> At 17, Armstrong began programming Fortran on his school district's mainframe. This experience helped him during his physics studies at University College London, where he debugged the programs of his fellow students in exchange for beer. While working for the Ericsson Computer Science Lab, he helped develop Erlang in 1986.
This is pretty legendary. He was quite the programmer.
RIP Joe. Every interaction I had with you in the past was always a great time. You'll be missed, but never forgotten. Thanks for everything you've done for the Erlang community and distributed systems design.
This is indeed sad news (I feel he deserves the black bar of mourning once confirmed). Although Elixir gets the most attention these days the solid engineering of BEAM is the foundation that Elixir is built on.
Joe was also a part of the Elixir community. He learned the language early on and participated in the mailing list and forums. Actually, he had just very recently become an admin at elixirforum.com. I will really miss him.
Candy-coated Elixir, which tries to hide concepts that are difficult for people who approach programming without the requisite mathematical skills, may be getting most of the attention, but serious mission-critical "nine-nines" projects are still being done in Erlang. I program in Erlang every day for a system that can't (and doesn't) have any downtime.
That's not what you're doing. You're shitposting about Elixir developers--while Joe was enthusiastic about Elixir to the point of being a mod on elixirforum.com.
"Keep your friends close and your enemies closer."
If you knew Joe, you'd know he was strict about feature-creep in Erlang. He removed Paramaterized Modules, even though it broke some popular Erlang projects -- including the Chicago Boss Web Framework. This was one of the reasons Elixir was created. To make it more "approachable" to people who didn't have the inclination to be FP purists.
That is so sad, he was in his late 60s and it seemed like he had a lot of life left in him, talking about how software can get better and having a great (and sometimes snarky) outlook on the profession.
Highly recommend his thesis (2003) and a few of his great interviews/presentations for anyone who isn’t familiar with Joe, it captures a lot of what he thought about and pushed for in his professional life.
It goes much further than that, it shows how to tackle reliability even in systems that are not distributed. The primary insight is that all software will be buggy so you need to bake reliability in from day one by assuming your work product will contain faults.
Yes, I know. Erlang was not distributed till 1991, roughly 5 years after it was born.
It's also really illuminating how they implemented the first versions of Erlang as a reified Prolog [1]. But that is not explained in the thesis, just in his 1992 paper which he briefly cites.
> That is so sad, he seemed like he had a lot of life left in him talking about how software can get better and having a great (and sometimes snarky) outlook on the profession.
I'm in actual shock, he was tweeting about pretty much that (also brexit and playing with his phone's voice recognition) just 2 weeks ago… He wasn't even 70…
"And then we've got this sort of dichotomy between efficiency and clarity. You know, to make something clearer, you add a layer of abstraction and to make it more efficient you remove a layer of abstraction. So go for the clarity bit. Wait ten years and it will be a thousand times faster, you want it a million times faster, wait 20 years."
That's... one way to make something clearer. It's also a way to hide complexity behind leaky facades. "Just add a layer of abstraction" is horrible advice.
It is a great talk. Based on that I took his course on Erlang at FutureLearn, which was also very good. Sadly, when I sent that link to some colleagues at work they shrugged and have since kept adding more and more dependencies and complexity to the front-end build. It's very hard to stop momentum once it's got going, culturally.
Thanks a lot, I never watched this, it’s absolutely awesome.
I laughed a lot and
I actually cried in the slide about legacy code when the first line was “programmers who wrote the code are dead” :(
Guess I'm alone in not seeing what's so good about this talk. He only presents problems, and in an extremely disorganized way. The closest thing to a solution is "wait, hardware advancement will eventually make your slow code fast."
It's definitely entertaining, but that's about all I can give it.
There's a time and place for things. Posting multiple critical comments in a commemoration thread is maybe not the best time or place, and better kept for later?
The idea of "no copies – everything exists in one place and has a unique ID" was new to me. I still don't know if it's a good idea, let alone practicable, but it's great food for thought!
Very sad news. Thank you to HN mods for commemorating Joe's passing with the black bar. His contribution to our world is immeasurable. Although I never had the pleasure of meeting him in person, by all accounts he was much loved. Joe you will be missed.
I like the black bar, it shows both mourning and respect.
Occasionally I'll see the black bar and the notice and not even know who the person was - at which point I get to find out about someone amazing I didn't know existed and to appreciate their contribution.
Obviously I'd prefer that there not need to be a black bar though.
This is incredibly saddening. Joe has left an incredible legacy for us to continue on. I'm saddened that I never got to meet him and thank him for all that he has done.
I was following him on Twitter, and it was always amazing that a genius like him took the time to respond to people on Twitter with insightful comments. Showed what a great human being he was. RIP Joe.
Not just that, but he also struggled in public with things we all struggle with, often with incredible humor. His struggles with Grunt.js come to mind. His humanity and charisma was in full force, and you can tell he worked in full grace and gratitude. His work of course is a testament to this.
RIP.
Thank you for Erlang, the great programming language, it saves me a lot of time. And thank you for the book: Programming Erlang. It guides me to think the programming world in different, interesting, and simpler way.
This recent panel (Armstrong, Hewitt and Hoare) discussion on concurrency was the first time I got to know about him and it gave me a lot of food for thought.
I met Joe in 2012 when I spoke at the Erlang Conf in SF. he is the reason I discovered functional programming and the whole concept of message passing (actor model) in distributed systems. Quite a legend and contributor to the world of computer science. RIP joe.
I also met him at a conference in '16. He gave a talk but mostly wanted to shoot the breeze in the hallway with fellow hackers. He lamented scheduling gripes with his PM, being a slave to Jira, and some new things he was playing with. Very approachable and humble.
He asked a few of us hanging around what talks we were thinking of attending and we agreed to check out something about concurrency in some modernish platform. The speaker was working on issues with handling failures, message passing, memory management, etc, but then realized Joe was in the back and pointed out it had all been solved many years ago by that guy over there. Everyone laughed and gave Joe the acknowledgement.
I don't feel like we've finished learning from and building on his work. RIP.
Life is short and Joe went much younger than one would expect. When I find myself doing something that I do not enjoy, I'm always trying to meditate to the moment of my future-death in order to psychologically slap myself in the face and re-evaluate what it is that I'm doing. We only get so many years and we should strive to use them to the best of our ability and potential. Joe Armstrong certainly did that. RIP.
205 comments
[ 3.5 ms ] story [ 195 ms ] threadBlack Ribbon please.
https://twitter.com/bodil/status/1119644170816303104
1. https://pbs.twimg.com/media/DBs4Q53XkAI6yKd.jpg
2. https://pbs.twimg.com/media/DAs_wqVXsAEvTor.jpg
3. https://pbs.twimg.com/media/Ds19NLdXgAAMXho.jpg
This is pretty legendary. He was quite the programmer.
Let people mourn how they will.
Stop.
If you knew Joe, you'd know he was strict about feature-creep in Erlang. He removed Paramaterized Modules, even though it broke some popular Erlang projects -- including the Chicago Boss Web Framework. This was one of the reasons Elixir was created. To make it more "approachable" to people who didn't have the inclination to be FP purists.
–Hello Joe
https://www.youtube.com/watch?v=uKfKtXYLG78&t=2m7s
https://www.youtube.com/user/gar1t/videos
The 'MongoDB is web scale' one is pretty funny and was widely watched shared in the day.
I wish he'd make some more videos.
Thanks, great link.
Highly recommend his thesis (2003) and a few of his great interviews/presentations for anyone who isn’t familiar with Joe, it captures a lot of what he thought about and pushed for in his professional life.
http://erlang.org/download/armstrong_thesis_2003.pdf https://youtu.be/fhOHn9TClXY https://youtu.be/lKXe3HUG2l4 https://youtu.be/rmueBVrLKcY
I hope his family and friends can find some comfort in how much he was appreciated and admired in the development community.
It's also really illuminating how they implemented the first versions of Erlang as a reified Prolog [1]. But that is not explained in the thesis, just in his 1992 paper which he briefly cites.
[1] https://citeseerx.ist.psu.edu/viewdoc/download;jsessionid=F4...
I'm in actual shock, he was tweeting about pretty much that (also brexit and playing with his phone's voice recognition) just 2 weeks ago… He wasn't even 70…
https://elixirforum.com/t/introducing-our-new-moderators-and...
https://youtube.com/watch?v=lKXe3HUG2l4
[0] https://gruntjs.com/
Delivered that line with perfect timing and humility. What a great sense of humor!
- clarity vs efficuency on abstraction
- the need to unwind (not sure if it is the right word) entropy
- naming and comments
- etc
Things were not so obvious to me when just started my career but are so important to software engineer's day to day
RIP Joe. Your ideas will live with us, forever
It's definitely entertaining, but that's about all I can give it.
I kind of doubt Joe Armstrong would support N days of mourning requiring uncritical acceptance of everything he's ever done.
Joe called things like he saw them. I think we better honor him by doing the same.
Occasionally I'll see the black bar and the notice and not even know who the person was - at which point I get to find out about someone amazing I didn't know existed and to appreciate their contribution.
Obviously I'd prefer that there not need to be a black bar though.
Feels extremely appropriate to me, and honestly, it was the first thing I looked for after I saw a tweet about this. RIP, Joe :(
Not to diverge too far, but does anyone know who the first black bar was?
https://twitter.com/FrancescoC/status/1119604272000315394?s=...
I'll remember Joe by the several insightful, entertaining talks he's given in recent years. https://www.youtube.com/watch?v=lKXe3HUG2l4 in particular.
This recent panel (Armstrong, Hewitt and Hoare) discussion on concurrency was the first time I got to know about him and it gave me a lot of food for thought.
Afterwards looked up talks by him and was fascinated by Erlang, sad to see him pass away.
He asked a few of us hanging around what talks we were thinking of attending and we agreed to check out something about concurrency in some modernish platform. The speaker was working on issues with handling failures, message passing, memory management, etc, but then realized Joe was in the back and pointed out it had all been solved many years ago by that guy over there. Everyone laughed and gave Joe the acknowledgement.
I don't feel like we've finished learning from and building on his work. RIP.