Functional Programming from First Principles • Erik Meijer • GOTO 2012

[@reyou7]

You may not like every part of this video but, at some parts, he just says such valuable things that you cannot hear somewhere else. I encourage you to watch it all to get whole vitamins!

[@NicholasShanks]

6:31 Everyone laughs at the mention of "2020". Heh… *sideways glance*

[@grawss]

12:51 The Microsoft logo changes because a program made that happen. And the code that displays it has a 100% chance of avoiding interaction between mutable states. Am I going to use the Microsoft logo as data input? Am I going to use that teddy bear as data input? Then I'll program it in a way that controls all variables and verifies the data with as much confidence as possible until output. I'd have loved to see someone else run up on stage and take the teddy bear from him during his presentation, and say, "sorry, a side effect you didn't plan for stopped your program from running! That's how the real world works."

[@AK-vx4dy]

Very funny and informative video!

[@drewh22]

The video helped me so much. So enlightening.

[@JoelSjogren0]

Enlightening and funny!

[@clementdato6328]

The more I write programs, the more I find that I don’t want to adjust my code just to please the compiler. I want to write code like it is in Python, and the precision like in C++, and if ever the compiler is not happy, I don’t have to change my runtime behavior, but instead I can prove to my compiler that my code is correct, with zero cost or sacrifice in runtime. I think dependent type and effect system and a generalized resource managing type system all combined is the only way I could happily code in a functional programming language.

[@burakcopur3841]

I think, what he talks about is generic programming and not functional programming. Both OOP and FP are trivial compared to generic programming. They work at a lower-level. This is why you can do generic programming all in OOP, FP, Logic programming etc. because it sits at a higher-level of abstraction. mainly that of interfaces.

[@curiosdevcookie]

06:34 Well, me from future telling you you “ Yes, Java has Closures by now!” Well, almost. You have to make some mental arrangements, though.

[@paganaye]

This world is imperative. But what if in a parallel universe the toy survived?

[@kevinl.4970]

Now Java 8 includes lambda expressions.

[@LIB3RTARIAN1337]

+Kevin Leonel Lopez They aren't closures though!

[@captainheretic]

How does this compare to functional programming languages?

[@ZapOKill]

@@WolfJ no not really... now you also have a complete messed up typesystem

[@viacheslav5574]

Looking at the real world to find out how thing should be is a very bad idea. The world is not an ideal place. Such unconditionally bad things as death, illness, suffering are consequences of "imperativity" of the world. Human memory is short, should we throttle computer memory and processing power to model human faculties? Life is driven by natural selection. Should we follow the example and kill everything that is weak? Horrible

[@natepepin09]

It isn't a matter of how things should be yet of how they are. If you know what something is, you then can address it in the proper way. This is his whole point about making things explicit and how they are just a manner of handling things as they are instead of some idea of what they should be. You might think you are ridding yourself of ridding yourself of elements of thr world by programming in a functional manner, but really you are just handling them in an extreme way. You acknowledge the imperative nature and try to reduce it to 0.

[@heroe1486]

Yet if you could grasp 1/10000000 of how the world or the universe work and be able to translate it into computing you wouldn't have been here watching a 5yo talk. Btw your whole text is nonsense, taking humans' weaknesses and ignoring all of their capabilities that could make programs better, the world is more magnificent than "horrible", such a nonsensical thing to say when you're part of the most dominant specie.

[@a46475]

Functional programming and it's advocacy is psychosis. How many talks have you seen on KZbin advocating functional programming (real "no side effect" FP, not map, filter, etc) end up going exactly nowhere like this? Answer that for yourself. They don't know the difference between math and computation. They don't know the difference between a thing and a model of the thing. They don't know the difference between a mathematical function and a procedure that models that function. They end up saying retarded things like a mathematical function has no side effects, thus this procedure which models the function shouldn't either.