/tech/ - Technology★

yes hello this is him

my main os is AmigaOS, and I only program in BASIC

bing bing wahoo fagioli fagioli give me the botnetoli

Daily reminder that Lisp will compile this without raising any warnings or throwing an error:

(defun add-5 (x)
  (+ x 5))
(defun this-crashes ()
  (add-5 "hello"))

You will have no idea that you have a type error until you run it. That is, unless you have provided type annotations and set optimizations to speed, in order to gain the fabled "performance even faster than C", in which case Lisp could make demons fly out of your nose.

>>918026

Here is what SBCL does:

* (declaim (optimize (speed 3) (safety 0)))

NIL
* (defun add-5 (x)
(declare (type fixnum x))
(+ x 5))

ADD-5
* (add-5 "hello")

34385326708

Lisp will happily chug along with the wrong result and you'll have no idea where the error lies.

>>918028

Dude, that's useful to find what's wrong. What was anyone thinking?

>>918026

>Lisp

"Lisp" is not a compiler, nor is it a language. Your complaint sounds like one about a specific compiler.

>>918028

<If I turn off the safety, then my program is unsafe

Certainly.

>>918048

> "Lisp" is not a compiler, nor is it a language. Your complaint sounds like one about a specific compiler.

All Lisp languages are the same in this regard. As for compilers, the Common Lisp standard says that type annotations exist, but now what they do. Find me any Common Lisp compiler that will give me the performance and the type safety.

> Certainly.

The point is that in order to get Lisp to be useful for systems programming where performance matters you have to throw out all the great things about Lisp. People say that Lisp can be as fast as C, but they never mention that you have to take the safety off, making it even more unsafe than C. That's like saying that a motorbike can be as fast as a car, but you have to take off your helmet to reduce air friction for it. You would have to be crazy to do that.

>>918057

> Find me any Common Lisp compiler that will give me the performance and the type safety.

I don't know of any that currently exist. It is also very difficult to solve this for all cases, but for simple ones like your example it should be possible.

I like Wirthian languages. I should play around in Oberon

http://www.projectoberon.com/

>>917987 (OP)

If only lisp machines did not always suck.

>>917987 (OP)

Define Unix.

>>918026

>>918028

>Lisp will happily chug along with the wrong result and you'll have no idea where the error lies.

ofc lisp will do what you tell it to do. if your program has a bug in it (like your little snippet has) then the problem is that you are an idiot.

>>918283

I am not op, but I am pretty sure that if someone would make a modern lisp machine, then it wouldn't suck

>>918276

Thanks for the link! Oberon looks interesting.

>>919018

>ofc lisp will do what you tell it to do. if your program has a bug in it (like your little snippet has) then the problem is that you are an idiot.

That's pretty dumb. Of course some kind of bugs are hard to prevent, because the compiler can't know for sure what you really want to do, but with a properly designed language, it should be able to tell what you want to do. Certain types have certain associated operations; this is type theory 101.

>>919288

*it should be able to tell what you don't want to do

Goddamn.

use Windows, games work and you never have to see the command line

peace

>>919314

use xbox, games work and you never have to see the command line

peace

>>919337

use pong, game works and you never have to see the command line

peace

>>919337

exactly friend

cant wait for the telemetry dildo up my ass