Computer Science

loltospoon

00:12

Hello all

What are some opinions on creating an operating system that can control another operating system like Windows or Linux? I'm a huge dreamer so I think of things like making JARVIS (from Iron Man) that will be able to help me control other systems

Gilles

01:08

@loltospoon That already exists. You can control Windows or Linux from Windows or Linux.

No idea what JARVIS is.

(“control another operating” system is an extremely vague description, but general-purpose OSes can control each other for any reasonable interpretation of the term that I can think of)

loltospoon

01:32

@Gilles maybe I should have added "assume we are looking for the non-trivial answer here". Thought that was implied.

In windows, you can make programs to control things in the Windows environment.

Now, let's say that I built my own operating system. Is it unreasonable to think that I could integrate it to work with Windows? Unreasonable meaning that the amount of time it would take to make that a possibility is too large to be fathomable.

Bounty!

08:12

6

Q: Type-classes for type inference

I'm creating a semantic analyzer with type inference. For the basics I've got a type variable and a type construct with name and a list of types. I want to support overloading and I know that Haskell uses something called type-classes. So if I understand correctly, when I call an overloaded func...

Evil

14:48

@loltospoon one tiny app, a socket with popen to shell, and you have remote access and if something is missing transfer it. What does it mean to integrate two systems? It seems that you want to execute commands from one system to another. It doesn't even need any integration (simple running app would do).

Otherwise your system would have to be compatible with targeted OS and vice versa. The second part is problematic. And if you want to dream big, why even bother to understand them, teach your AI how to use it from users and programatic perspective.

Devilius

15:34

Hello, about quicksort median of three: I understand the logic of median of three for an already sorted array: it lets us pick the middle element, and thus avoid the worst case O(n^2). However, why do we do the median of three? What would be the problem with always picking the middle element? what is the advantage in sorting these three values?

vzn

15:49

0

Q: CS angle AMA "ask me anything" jun13 Bernardo Meurer

earlier this year Michael Wehar proposed a CS AMA "ask me anything" session that had strong support but which has not gone further/ advanced so far beyond conceptual stage. this is to announce not exactly that event but one related. on jun13th Bernardo Meurer will host the Physics AMA. he has str...

discussion chat

↑ plz star this! :)

@Devilius welcome to CS. the median-of-3 is a quick way to estimate the average/ median value of the entire set of data without much work. quicksort is optimum speed when the "pivot" element is near the average and worst case scenarios are when it isnt. you can try this out with experiments. to the contrary the "problem" is more with picking elements that are not close to the median. ps whats up with your av, is that you or an actress etc?

Devilius

it's ginny weasley

from harry potter

film 2

vzn

@Gilles missed my posted links/ earlier dialogue on that huh? :( :P theverge.com/2016/12/20/14028744/…

Devilius

also, thanks for your answer. I've been reading about it and am slowly understanding exactly what you've. Essentially, we do median of three because we want to get the median, but cant do it really efficiently, so settle on an "approximate" median based on three integers. correct?

vzn

@Devilius exactly correct. what language do you use? highly encourage you to try some experiments that quantify this effect/ balance, its not hard at all to do.

Devilius

Java

I've been running tests, that's pretty much what made it clear as to why the random pivot choice also had this pathological case

vzn

15:57

@Devilius yeah quicksort is built into the std library in Java & many others.

@Devilius good! its also possible to show that increasing the time to get a more accurate median is "generally" at the overall expense/ performance tradeoff of the algorithm. ie in some algorithms, approximations actually work better than exact values so to speak.

Devilius

ah right

that makes a lot of sense

vzn

@Devilius which reminds me there is a whole new science of building cpus with inexact computations which decreases energy consumption yet is "good enough" accuracy. think it will continue to grow. it seems to be a quite general concept/ phenomenon/ theme in CS vzn1.wordpress.com/2017/05/19/…

@loltospoon yes this is rather vague and general area is already very highly developed. you might also look into virtual computing.

Devilius

thanks for the link

nbro

@loltospoon First of all, bear in mind that creating an operating system is a complex task. Second, it depends on what you mean by "integrating it to work with Windows". It's very difficult to give you a concrete answer, since you're asking a vague question.

vzn

16:32

in theory salon, 2 mins ago, by vzn

British Airways Union Blames Massive IT Failure On Outsourcing IT Jobs To India / huffpost

in theory salon, 2 mins ago, by vzn

Are leaders ignoring automation and artificial intelligence? / govnews / see also attack of the jobkiller robots survey / Tmachine blog

Computer Science Meta

18:44

1

Q: CS angle AMA "ask me anything" jun13 Bernardo Meurer

earlier this year Michael Wehar proposed a CS AMA "ask me anything" session that had strong support but which has not gone further/ advanced so far beyond conceptual stage. this is to announce not exactly that event but one related. on jun13th Bernardo Meurer will host the Physics AMA. he has str...

discussion chat

Transcript for