher-level programming language called
"C" - the first such OS like this. Dennis Ritchie had produced "C" from "B".
Ken Thompson created "B" when he ported the Basic Combined Programming Language
or BCPL by Martin Richards at MIT. It was far easier to program in C than in
low-level machine assembly language. Because it was written in C, Unix was easy
to port to different hardware architectures outside of the PDP minicomputer family. C
also made Unix easier to modify and enhance. Third and most importantly, Unix did not cost an arm and a leg to acquire. AT&T and
Bell Labs sold the Unix source code to nonprofits like universities for
something like a few hundred dollars. This was in contrast to other software companies, which zealously guarded their programs’
source code. And even if users were able to see the code, they were not allowed
to modify it. Not the case with Unix. ## The AT&T Consent Decree This behavior wasn't exactly out of
Bell Labs' and AT&T's good heart. Back in 1956, AT&T settled an anti-trust lawsuit by the US government - the
1956 AT&T Consent Decree. The settlement decreed that in exchange for
a legal monopoly on the US telephone system, AT&T would make available all of its
inventions to the academic community at no charge, or license them
for fair and reasonable terms. The Consent Decree also barred AT&T from
entering the computer business - so a computer operating system like Unix
was not commercially useful for them. Funny enough, the Consent Decree had already
facilitated the spread of one ground-breaking electronics invention a few decades earlier - the
transistor. And now it has done it for another. Whichever university asked for the Unix
source code from AT&T had it … or more like had it thrown at them from out the window
of a passing truck. Wary of violating the Decree, AT&T managers shied away from
providing support to their licensees. Again, another piece of fortune because
it encouraged university students with more cleverness and time than money to work
together implementing the features they wanted. By the early 1970s, Unix had gone worldwide.
University computers in Australia, the United Kingdom, Belgium, and
the Netherlands were running it. ## Berkeley Unix In November 1973, Ken Thompson and
Dennis Ritchie presented the first Unix paper at the Symposium on Operating
Systems Principles at Purdue University. UC Berkeley Professor Bob Fabry was at
the talk and obtained a copy. Since Unix then only ran on PDP minicomputers, several
departments pooled their resources to get one. A few graduates and professors started
working with it. And Ken Thompson himself joined Berkeley as a visiting professor in
1975 to help. He helped the team install the latest Unix version - Version 6 - on
a newly acquired PDP 11/70 minicomputer. Two students in particular became quite
familiar with this Unix - Chuck Haley and Bill Joy. They worked on finishing a
Pascal implementation that Thompson started, which would allow Unix to support
this higher level language. It turned out quite well, widely admired
because of its excellent error handling. The Pascal implementation Thompson had apparently did not turn out error messages that
were all that easy to understand. Haley and Joy also added a few utilities
including a very capable wysiwyg text editor called "ex" for "EXtended". Ex later became
"vim", a text editor that some people like. Word got around thanks to how well the Pascal
compiler recovered from errors and how fast it compiled. People started requesting their
own copies of the Berkeley variant of Unix. So early in 1977, Bill Joy put together
30 copies of what he called the "Berkeley Software Distribution" or BSD and sent them
out for about $50 per tape. From there it got quite popular, which made good money
for Joy since a tape cost like $10 each. A second version, "Second Berkeley Software
Distribution" or 2BSD, quickly followed. Then in 1978, DEC introduced a new
computer called the VAX-11/780, a 32-bit computer with more memory and
compute power. AT&T released a version of their own UNIX/32V for it, but it did not
support the VAX's virtual memory capabilities. This limited processing to the computer's
physical memory, diminishing its power. In a game-changing move, Bill Joy and
the Turkish grad student Ozalp Babaoglu added that virtual memory feature and
shipped it in December 1979 as 3BSD. With this, Berkeley solidified its position as the coordinating gateway for new,
leading edge Unix releases. ## DARPA & the Internet
BSD then took another big leap thanks to DARPA. Up until then, DARPA ran on a legacy mishmash
of software written in different languages for different computers. In 1979, DARPA
sought to consolidate to a single "universal computing environment" so they can
share their software across the organization. They chose Unix to be that computing
environment because of how it capably handled different hardware. In 1980,
Fabry received an 18-month contract from DARPA to add some features
to the 3BSD release of Unix. So he along with Bill Joy set up a new
organization called the Computer Systems Research Group or CSRG for this. They had
a small group of core programmers - the "steering committee" - coordinating a
global network of volunteer contributors, working on the 4BSD or BSD Unix series. This new Unix had to support several
of DARPA’s protocols including those for the Internet. For instance, 4.2BSD
fully supported the Internet protocol stack TCP/IP and played a significant role in
popularizing the Internet as we know it today. 4.2 was very popular. Over a thousand
licenses were issued just a month after its official release in April 1983 -
more than all of the other previous distributions combined. Momentum was accelerating. ## From Hobby to Industry Prior to that release though
- in the summer of 1982 - Bill Joy announced that he would leave the CSRG
to join Sun Microsystems as a full-cofounder. His work at BSD was earning him
tens of thousands of dollars, but he felt the academic university
atmosphere at Berkeley constrained efforts to grow. In his own words, "it
needed to be a commercial activity". Sun Microsystems is famous for
pioneering and popularizing the workstation computer which included - Unix
OS with scientific/engineering applications, the Motorola 68000 microprocessor,
and other off-the-shelf hardware. They later developed their own closed-source
Unix variant for their hardware - SunOS, branched off from the 4.2BSD version of Unix. A single Sun workstation by itself could not match
up with a mainframe or even a minicomputer. But these things were meant to be networked, and
in doing so became immensely valuable. Sun in turn reaped the benefits to become one of the
fastest growing companies in Silicon Valley. Sun quickly rose above, but they were far from the
only company to commercialize Unix. For instance, we have the small, Berkeley-based
software company Mt. Xinu. They sold a commercially licensed version of
the BSD for the DEC Vax minicomputer. The interesting name fits well with its
slogan, "We know Unix(TM) backwards and forwards". They handed out some pretty
fun posters and calendars at events. Other notable Unix-based startups
include Santa Cruz Operation, which sold Unix variants for
x86 computers. And Onyx Systems, which marketed a variant of Unix
for Zilog-based Personal Computers. Even Microsoft got into the fun with their Unix
variant for 16-bit microcomputers called Xenix. And of course, we have Steve Jobs' workstation
computer startup - NeXT. Founded in 1985, their operating system NeXTSTEP was derived
from BSD, 4.3BSD Tahoe if I recall correctly. ## Conclusion Unix pioneered these powerful concepts that helped
make software the powerhouse industry it is today. Its early development and foundations came
about through the efforts of people just interested in it, not really making money from it. But Bill Joy clearly said the quiet part out
loud. As the community grew and its potential became more apparent, the work surrounding
Unix needed to be a commercial activity. 1983 marks the end of Unix's awkward growth years. A bounty of wealth laid seemingly
ahead for the Unix industry - but who will reap its gains? The stage is set
for the blood and fury of the Unix Wars.
