OSDev.org

There are many great men do so much work on developing an OS in this forum, you can implement thread, process. and you could write hard disk driver, NIC driver, etc. so I doubt that What kind of job you can find, you know, there are a few companies who develop OS in the world. What do you think?

Developing an OS in your spare time does not necessarily mean you are an OS developer (exclusively). It just means you are a developer.

You handle tool chains, build systems, intricate testing and debugging, low-level code, working with technical specifications. For many of us here, specs in non-native language. All of it things that look good on the resume of any software developer.

yes, I agree with you, It's very useful for our technology and finding job.

but if your job is developing an OS, I think you will feel more happy than developing other softwares..

so we want to develop OS, but we have to earn money and develop other software, we can't work attentively. and unhappy!

OS developing is hobby.

Haven't you read the Beginner's Mistake's page? Have you even taken the effort to read the header about 'Commercial OSDev' and how it will most likely never work out?

Solar is right. And besides that, you can't really estimate what kind of OS developer they would be looking for. Most people working solely on their hobby OS' try to handle every part of it. Others are mostly capable of doing mainly network drivers, video drivers, or other specific parts of the operating system. There is no way anyone can do and know everything equally well at the same time.

It'd be more reasonable for a company to ask a specific kind of developer instead of a 'general OS developer'. There also aren't many companies in the world that are still looking for people that can code operating systems. You mainly have Microsoft and Apple as commercial OS companies (Linux isn't proprietary). Besides those, there are a few other smaller companies that make more specific operating systems like text-based systems for terminals in shops and such.

Actually, I don't think I would want to work as OS developer.

It's more of a gut feeling, and sure as hell I've once thought different, but I don't think it's an environment I would feel comfortable in.

One thought there is that changes are slim that you work the same job from graduation to retirement. As "generic" software developer, I have a wide range of job openings to chose from if change becomes necessary. As an OS specialist, I'd be facing few choices, certain relocation (which I am loath to do (*)), or the alternative of taking a job outside my speciality.

(*): Owning, and living in, the house of my childhood, roomy bordering on the ridiculous, huge and beautiful garden, everything renovated to our tastes, 50m to the kindergarten, 70m to the elementary school yet quiet and green surroundings, friends and family no more than half an hour away. I cannot imagine better living anywhere else on the world. Would you sell or rent that to live in a rent apartment in a big city where you know no-one? Or go away every Monday to return on Friday evening, not being part of your kids growing up? I wouldn't.

david wrote:There are many great men do so much work on developing an OS in this forum, you can implement thread, process. and you could write hard disk driver, NIC driver, etc. so I doubt that What kind of job you can find, you know, there are a few companies who develop OS in the world. What do you think?

Most hobby OS developers do it for the hell of it.

If you see it as a way of helping you into employment, writing your own OS is unlikely to get you a job developing the next version of Windows, but it might help with embedded programming, where you are basically writing a very crude OS to control something like an industrial power supply. (Crude because the processors tend to be crude, compared with something like a Pentium, and the associated PC hardware.)

Actually, if your main ambition is to have better chances in the job market, I strongly advise against hobbyist OS development.

Many of the things we do here are very specific to OS / driver development, with little use outside of kernel space. Most hobby OS projects don't do client / server architectures, database connectivity, design patterns, Java or C# .NET coding, application GUIs or such things. You would be much better off starting, or joining, some userspace application project.

You guy's aren't thinking of the embedded systems industry which i'm sure will be growing since computers are being integrated into everything. Usually you work at a pretty low level when working with embedded systems whether you are working on the RTOS or not. Therfore, an OSdever is a much better candidate for working on an embedded system than a Java programmer.

Plus once you're out of kernelspace, you have to go somewhere (if you continue at all, of course). Reimplementing the usermode foundation of a full operating system from scratch is definitely a more relevant task. It requires knowledge of every piece of the system, as well as the application and user interface levels, and down to the bare hardware, because you wrote the kernel. Personally, I'm eagerly awaiting the chance to write a truly native toolchain for my OS - doing anything similar wouldn't be the same for another crappy Linux distro.

I want to turn on my computer and know, "I wrote this," down to the very last bit. Only motivation like that can make you truly want to reinvent all the wheels of software - the most arduous task, but also the most interesting. What better practice for any developer position? A man who has bootstrapped himself from the depths of the hardware must truly be a force to be reckoned with, once he has reached his goal.

dude101 wrote:You guy's aren't thinking of the embedded systems industry...

Yes I am thinking of that. You know, at several points of time I applied for a job in that very industry. You know what? They don't care much for hobbyists. What they want is people with a master's degree in Computer Science or Mathematics, with strong focus on Automata Theory, Algorithms, and Digital Logics as well as (usually) experience with hard real-time environments. I don't say they won't ever hire someone with less formal education, but if that's where you're aiming, you should at least make your OS hard real-time (which pretty much kills its applicability as a generic desktop OS).

Hello,

I'm a new member here. I've been a professional OS designer since 1998. I've worked mainly with telecoms (infrastructure products and mobile phones). You pretty much write either drivers or debug software, i.e., real-time logging functions. As far as education is concerned, what is needed isn't really taught at university. You need to be able to read a dump - read a stack and figure out what instruction caused the exception then get the registers info from the dump and figure out what crashed the machine. You also need to be able to understand race conditions with no problem whatsoever - i.e. know when to turn interrupts off (I'm thinking uniprocessors here). Other than that just good C and assembly knowledge.

Anybody who has managed to build their own OS from scratch and get it to run on real hardware would be able to do my job. Any employer who says otherwise is just scared to hire you cause they can't understand what you can and cannot do - these higher degrees are a sort of insurance policy. The manager who hires a useless developer who has a PhD can always say "but they had a PhD!". The manager who hires a useless developer who has no education to speak of looks bad. Anybody who wants to work in OS design should be designing OSes to demonstrate their ability.

Barry
www.barrywatson.se

NickJohnson wrote: Personally, I'm eagerly awaiting the chance to write a truly native toolchain for my OS - doing anything similar wouldn't be the same for another crappy Linux distro.

I want to turn on my computer and know, "I wrote this," down to the very last bit. Only motivation like that can make you truly want to reinvent all the wheels of software - the most arduous task, but also the most interesting. What better practice for any developer position? A man who has bootstrapped himself from the depths of the hardware must truly be a force to be reckoned with, once he has reached his goal.

I agree 110% with that post. QFT.
--------------------------------------

With regards to job opportunities: I never really considered getting some bonus, or elevated privilege at work because of my OS-Dev skills. I do it mainly because I like to know that I made it work, and that I could do it. It's really just a hobby.

That said, I don't expect to gain anything much from it at work: being able to do complex system programming will definitely set me apart from the rest, to an extent, but it seems that the IT industry today is being take over by what I'd like to describe as the Horde of UniMonkeys.

Every 3 years, some institute spawns a new set of generic objects which have been programmed to think that Java is the essence of all programming, and execute random functions (UniMonkey.lookForJob(); UniMonkey.floodTheJobMarket(); UniMonkey.waveUniCertificate(cert cert_level); ), thinking that they are actually instances of the realProgrammer class.

So even though I may obviously be better than the rest of the employees in my environment, as many others here would be, I imagine, (I know WAY more than my Uni Professors already), without that UniMonkey.waveUniCertificate(cert_level) function, I can't rise very high. I'd need to still go and get course training.

The IT world today is trying to strip away those logical, 'irritating', 'stuck up' programmers they had to deal with about 40 years ago, and replace them with a horde of monkeys who will do as their employers say. Never mind that these Monkeys really don't have the skill to properly implement the spec: as long as they can be governed, and their actions controlled like the other employees, everything is fine.

holypanl wrote:

NickJohnson wrote: Personally, I'm eagerly awaiting the chance to write a truly native toolchain for my OS - doing anything similar wouldn't be the same for another crappy Linux distro.

I want to turn on my computer and know, "I wrote this," down to the very last bit. Only motivation like that can make you truly want to reinvent all the wheels of software - the most arduous task, but also the most interesting. What better practice for any developer position? A man who has bootstrapped himself from the depths of the hardware must truly be a force to be reckoned with, once he has reached his goal.

I agree 110% with that post. QFT.
--------------------------------------

With regards to job opportunities: I never really considered getting some bonus, or elevated privilege at work because of my OS-Dev skills. I do it mainly because I like to know that I made it work, and that I could do it. It's really just a hobby.

That said, I don't expect to gain anything much from it at work: being able to do complex system programming will definitely set me apart from the rest, to an extent, but it seems that the IT industry today is being take over by what I'd like to describe as the Horde of UniMonkeys.

Every 3 years, some institute spawns a new set of generic objects which have been programmed to think that Java is the essence of all programming, and execute random functions (UniMonkey.lookForJob(); UniMonkey.floodTheJobMarket(); UniMonkey.waveUniCertificate(cert cert_level); ), thinking that they are actually instances of the realProgrammer class.

So even though I may obviously be better than the rest of the employees in my environment, as many others here would be, I imagine, (I know WAY more than my Uni Professors already), without that UniMonkey.waveUniCertificate(cert_level) function, I can't rise very high. I'd need to still go and get course training.

The IT world today is trying to strip away those logical, 'irritating', 'stuck up' programmers they had to deal with about 40 years ago, and replace them with a horde of monkeys who will do as their employers say. Never mind that these Monkeys really don't have the skill to properly implement the spec: as long as they can be governed, and their actions controlled like the other employees, everything is fine.

That's why I stuck with the embedded world rather than the IT industry.

bwat wrote:I'm a new member here. I've been a professional OS designer since 1998. I've worked mainly with telecoms (infrastructure products and mobile phones). You pretty much write either drivers or debug software, i.e., real-time logging functions. As far as education is concerned, what is needed isn't really taught at university. You need to be able to read a dump - read a stack and figure out what instruction caused the exception then get the registers info from the dump and figure out what crashed the machine. You also need to be able to understand race conditions with no problem whatsoever - i.e. know when to turn interrupts off (I'm thinking uniprocessors here). Other than that just good C and assembly knowledge.

It is alarming that you think these things are not taught to university students studying computer science and engineering. Which university did you attend and for which degree (if you don't mind me asking)?

bwat wrote:Anybody who has managed to build their own OS from scratch and get it to run on real hardware would be able to do my job. Any employer who says otherwise is just scared to hire you cause they can't understand what you can and cannot do - these higher degrees are a sort of insurance policy. The manager who hires a useless developer who has a PhD can always say "but they had a PhD!". The manager who hires a useless developer who has no education to speak of looks bad. Anybody who wants to work in OS design should be designing OSes to demonstrate their ability.

It's not just about doing a job, it's about doing a job well. While I will concede that having success in building an OS from scratch is indicative of the ability to do other system-level programming tasks, it is extremely important to differentiate between the propensity to produce a naive solution versus a cognizant solution. Furthermore, the company typically expects the solution to be accompanied by additional deliverables, which requires a proper set of technical communication skills.

It's certainly plausible that someone without a degree could have otherwise developed an equivalent--or better, even--repertoire as someone else who has a degree, but from the company's perspective it simply isn't worth the chance.