The OS Graveyard

I was searching around in the Operating Systems subforum, and I found a graveyard.
Everywhere I look there is a discontinued OS made by a new member. What is the rush to make an OS? OSs most of the time add barely any functionality to the game. But still, every single new member I meet seems to try and make an OS. And they post it, get feedback on how they shouldnt be making the OS, lose hope, discontinue the OS and never post anything again. It is sort of sad. It's like OSs are new member traps, and good OSs are bait. They all see something cool like OneOS, then they get inspired, then shot down. It's pretty sad. People should stop focusing on making OSs and make cool programs, like a controllable RS gate, or a computer controlled base with piston doors and controllable lights! All I'm trying to say is, don't make an OS unless you are 100% confident in your skill, and don't post it till you have most of everything you want.

Now I have a question. What do you think drives the OS craze?

I honestly think people want to get recognized for something, look at the tutorials, learn what they think is enough to make an OS (due to looking at all the other failed attempts & NDFJay tutorials) and slap their name/image on it. I've even tried to make an OS a couple times, although they never amounted to anything I'd want to release. I'd think they would look at OneOS's source code and run for their lives, but sadly that hasn't happened.

I'm trying to remember back to why I made my first OS, PearOS. If I recall correctly I'd just played around with a Mac classic emulator and wanted to make something like that for ComputerCraft. Although there must've been a bit more than that.

I think you've sort of answered your own question really. People see OSs as the pinnacle and the coolest programs so they try and make them. It's also a lot easier to make a CC OS than a normal OS too obviously. But yea, I'm interested to see what others think.

There's no reason to discourage people from making OSs. Yes, there are already a lot of them, and yes, most of them aren't terribly useful. However, no one should be discouraged from writing the code they're interested in.

Separately, feedback on any posted program should be constructive. If you see posts that are just negative and offer nothing constructive, please report them. We try to stay on top of these posts as well as we can, but we can't be everywhere at once.

Lyqyd, on 19 November 2014 - 09:06 PM said:

There's no reason to discourage people from making OSs. Yes, there are already a lot of them, and yes, most of them aren't terribly useful. However, no one should be discouraged from writing the code they're interested in.

The thing is though, most of the time when people discourage them from writing Operating Systems, they're trying to encourage them to write something simpler, so they can slowly build up their skill and hone it. Then in a couple of months, they can try to write an Operating System again, and it might actually be the next OneOS.

It's a very safe bet that any given coder around these parts doesn't even understand the difference between an OS or a shell. It's rather difficult to muster any enthusiasm to "encourage" people to work on a project when they don't even know what it is they're trying to write… Simply explaining it is enough to put most people off.

Why do users want to write their own menu systems and whatnot? Because they think CraftOS is too difficult, clunky, or ugly to use. Which brings up another ironic point - if they're of that opinion, then odds are they're also in no way qualified to improve on it. That's not to say that it can't be improved on, but yeah, pretty safe odds.

So you've got a scenario where users try to pick a project before they even remotely understand what it entails. Odds are they would pick something easier, if they knew enough to understand what's easy, what's a decent challenge, and what's so completely out of their league that they'll lose all interest before getting anywhere!

Any coder can work their way up to the task, but you've first got to learn to crawl, and all that.

And then there's the whole matter of practicality. I can't speak for everyone, but nearly all my ComputerCraft systems boot a startup script then do nothing but run that one script. They might manage a reactor, or monitor players, or poll inventories, chop wood, act as a hopper, whatever… but that's all they do. Ever. If I want a different task performed, I install another system. I don't have to keep coming back and restarting them or whatever. The remaining units are there purely to test code on; if I even bothered to install a fancy OS, I'd never get around to using it.

On real computers, my view of a decent OS boils down to something with a decent file manager which otherwise stays out of my way.

Once upon a time door programs were being posted left right and centre, so it was interesting to see the shift to OSes, sadly the shift was largely caused by the NDFJay 'tutorial'.

Personally however, I do prefer the extra ambition that people seemed to have mustered to think they can create an OS, it's far more complex and you can learn a great deal more from it than from making a door lock. That being said though, too many people go into ComputerCraft programming with the knowledge of an if statement and shell.run and think they can make an OS, and that's the kind of thinking that — in my opinion — should be discouraged from an early stage, they should be redirected to focus on something simpler in order to hone their skills and learn new things, and then once they're feeling more confident with the language to move onto something more complex; look at teaching children at school, do we all start with calculus? no, we start with 1+1 and go from there until we're able to understand the grander, more complex concepts, the same thinking should be applied to programming. Take a look at any education institution that teaches programming, they start with something simple and work their way up, for me at University that was beginning with Pascal and learning about procedural programming and the concepts that revolve around that, then slowly moving through Object-Oriented programming, all the way until this year where we were tasked with the creation of a compiler/parser for our own language, something very few, if not none, of us could have done back in the first year.

But back to point, a lot of the reason the OSes fail is simply due to the complexity of it's source; they generally have very unorganised code which is all over the place, with little to no indentation, and a very high cyclomatic complexity, this causes them to have no understanding of their own code, so when a bug suddenly pops up that they don't know how to deal with, forcing them to abandon their OS.

A final point is that it's difficult to think of unique ideas for most people, it is far easier to get an idea someone else has done and try and improve upon it. Don't forget a lot of the forums (and Minecrafts) user base are very very young, they see something someone has made (like OneOS) and how much love it gets from the community, and they want a piece of that action, so they try their best to make something that their skill level just isn't capable of.

theoriginalbit, on 20 November 2014 - 01:44 AM said:

… look at teaching children at school, do we all start with calculus? no, we start with 1+1 and go from there until we're able to understand the grander, more complex concepts …

Ahem, I beg to differ: http://www.amazon.com/Introductory-Calculus-For-Infants-Inouye/dp/0987823914

Cranium, on 20 November 2014 - 08:20 PM said:

theoriginalbit, on 20 November 2014 - 01:44 AM said:

… look at teaching children at school, do we all start with calculus? no, we start with 1+1 and go from there until we're able to understand the grander, more complex concepts …

Ahem, I beg to differ: http://www.amazon.com/Introductory-Calculus-For-Infants-Inouye/dp/0987823914

Who ever thought of that as a business idea was a genius. I can't at anyone other than maths PHDs buying it though. There's even a quantum mechanics one too.