HN Theater @HNTheaterMonth

⬐ russellbeattie

I like Rob Pike a lot, but this definitely wasn't his best presentation. It's more like, "Things Rob sorta remembers about the early days of Unix."

⬐ dang

Discussed at the time (of the talk):

The History of Unix, by Rob Pike [video] - https://news.ycombinator.com/item?id=18417918 - Nov 2018 (47 comments)

⬐ s_dev

Sound seems to begin at 3:45

⬐ None

None

⬐ nix23

Sometimes i wonder how our os's would look like if the plan9/inferno idea could catch on.

⬐ pjmlp

Android is the closest to it, currently.

Managed runtime on top of UNIX like kernel and OS IPC everywhere.

⬐ hulitu

Lol. Accessing a file in android is a mess.

⬐ pjmlp

Depends on the point of view regarding OS APIs.

⬐ hulitu

As a user i do not use APIs. I use apps. And every attempt to save something is an exercise in pain.

⬐ nix23

>OS IPC everywhere.

Just inside the device ;)

⬐ pjmlp

Everywhere was intended on the context of OS stack.

⬐ None

None

⬐ trasz

That’s only tangentially similar to Inferno, and completely unlike Plan9.

Even for Inferno the idea was to make the system flexible, Android is everything but.

another one was BLIT https://en.wikipedia.org/wiki/Blit_(computer_terminal) https://www.youtube.com/watch?v=Pr1XXvSaVUQ also mentioned in The History of Unix, by Rob Pike https://www.youtube.com/watch?v=_2NI6t2r_Hs

then there was Star

Crabs was probably the first graphical prank program—most assuredly the first well-documented one: there are a lot of scattered anecdotes about it around. Here's an amazing talk by Rob Pike where it's casually mentioned, for example:

https://www.youtube.com/watch?v=_2NI6t2r_Hs

It's not something you see very often anymore, partially because modern windowing systems (and systems in general, really) were built with different security models in mind (you can probably remember a few similar programs released into the wild for real mode operating systems, they were especially common then), but also because windowing systems have moved away from how simple UNIX's windowing system was at the time.

It's actually kind of relevant to something that was on the front-page recently: the displays they were using (JERQ/Blit bitmap displays) depended on a routine that Larry Tesler co-invented (the bit blit, hence the name "Blit").

http://doc.cat-v.org/bell_labs/blit/

https://en.wikipedia.org/wiki/Bit_blit

If you want to give it a try, I think 9front ships with a Blit emulator these days; load up v8 and you should have a nice few hours poking around.

http://9front.org/

⬐ munificent

My favorite in this category was an old Macintosh extension named, I think, "Minitaur". Everytime you restarted the machine, it shaved off a single pixel of screen space, replacing it with black. So your screen got imperceptibly smaller over time until eventually you realized something was really not right.

⬐ msla

The Blit (begat of Jerq, productized as the DMD 5620) was an interesting system: The thing could be a normal text terminal, but the magic came when the host sent it software to run locally. In short, if the IBM block-mode terminals were the equivalent of filling out HTML forms, the Blit was closer to WASM.

https://loomcom.com/blog/0115_emulating_the_dmd5620.html

Videos of Ken Thompson and Rob Pike telling Unix history stories:

Ken Thompson (interviewed by Brian Kernighan, 2019, starts after 7m38s): https://www.youtube.com/watch?v=EY6q5dv_B-o#t=7m38s

Rob Pike (Unix History presentation, 2018, starts after 3m40s): https://www.youtube.com/watch?v=_2NI6t2r_Hs#t=3m40s

⬐ clSTophEjUdRanu

Ken Thompson interviewed by Kernighan is a real treat. I love this interview.

⬐ pmoriarty

Thank you soooo much for posting the links to these. I'm just starting to dive in to the Thompson interview and it's amazing.

I strongly recommend anyone with even a vague interest in UNIX (well, obviously that would be everyone reading this HN thread) to watch these!

https://www.youtube.com/watch?v=_2NI6t2r_Hs among other things.

These were places where old-school (with classical understanding, if we follow Robert M. Pirsig) people tried to reduce grounded in reality concepts to optimal (as humanly possible) implementations, as opposed to the modern tendency to pile up more and more crap.

I think Rob Pike mentioned it in a recent talk on the history of Unix https://youtu.be/_2NI6t2r_Hs?t=1901

⬐ ken

He says this is the first case of regular expressions on a computer, which I believe, but I'm not hearing anything about runtime optimization or compilation.

⬐ chombier

less than one minute later he mentions "compiling on the fly" some assembly code https://youtu.be/_2NI6t2r_Hs?t=1944

⬐ ken

Sure, but he never claims that's the first instance of runtime compilation, just that the implementation is "pretty interesting".

⬐ mcnichol

What a fantastic talk.

I think there would be a lot of value in capturing a "History of..." type series from Rob Pike (Alongside other greats).

There is one similar to what I am referring called Web of Stories where they did this with Donald Knuth.

https://www.webofstories.com/play/donald.knuth/93

⬐ Aloha

The first time I saw a Blit/5620 and used one, I was amazed - it was like an alternative development branch of evolutionary development that never fully developed. I might prefer living in that world where it had however.

The fact that I can get several sessions going on at once, with interactive data, is fascinating - and all over a relatively slow serial data link too.

⬐ gonzo

You would have enjoyed Lisp Machines

⬐ AnIdiotOnTheNet

This is a feeling you'll experience a lot looking back on the history of computing. Sadly, it appears we live in, if not the darkest timeline, certainly one of the worst.

⬐ vram22

Related:

Unix History:

https://www.levenez.com/unix/

⬐ ChristianBundy

Audio doesn't start until 3:42.

⬐ wging

I opened this earlier, and it linked to https://www.youtube.com/watch?v=_2NI6t2r_Hs&feature=youtu.be... -- I think a moderator may have 'fixed' it in a really silly way.

⬐ aap_

Great talk and quite a few photos of things I had found photos of before.

⬐ type0

I wish he would revive the crabs!

⬐ cronix

I guess this is "the first screensaver?" I doubt it saved any screens, but neither do current screensavers lol.

⬐ fermienrico

Off-topic: Growing up in the late 80's and going to college in 2000's, I've always romanticized about places such as Bell Labs and Xero Parc in terms of their talent pool, innovation and revolutionary ideas. It feels like all the cool stuff happened 30-40 years ago.

What are some of the modern equivalent of such places?

I can think of Boston Dynamics.

⬐ bch

Isn’t it early to say Boston Dynamics has had any impact, outside of the occasional cool YouTube video? Neat, without a doubt...

Re: other modern equivalents - would free software, in general, count? We get to observe first hand, tinker-with, talk-to-creators-of, and contribute-to some of the heaviest-lifting software ever. Linux and BSDs, MySQL, Postgres, Cassandra, Python, Blender... any of those, or others, and the right research paper or YouTube video or book, and there’s a good chance you can have your mind blown.

⬐ int0x80

I think open source is indeed the continuation of that. The hole linux community, gnu/gcc, bsd etc. There are really very talented people working there making huge impact on how things work today and will work in the near future (at least).

And I think the key here is the community actitute and ecosystem. The hole 'huge communities driven by the goal of creating good/the best possible technology and code' and not directly (I know big companies are in, but they still have to play the community game/rules) by comercial or political agendas.

⬐ fasquoika

Alan Kay has asserted repeatedly that there's no place like Xerox Parc in the modern world. Not necessarily authoritative, but something to think about.

⬐ TheOtherHobbes

PARC wasn't just a ridiculous concentration of money and very smart people with a budget, it was a ridiculous concentration of money and very smart people who had creativity and vision.

There's nothing like it today, because the only vision you'll find in the current crop of smart+budget shops is how to make more cash from ads and/or the web.

Even at SpaceX, the getting-to-Mars vision is more than fifty years old.

PARC was about enhancing human creativity, cognition, and potential, and everything else came from that.

Not only is no one is working in that space today, but all the big smart+budget shops are more likely to pollute that space than contribute to it.

⬐ Aloha

PARC also had a huge amount of trouble commercializing any of its work - part of it was the intrenched bureaucracy at Xerox, and the way Xerox looked at products and sales. That said, the Laser Printer alone, paid for the money spent on PARC several times over.

⬐ stickfigure

This is unreasonably uncharitable.

I have doubts about its commercial viability, but Magic Leap seems pretty ambitious. Microsoft has been exploring all kinds of exotic technologies, some of which make it into production (kinect, hololens) and some of which don't (the original 'surface'). Alphabet has a bunch of moonshots that sound like vanity projects to me but if even one of them works out it may change the world (Waymo seems closest). Even the much-maligned Google Glass is a preview of the future.

It sounds like your standard for creativity and vision requires rewinding time to a more naive state. Peter Norton made a fortune by selling software that would undelete files. At the time it was amazing!

⬐ danso

I agree that Microsoft is doing some interesting work (including in open source) that may cause more buzz if Microsoft wasn’t saddled with the legacy of Windows and cubicle computing. I wonder how exciting/sexy Xerox’s corporate image was back in the heyday of Xerox PARC?

⬐ RickSanchez2600

AT&T Bell Labs made videos like this: https://youtu.be/tc4ROCJYbm0?list=FLHa3ljE2SLOERhULuoHNx-Q

⬐ Aloha

That intro is clearly the result of too slowly delivered cue cards - it was painful to watch.

⬐ jonsen

I easily got each and every word spoken. And had time to think. A lot of old videos have a pleasant pacing.

Rob Pike 2018 — on average i got two out of three words and had to spend mental effort on interpolating — that was painful.

⬐ ontouchstart

I totally agree with you. I watched the video so many times that I "bookmarked it" with a one-cell Jupyter Notebook:

https://nbviewer.jupyter.org/github/ontouchstart/colab_noteb...

⬐ andyjohnson0

IBM's research division used to have a good reputation for innovative, long-term work - especially in maths/CS and physics. I'm not sure how much of that has survived the recent changes in the company. Anyone want to comment on what it's like now?

⬐ WestCoastJustin

Amazon, Alphabet, Intel, Microsoft, Apple, SpaceX, etc [1]. Just look at what companies are pumping 10's of billions of R&D dollars into. But, they are doing some very cool things at Google & some of the X bets. I suspect that's why Rob Pike is there today. Personally, SpaceX looks like they are also very much pushing the envelop technically and must be a very cool place to work [2]. By the way, this was a pretty awesome talk. Rob also helped created UTF-8 & Go [3]. Pretty amazing career.

[1] https://www.recode.net/2018/4/9/17204004/amazon-research-dev...

[2] https://youtu.be/u0-pfzKbh2k?t=19

[3] https://en.wikipedia.org/wiki/Rob_Pike

⬐ atq2119

The list of companies you give doesn't quite work. Bell Labs and Xerox Parc were small organizations within a larger company.

The vast majority of people in the companies you listed are doing work that has nowhere near the density of interestingness that happened in Xerox Parc, for example. There might be individual departments in those companies that are comparable -- in Microsoft Research perhaps, and a sibling comment points out Alpha Go.

Anyway, the point is: Working at Alphabet or Intel or any of these is nothing like working at Xerox Parc, unless you happen to be in a very special place within those companies. But then, such special places are bound to exist elsewhere as well, in particular in small startups that are doing genuinely interesting things.

⬐ ori_b

> Amazon, Alphabet, Intel, Microsoft, Apple, SpaceX, etc [1]. Just look at what companies are pumping 10's of billions of R&D dollars into.

What fundamental research are they pumping money into, regardless of imminent commercial promise? What set Bell Labs apart was the level of fudnamental "Well, we have no idea where this will go, but we can afford to throw some of our monopoly money at it" environment.

⬐ rpeden

For Microsoft, at least, take a look at work done by Leslie Lamport and Simon Peyton Jones while working for Microsoft Research. Look at the papers they've published, and also what they work on - things like TLA+ for Lamport and Haskell for SPJ. I'd argue that both of these guys do interesting and important CS work that doesn't have an immediate commercial payoff for Microsoft.

They're far from the only ones doing good CS research at MS; they're just the first that came to mind.

⬐ jplayer01

I'd say the only one really doing blue sky in any sense of the word is Microsoft. Most of their stuff in R&D never comes to market and is never intended to as its entire purpose.

⬐ ezoe

You mean, Windows?

⬐ ygra

Windows isn't exactly a Microsoft Research project ...

⬐ TheOtherHobbes

Not deliberately.

⬐ pjmlp

However, many technology stacks in use on Windows like WSL picoprocesses, .NET generics, MDIL later replaced by .NET Native, Secure Kernel, Z3 theorem prover for drivers, P language for Windows 8 USB stack, ... have their origins in MSR work.

⬐ coldtea

Xerox Parc fits your question, but the interesting parts of Bell Labs, i.e. the ones we use today (UNIX, C, etc) wasn't any "fundamental research" either.

⬐ pjmlp

Actually had Bell Labs been allowed to sell UNIX and we most likely wouldn't be using it today.

⬐ AsyncAwait

> but the interesting parts of Bell Labs, i.e. the ones we use today (UNIX, C, etc) wasn't any "fundamental research" either.

What about transistors? Don't we use them today?

⬐ coldtea

We do, but that's not the part of Bell Labs people usually reminisce about in a programming context or the era TFA talks about.

⬐ chillee

I'd say things like AlphaGo are "fundamental research". Really, any of the top tier AI labs do what's often considered blue skies research, and have a huge concentration of talent.

Microsoft also has a long legacy of funding a large variety of CS research.

⬐ ori_b

Bell labs didn't just do CS research, and the CS researchers weren't separated from the other researchers. The physicists worked with the chemists who worked with the computer scientists who worked with the theorists who worked with every other field.

⬐ menudo

Is it me, or are there 3:42 (three minutes, forty two seconds) of silence at the start of the video?

⬐ yesenadam

It's not you.

⬐ apetresc

It's a recording of a Google Hangout, which starts recording once the organizer starts the meeting. But then they waited a while for all the participants to join up.

⬐ hhh

This was an absolute treat, and I'd love to see more talks like it.

Huge amount of personality and a long-standing relationship with every topic and person brought up. I'm happy I got to see it. I hope to personally be able to talk about modern equivalents of such amazing topics in the future.

Great voice, great topic, great personality (at least what you can ascertain from a presentation. Worth a listen/viewing.

⬐ nialv7

> great personality

Not to disagree, but what you can't see from this lecture is that Rob Pike is... very opinionated.

For example: http://harmful.cat-v.org/software/dynamic-linking/ and http://harmful.cat-v.org/cat-v/

⬐ danso

Made me think of Brian Kernighan, another Unix guy who’s still kicking and quite engaging. Googling their names together brought up the fact that the two authored a 1999 book named “The Practice of Programming”, and then -$30 from my wallet for the impulse buy.

⬐ vram22

They also authored "The Unix Programming Environment".

⬐ tcard

Both are worth every penny.

They showcase delightfully a way of understanding and writing software, a kind of down-to-Earth simplicity aesthetics that is usually ditched in favor of sophistication.

I believe we're mostly past the point of favoring such philosophy (the Go programming language being maybe the most recent attempt, only partially successful at it), as it has fallen short of what the world demands from software, but it remains a very valuable influence.

⬐ vram22

>Both are worth every penny.

Totally agreed. I cut my programming teeth on UPE and the K&R C book (2nd, ANSI edition), among others. I try to model the command-line tools I write on the style shown in those two books (Kernighan is a co-author of both); being doing that for a long time. Here's an example, a tutorial I wrote for IBM developerWorks - I mentioned it in an HN thread a while ago:

https://news.ycombinator.com/item?id=17327807

>They showcase delightfully a way of understanding and writing software, a kind of down-to-Earth simplicity aesthetics that is usually ditched in favor of sophistication.

True.

>I believe we're mostly past the point of favoring such philosophy (the Go programming language being maybe the most recent attempt, only partially successful at it), as it has fallen short of what the world demands from software, but it remains a very valuable influence.

Right, and The Go Programming Language book has Kernighan as a co-author too.