Computer Science - A Guide for the Perplexed • Joe Armstrong • GOTO 2018

Computer science is not a science ?
Errrh. I stopped watching at that point.

Computer science is not about computers, and nor about science

A must listen for someone starting in this field and something I wish I was adviced when I was starting. Over the years working in the field I've come to same conclusions on lot of topics that he addressed here.
This and Richard Hamming's You and your research are 2 great advice talks .

you not only need to restart your cable modem, you also need to restart your Dreamliner aircraft every 248 days as it turned out recently, to avoid an integer overflow.

Funny, I have those three books on my bookshelf.

Computer Science is not a science and there are far too many ideas floating around.

My backlog is now increased :)

These old timers are so negative...... damn young folk and walkmans and boom boxes.....

RIP Joe Armstrong

RIP Joe Armstrong...you will be missed...

A true gentleman who has made the world a better place. Everyone should Joes thesis, I think we are obligated.

I met him in 2008. He was a nice guy.

He maybe wrong about pipes in some or other way they are being used in modern day app

I'd add interpersonal communication to the list, even though he did mention it implicitly. Sad to hear about his passing... love his work.

RIP

this choosing problem is very relevant. in 1995 nobody in CS had to do that. now, since many teachers have 20 years of experience, they somehow still cannot do that and thus very difficult for students, that often pushed 100+ list of things in courses and same problem: "what is the most important thing? - I dont know". also in other fields in university, since explosion of information and many things have been there also(computers helped create more research and literature in all kinds of fields, previously eg. sociology experiment data had to be on paper managed so this made it expensive and slow, which helped keep information and material very limited).

in current job, this also applies (choosing and deletion problem) to database. Since database has gathered many "evolutions" of program data, some older data is not valid by newest program and company specs. but it is very difficult to remove it or even update it. and this old data can easily cause unnecessary trouble in many ways.

Fantastic talk

Any one ELSE walking like he did would made me crazy. Not Joe... RIP Joe, you are missed.

Hardware doesn't become faster anymore, so we have to switch languages again to take advantage of distributed and concurrent execution.

Absolutely brilliant talk! The world misses people like Joe Armstrong!

Legend

Cameraman tracking Joe must be tired after the talk

'now the program compiles in the time it takes the spring to recoil from the enter button' - er, please tell me what build system this guy is using

All the things he says are more of less true for every industry.

Is he drinking Walmart's Sam's Choice branded water?

The Mythical Man-Month reminds me of the IBM man-year; that's 730 programmers trying to finish a project by lunch time.

What a great sense of humor, what a great mind, what a great speaker. The world lost a great one. RIP professor, you shall not be forgotten.

The computer revolution has not happened yet - Alan Kay
Computers for Cynics - Ted Nelson
Free is a lie - Aaron Balkan
How a handful of tech companies control billions of minds every day Tristan Harris (TED-Talk)
Matt Might - Winning the War on Error: Solving Halting Problem, Curing Cancer - Code Mesh 2017

YAY... I've been saying "C.S. is not a science" for a good few years now... nice to hear someone with far more knowledge and kudos agreeing with me. :)

This is waste of time

I’m not sure what emacs was like when Joe Armstrong was learning it, but this was my experience literally today:

I made a minor change to my emacs config, named org-roam.el, and proceeded to compile my packages.

Then I noticed a typo, fixed it and recompiled my packages (I don’t compile my configs, so why I often need to compile my packages to solve some issues is beyond me)

Trying to use org-roam, it told me that 1. compiling emacssql failed and 2. the package org-roam didn’t provide the previous typo.

The correct answer was to manually delete the stale elc-file, but I didn’t catch the first error message on my first try, and had to search through my private config and the source code of org-roam for the typo. I couldn’t get a backtrace because nothing failed when trying to make a new note (nothing worked either…) and I obviously didn’t find the typo I already fixed.

Apparently my mistyped hook messed up the .elc file and it was left as a stale file despite recompiling my packages?

I love emacs, and especially the modern features. But it definitely is no exception when it comes to software getting more complex.

I miss Joe

Charles Moore has also made a programming language, operating system, and multi-core chips. His software has run on Mars! But he doesn't write books.

Great presentation. Thanx

"Computer Science is not a science."

SICP folks said that as well.