Is embedded development comfier than most other software jobs?
Bumping for cozy confirmation
>>55020497
It's most likely harder to get into, the cash is alright, and if you become an expert you're then you're set.
>>55020497
Is that a cherry blue switch?
If you deploy bad code then you're nothing less than never getting a job again.
Even one fucking bug.
>>55020805
>>>>>implying
>he deploys his embedded software with less than 100% test coverage
>>55020497
obviously
>>55020799
no you dumb ass, that's android
>>55020799
Probably just a push-on/push-off switch
>>55021138
I would laugh so hard if it really was a mechanical blue switch
>>55020497
>asks about embedded development
>posts an FPGA devboard
Now, if I were that special kind of asshole, I'd ridicule you for this and do my best to make you feel stupid on the internet because I'm that insecure.
But I'll just simply mention, since they do get confused for one another all the time, SBC's, Microcontrollers, and FPGA's all tend to be interchangable to the untrained eye, BUT, FPGA's aren't "embedded" per se, what you want to refer to is "digital design"
>>55020497
>embeded
>posts FPGA
Nice meme.
Also
>Is embedded development comfier than most other software jobs?
Depends on the complexity of the project you need to develop if it's a simple controller board it's just defining a few functions and your done.
If you need to control something massive and do automation then you're fucked with the amount of circuits and backups you gotta make.
>>55021402
Actually those buttons are mechanical push spring buttons with a holding mechanism.
However they are mostly used for powering on or off devices.
>>55022524
Do you get to use any cool math like control theory?
>>55020805
>If you deploy bad code then you're nothing less than never getting a job again.
>Even one fucking bug.
Most of the embedded systems that my work uses run old, exploitable services. The RFID readers run an exploitable version of httpdpd or whatever the fuck.
DOES THIS LOOK LIKE THE FACE OF MERCY?
>>55023278
yes
>>55023278
YOU CANT EVEN BEGIN TO IMAGINE THE SUFFERING THAT IS XILINX ISE
DO YOU WANT TO GENERATE A CORE? FUCK YOU, SILENT CRASH WITH NO TRACE
YOU WANT TO JTAG? FUCK YOU SILENT RANDOM CRASH WITH NO TRACE
YOU WANT TO USE THE SHITTY IDE INSTEAD OF A FUCKING MAKEFILE JUNGLE? HERE'S SOME THREE SPACES PER INDENT DEFAULT. ALSO .XISE IS CORRUPT SO YOU HAVE TO REMAKE THE ENTIRE PROJECT BECAUSE NO ONE KNOWS HOW TO FIX IT.
REEEEEEEEEEEEEEEEEEEEEEE
>>55023278
>>55023301
FUCK XILINX
>be me
>love low level stuff
>1st year university
>think about how embedded must be so cool
>FPGA exam is easy, just use already built adders and multiplexers to add numbers depending of which button is pushed
>fail completely with 10/100 mark
>2nd semester
>Have C class
>Though I'd love it since I like fairly low-level shit
>Fail the class at the end because I couldn't do the final project "travel agency" of some sort
Embedded looks really comfy but I see I'm fucked.
>>55020497
Embedded is that one thing where it all seems really straightforward and comfy, and programming seems simple.
Until your program bugs the fuck out for no apparent reason at all, systems crashing everywhere.
Not to mention the need to sometimes use retarded ASM languages.
>>55023226
Fuck me if I know.
I just make shit work, and by making it work is making all the instructions as fast and simple as possible.
>>55023326
>exam is easy
>fail completely
How, anon? Did you mess something up? How?
Also, don't worry too much about the future. Sort out the present and the future will sort itself out.
>>55023698
This guy knows. I'm an embedded software engineer for an automotive engineer and we've hit a strangely infrequent bug with our data flash system. It's the most aggravating process to even get it set up to run step by step to figure out what's happening.
>>55023326
Fuck you for getting into that program without really knowing what you were getting into. It's chucklefucks like you that caused my professors to make their low level software courses easier.
>>55023301
The Vivado Suite isn't so bad these days. 15.4 was pretty stable as long as you followed the build steps that it wants you to do.
>>55023698
>Until your program bugs the fuck out for no apparent reason at all, systems crashing everywhere.
Only 2 things cause this in embedded:
>Your programming errors (i.e. fucking up a pointer)
>Didn't read the manual properly.
>Not to mention the need to sometimes use retarded ASM languages.
What's retarded about Assembly? It's very basic & direct because it has to be. It's also very transparent. If you understand the hardware & asm then you can see exactly what is going on.
>>55023940
Flash memory is a shit. I've written a few Flash drivers (1 in Assembly, the others in C) & there's always catches & conditions that have to be met before you can program or erase. Plus you can't debug them because you can't access the CPU when it is rewriting itself.
Recently been introduced to TI MSP series & FRAM (ferroelectric RAM). 1 #pragma before the variable you want to store & job done.
>>55024005
>Only 2 things cause this in embedded:
That's pretty much the case in all programming. But it's harder to fix in embedded since sometimes you don't get access to a debugger and have to rely on logic analyzer output.
>>55020497
Former embedded dev reporting in. Absolutely not. Salary is shit, colleagues are shit and work is shit.
>>55023278
ISE is just a huge bag of dicks
>>55024570
What sorts of qualifications do you need to get a job in embedded?
>>55024710
You need to figure out how to do complex arithmetic only being able by dividing or multiplying in factors of 2
>>55020497
Depends on what it is for.
I only have unresearched second hand knowledge, but my friend has a masters degree in EE and he didn't make shit as an embedded developer. He last job in the field was writing laser printer firmware for one of the big manufactures. He now does Java web development and just got hired where I work and will make a killing.
>>55023906
Not him but I have dyslexia and dysgraphia. I have failed many exams due to reading the question wrong or writing something other ( usually to nonsensical results, replacing words with similar sounding ones etc. ) than what I was intending to write.
>>55024710
If you survive riding the lightning, you'll do fine.
>>55024802
That's rough, sorry man.
>>55024551
can't you just simulate the hardware?
>>55024920
In Proteus, maybe, but you can only use/insert a very small number of embedded microcontrollers
>>>55024570
current electrical eng student in final year.
Always thought doing embedded stuff would be pretty fun. is it really that shit?
should I just learn how to software eng?
>>55025334
It was for me. Embedded jobs are hard to come by, companies that have them know this and basically extort their employees and push salaries down.
The thing about embedded is also that you usually get into some niche market and once you get a customer they're suddenly fully dependent on you. It's not like software, where you can just replace the entire software or even cloud service, it's physical things that your customer or client usually incorporate into their own equipment and become totally dependent on.
Which is why you can deliver horribly low quality shit, and still get paid. In my job, the CTO was an incompetent fool that had taught himself C++ and couldn't code for shit. He didn't "trust" git and refused to use it. He didn't "trust" the standard library, so he had written his own containers. Stuff like this. It was horribly, and half of my colleagues couldn't code for shit and the other half could but just chose to do nothing about it because they were dead inside.
>>55024710
I had a master's degree in CS and a bunch of relevant courses as well as my master thesis being on modifications to TCP in the linux kernel
>>55025401
>Which is why you can deliver horribly low quality shit, and still get paid. In my job, the CTO was an incompetent fool that had taught himself C++ and couldn't code for shit. He didn't "trust" git and refused to use it. He didn't "trust" the standard library, so he had written his own containers. Stuff like this. It was horribly, and half of my colleagues couldn't code for shit and the other half could but just chose to do nothing about it because they were dead inside.
This saddens me.
>>55025401
so what happens if you do a good job?
nothing?
seems like its good if you can just lazily code stuff
>>55025429
>so what happens if you do a good job?
If you're lucky, you get hired by another company that is slightly better than your old one.
>seems like its good if you can just lazily code stuff
That's not good, anon. They didn't even do testing, meaning that there were bugs that made it to production which cost us a bunch of late hours AT THE CLIENT trying to debug and test for no increase in salary at all.
>>55025788
so what exactly are the perks of working in embedded design then?
you cant get fired once you're hired?
That it?
>>55025788
also what do you do now? whyd you swap out of it
>>55025814
>so what exactly are the perks of working in embedded design then?
CV fill
>you cant get fired once you're hired?
That too, yeah.
>>55025834
Well, the reason I got into it was because I wanted to do something more meaningful with my life than making DAM systems. The pay was good, I got to do C++ programming, the bosses were super nice, the benefits were extremely good, but it was basically glorified web development with some minor image and video processing.
The problem was that everywhere I went, people just saw "WEB DEVELOPER" because of my experience, nobody gave a fuck that I actually wrote C++ in my work and had submitted a patch to the Linux kernel during my work with my masters and all my courses/classes were basically stuff like "real-time embedded systems" and "pc-based instrumentation and microcontrollers".
I went to this company because I knew a guy that worked there (and he was keen on getting a signing bonus), I showed off my networking skills and programming skills and got the job.
But shit was so horrible there, had to leave before 6 months.... Not only from a technical point of view, but also the CTO was an arrogant dick that thought he could treat all his developers as shit because he was delusional and thought he paid us well, and the director of sales was an alcoholic (as in actual alcoholic, he came to work drunk and usually drank half a bottle of whiskey every day at work). The CEO was a nice sympathetic guy though, and also in charge of the hardware devs, I don't know how he ended up with the others.
Anyway, I digress. My old supervisor from where I did my masters called me and said that they had a PhD position open, so I saw an opportunity to get out from that hell hole and ended up doing that instead.
The pay now is of course
>>55025937
*The pay now is of course less than ever before, but at least I get to do interesting stuff and knowledge, increasing performance and stuff like this is actually valued.
>>55025954
>>55025937
thanks for the story anon.
im actually considering moving to sweden for their ICT sector for a few years
>inb4memes
that or just stay at home in aus.
>>55025937
>>55025788
>>55025401
>>55024570
I thought embedded development was the pinnacle of software development.... Are you saying that /g/ is full of shit and that I was meme'd into doing computer engineering despite the fact that I hate everything about it except the programming classes?
>>55026146
we got memed bro....
>it was just a meme bro
>>55026146
>I thought embedded development was the pinnacle of software development....
Sure, if you're somewhere good.
>>55026181
Like where?
>>55026350
A place that makes critical systems. Places that you aren't likely to be able to get into.
>>55026374
>A place that makes critical systems
I'm >>55025937 and as part of my PhD I work with a company that is a subcontractor to Lockheed Martin (via an American contractor).
You would be surprised to know that the state of the software in the F35 is extremely bad and that there has been cases where the pilots have had to reboot some of the computers because they crashed.
>>55026403
>>55026374
Which is obviously bad for modern fighters (or bomber-fighters) which rely on computers to even stay stable because they are aerodynamically unstable.
Software developers are basically code monkeys whether they are named Pajeet and work for some outsourced Microsoft project or if they're named John and graduated from MIT and work at LM.
>>55026403
>You would be surprised to know that the state of the software in the F35 is extremely bad
Not really, the F35 is a clusterfuck.
Even so, I imagine it's better than churning out boring shit like dildo controllers...right?
>>55020497
Embedded is terrible. The pay is worse than pretty much every other field and there is no market for it. You'd have to be lucky to get a decent job that isn't at an ancient company that doesn't understand software or you are at some startup that is trying to make another shitty iot device that nobody would want.
The actual programming is shit because you are literally reinventing the wheel with every little thing depending on the controller/DSP you are working with often with errors in documentation. The overtime debugging will drive you insane because you can never be sure if there is fault in your logic or the hardware or the shitty documentation. You also don't get to Google errors because chances are, your environment is too specialized to find anything.
Only work in embedded if you just want a shit job and dont want to get anything out of your career.
>>55026653
So what do?
>>55026403
>You would be surprised to know that the state of the software in the F35 is extremely bad and that there has been cases where the pilots have had to reboot some of the computers because they crashed.
It's probably because they're moving from Ada to C/C++ for some retarded reason
>>55024570
>>55025401
Please tell me this is just a big prank. I was really looking forward to getting into firmware development. S-s-shut up! What state do you live in? I-I bet you live in some mid-western shit hole.
>>55026724
>for some retarded reason
That reason being that they had some serious issues in their previous software version and had to scrap everything and start from scratch.
>>55026604
Well, as I work for one of the subcontractors, I'm not involved with that at all. The company I work with is making PCIe-based interconnects, from rugged environments to high-speed computing clusters. I'm more in the latter camp, but I've heard one or two stories about the problems with the F-35.
>>55026786
It is the most saturised programming industry and that completely crashes the salary levels.
This poster gets it: >>55026653
>>55026879
so what field needs people?
>>55026665
learn Java or C#
>>55026911
I'm probably going to get memed to death by /g/ with this, but: Java, C# and anything involving web development.
At least from the programming side of things. As for other ICT stuff, lets just say that I've never EVER heard about an unemployed system or network administrator (especially the latter), security expert or even DBAs.
>>55026911
fields that are growing like web and mobile and just general business side logic
>>55026974
>>55026949
>>55026933
These, and especially >>55026974
>>55027008
>>55026974
>>55026949
>>55026933
BUT THESE FIELDS ARE SATURIZED BY PAJEETS AND RASHIDS
REEEEEEEEEEEEEEEEEE
>>55027050
Protip: Getting a job in this industry isn't some easy-mode walk in the park, you actually have to make an effort.
But once you do make an effort, you'll see that it's certainly not impossible to get a job. You really shouldn't let yourself be influenced by /g/ on these things, it's full of NEETs and college students who never had to work for anything in their lives.
>>55020805
This is true if the bug actually fucks up something for end users. And you should be building test suites with some kinda of testing strategy.
>>55027220
that is if your employers believe in testing
>>55027391
Indeed, most of them don't (because they're hardware engineers that believe software is just something you hack together the night before production).
>>55027477
testing and debugging software is childs play compared to hardware
>>55027492
Thank you for proving my point. This is the exact kind of attitude you are likely to encounter in the industry from decision makers.
Who /Leipzig University/ here?
I know at least some of you goys are posting here because of the hardware lab
>>55020497
In my experience, sysadmin jobs for small-medium sized companies are comfier than most programming positions. Sure, there's no separation of duties and sometimes I end up having to do programming or help with web design or run cables or whatever, but at least it's comfy. Also there's a ping-pong table and we can drink at work :D
>>55027575
>and we can drink at work
>>55027527
What do you want me to say?
There are literally thousands of tools and methods to test your shitty software. a small off by one bug deep in your software is a lot easier to find by unit testing whereas a single miscalculated resistor will fuck the entire thing up. Good luck unit testing a single resistor.
Also, hardware simulators are useless if you are doing anything besides making an led blink
>>55027593
Family-owned companies ftw, dat earned nepotism
Last fall we flew our clients from all over the country and the bahamas to our office, had a taco truck in the parking lot with several kegs of beer and everyone got drunk and played sandbag toss/pingpong. We even designate certain employees to go out drinking with our Russian clients when they're in town.
I'm literally never leaving this job.
>>55026146
Yes. I too fell for the meme and now I'm scrambling to teach myself C# and Qt and Android and whatever the hell else shows up on all the generic "software engineer" job postings.
>>55020497
Seems like it. I work in embedded at Intel and everyone just jerks off all day.
>>55026879
>It is the most saturised programming industry and that completely crashes the salary levels.
But... but /g/ told me it would be a good choice because everyone else is going into web and phone apps...
>>55027877
I almost fell into that trap but after an internship doing firmware at a startup trying to do an smart meter with competition that was already on the market and sponsored by the government. I end up realizing its either that bullshit or from what I hear at other places, doing the same shit but with more people working on the same project that can fuck shit up.
I ended up switching my courses to focus on things that uses java and never looked back to low level shit again.
>>55027977
the only thing is, there is 50 web/mobile/java job to every embedded job and they pay more than the embedded bullshit.
Sup niggaz,
Have prof 2 + years experiance in soft/hw embedded design (industrail automation) and 2 years in fpga/asic dev/verification. Both jobs are fine, coleges are alcoholics but there is not big comunity, and the job is not well paid im my country ofc and the market is shit. Atm im soft dev in kernel/bios apps. RIP life
>>55028378
never too late to switch to doing higher level development
>>55022483
And since we both are gentlemen I would not think of ridiculing you for not considering the common possibility of FPGAs with softcore CPUs that requires, well, programming.
>>55020497
>comfier
Well, it seems to me that there is less danger of outsourcing to India in the embedded market. However embedded DSP is not a good job market these days so it depends heavily on the field within embedded systems.
>>55027610
>a lot easier to find by unit testing
I worked for a few years in embedded programming, chip design for embedded deployment and I have also worked in software testing and QA.
People *say* testing is easy. Experience shows it is not. I had one case where we fond the root cause of problems was that people were simply burned out. Management was very much not interested. End result: both of the important customers very angry, most programmers burned out, entire division "reorganized" by firing most and moving some, while management got promotion. This is a world where Dilbert lives.
>>55028773
I'm aware of soft-cores.
I'm just surprised people would bother with FPGAs when they're just going to go back to CPU-like logic structure.
Kinda like when people learn how to program a 3D environment and all the tools to make a video game...to make Gone Home.
>>55020805
>what is preliminary hardware testing
Holy shit, what a summer
>>55026403
>state of the software in the F35 is extremely bad
That is well known, GAO has made it public.
They even had a regression in the radar code in flight, practically a shooting offence.
L-M has a bad reputation for work. Staff turnover is simply enormous and every time you visit them half the people are new. F-35 will leave a permanent stain on your CV unless you evac. pretty quickly.
Embedded can be nice but some companies are slave ships. TI used to have a gruesome reputation, not sure how it is now.
>>55028889
There can be many reasons and I fear the horror known as "legacy code" is the most common reason for this. Then much of the rest of the FPGA is basically expensive interfacing. From a professional point of view it would be more satisfying to create more logic in the FPGA fabric rather than grunting out more banal code. However risk management often trumps the issue and softcore it is.
>>55024737
You can easily write a program in prolog that does all the backtracking for you.
>>55020497
Only if you are working on control software for android girlfriend appliance.
>>55026879
The F22 ran on Ada and never had as many problems as the F35 is having right now.
>>55028378
How often do you use formal methods vs. random testing in verification?
>>55029295
was long time ago (more then 2y ago). Idk rly, there was some testcase methodology based on constr random verif. For some specific cases were used directed env, you know homie
>>55023278
Well shit
>>55023278
Do people actually use VHDL outside universities anymore?
>>55030454
Yes
>>55030454
I see more places use verilog but vhdl is still used
>>55030943
>>55030692
maybe a better question is 'why does anyone use VHDL anymore?' then
>>55030982
VHDL is preferred for Altera systems
Verilog is preferred for Xilinx systems
The vendor tools tend to favor one or the other.
>>55030982
Strong typing and easier to get certified.
What's better: Computer Engineering or Computer Science?
>>55032785
software engineering
Probably, since they use C instead of Pajeet languages.
>>55020497
well they should all be gassed for writing garbage, nonfree, shitcode that literally has to be fixed in higher systems that actually interact with these shit devices running this shit code.
FUCK ALL HARDWARE FAGGOTS
>>55023278
wtf does line 23 do. the others => '0' bit.
>>55033930
Sets all 48 bits to 0. Say you wanted to set the 8th bit to 1 and the rest to 0, then you'd write:
signal counter : std_logic_vector(47 downto 0) := (7 => '1', others => '0');
How do I get an embedded job anyway? Everything I've seen is Java or boring web stuff with Rails or Angular and Node or whatever shit.
Where do I even LOOK? I love embedded stuff ;-;
>>55020799
it's a double q way toggle switch
>>55036165
2 way *
>>55023311
Vivado is pretty good
>>55035536
Sorry for the (You), but, I have the same question here.
I know that if I can't figure this stuff out on my own, there's probably no helping me. But this is just a casual curiosity, because I'm happy with what I'm doing now, and I'm curious about what the path towards embedded dev looks like.
>>55036477
>>55035536
go do an intership at a company that does embedded software then hopefully they hire you.
telecommunications, electronics or software engineering?
>>55036773
software engineering
>>55026653
but anon how will i ever realize my dream of making robots true
it's like you're saying webdev is better
IMO I think that embedded is trending in power electronics field.
>>55033670
>they use C instead of Pajeet languages
top kek. C is the most pajeet language one can imagine. Just go on youtube and search for "C programming tutorial".
C programmers are getting cucked by dirty sandniggers from india.
>tfw aspiring for a job in embedded
>tfw you've never done anything above programming AVRs
>>55020891
No such thing as 100% test coverage
>>55020805
This implies a whole team failed. Everyone makes mistakes, and things like regression testing and code reviews are there to catch them.
>>55020891
>he doesn't formally prove his code correct
>>55023987
Vivado is alot better than alteras solution
>>55026653
>Cont Google errors
This, most web devs would shit themselves in these positions. Basically u have very small communities supporting each hardware. Also MCUs have errata that is only fixed with new release of hardware
>>55020772
>the cash is alright, and if you become an expert you're then you're set.
So, pretty much any real job ever.
bump for advice how to get into this field as an entry level eng.
