As a developer I can freely admit that without the operations people the software I develop would not run anywhere but on my laptop.
I know as much about hardware as a cook knows about his stove and the plates the food is served on – more than the average person but waaaay less than the people producing and maintaining them.
As a devops manager that’s been both, it depends on the group. Ideally a devops group has a few former devs and a few former systems guys.
Honestly, the best devops teams have at least one guy that’s a liaison with IT who is primarily a systems guy but reports to both systems and devops. Why?
It gets you priority IT tickets and access while systems trusts him to do it right. He’s like the crux of every good devops team. He’s an IT hire paid for by the devops team budget as an offering in exchange for priority tickets.
I've always found this weird. I think to be a good software developer it helps to know what's happening under the hood when you take an action. It certainly helps when you want to optimize memory access for speed etc.
I genuinely do know both sides of the coin. But I do know that the majority of my fellow developers at work most certainly have no clue about how computers work under the hood, or networking for example.
I find it weird because, to be good at software development (and I don't mean, following what the computer science methodology tells you, I mean having an idea of the best way to translate an idea into a logical solution that can be applied in any programming language, and most importantly how to optimize your solution, for example in terms of memory access etc) requires an understanding of the underlying systems. That if you write software that is sending or receiving network packets it certainly helps to understand how that works, at least to consider the best protocols to use.
. I think to be a good software developer it helps to know what's happening under the hood when you take an action.
There's so many layers of abstractions that it becomes impossible to know everything.
Years ago, I dedicated a lot of time understanding how bytes travel from a server into your router into your computer. Very low-level mastery.
That education is now trivia, because cloud servers, cloudflare, region points, edge-servers, company firewalls... All other barriers that add more and more layers of complexity that I don't have direct access to but can affect the applications I build. And it continues to grow.
Add this to the pile of updates to computer languages, new design patterns to learn, operating system and environment updates...
This is why engineers live alone on a farm after they burn out.
It's not feasible to understand everything under the hood anymore. What's under the hood grows faster than you can pick it up.
yeah i wish it was a requirement that you're nerdy enough to build your own computer or at least be able to install an OS before joining SWE industry. the non-nerds are too political and can't figure out basic shit.
I think software was a lot easier to visualise in the past when we had fewer resources.
Stuff like memory becomes almost meaningless when you never really have to worry about it. 64,000 bytes was an amount that made sense to people. You could imagine chunks of it. 64 billion bytes is a nonsense number that people can't even imagine.
and I don't mean, following what the computer science methodology tells you, I mean having an idea of the best way to translate an idea into a logical solution that can be applied in any programming language,
It very much depends on how close to hardware they are.
If someone is working with C# or JavaScript, they are about as knowledgeable with hardware as someone working in Excel(I know this statement is tantamount to treason but as far as hardware is concerned it’s true
But if you are working with C or rust or god forbid drivers. You probably know more than the average IT professional you might even have helped correct hardware issues.
The devops team set up a pretty effective setup for our devops pipeline that allows us to scale infinity. Which would be great if we had infinite resources.
We're hitting situations where the solution is to throw more hardware at it.
And IT cannot provision tech fast enough within budget for any of this. So devs are absolutely suffering right now.
I know this is not everyone and there're some unicorns out there but after working with hiring managers for decades i can't help but see cheap programmers when I see Devops. It's ether Ops people that think they are programmers or programmers that are not good enough to get hired as Software Engineers outright at higher pay. It's like when one person is both they can do both but not great at ether one. Devops works best when it's a team of both dev and Ops working together. IMO
Our IT team would rather sit in a room with developers and solve those problems, than deal with hundreds of non-techs who struggle to add a chrome extension or make their printer icon show up.
As a developer I like to mess with everything. Currently we are doing an infrastructure migration and I had to do a lot of non-development stuff to make it happen.
Honesly I find it really usefull (but not necessary) to have some understanding of the underying processes of the code I'm working with.
Simple example, our NASes are EMC2. The devs over at the company that does the software say they're garbage, we should change them.
Mind you, these things have been running for 10 years straight 24/7, under load most of the time, and we've only swapped like 2 drives, total... but no, they're garbage 🤦....
I work on a team with mainly infrastructure and operations. As one of the only people writing code on the team. I have to appreciate what IT support does to keep everything moving. I don't know why so many programmers have to get a chip on their shoulder.
Sorry, this comment is causing me mental whiplash so I am either ignorant, am subject to non-standard circumstances, or both.
My personal experience is that developers (the decent ones at least) know hardware better than IT people. But maybe we mean different things by “hardware”?
You see, I work as a game dev so a good chunk of the technical part of my job is thinking about things like memory layout, cache locality, memory access patterns, branch predictor behavior, cache lines, false sharing, and so on and so forth. I know very little about hardware, and yet all of the above are things I need to keep in mind and consider and know to at least some usable extent to do my job.
While IT are mostly concerned on how to keep the idiots from shooting the company in the foot, by having to roll out software that allows them to diagnose, reset, install or uninstall things on, etc, to entire fleets of computers at once. It also just so happens that this software is often buggy and uses 99% of your cpu taking it for spin loops (they had to roll that back of course) or the antivirus rules don’t apply on your system for whatever reason causing the antivirus to scan all the object files generated by the compiler even if they are generated in a whitelisted directory, causing a rebuild to take an hour rather than 10 minutes.
They are also the ones that force me to change my (already unique and internal) password every few months for “security”.
So yeah, when you say that developers often have no idea how the hardware works, the chief questions that come to mind are
What kinda dev doesn’t know how hardware works to at least an usable extent?
What kinda hardware are we talking about?
What kinda hardware would an IT person need to know about? Network gear?
When IT folks say devs don't know about hardware, they're usually talking about the forest-level overview in my experience. Stuff like how the software being developed integrates into an existing environment and how to optimize code to fit within the bounds of reality--it may be practical to dump a database directly into memory when it's a 500 MB testing dataset on your local workstation, but it's insane to do that with a 500+ GB database in production environment. Similarly, a program may run fine when it's using a NVMe SSD, but lots of environments even today still depend on arrays of traditional electromechanical hard drives because they offer the most capacity per dollar, and aren't as prone to suddenly tombstoning when it dies like flash media. Suddenly, once the program is in production, it turns out that same program's making a bunch of random I/O calls that could be optimized into a more sequential request or batched together into a single transaction, and now it runs like dogshit and drags down every other VM, container, or service sharing that array with it. That's not accounting for the real dumb shit I've read about, like "dev hard coded their local IP address and it breaks in production because of NAT" or "program crashes because it doesn't account for network latency."
Game dev is unique because you're explicitly targeting a single known platform (for consoles) or targeting for an extremely wide range of performance specs (for PC), and hitting an acceptable level of performance pre-release is (somewhat) mandatory, so this kind of mindfulness is drilled into devs much more heavily than business software dev is, especially in-house dev. Business development is almost entirely focused on "does it run without failing catastrophically" and almost everything else--performance, security, cleanliness, resource optimization--is given bare lip service at best.
Game development is a very specific use case, and NOT what most people think of when talking about devs vs ops.
I'm talking enterprise software and SaaS companies, which would be a MUCH larger part of the tech industry then games.
There are a large number of devs who think public cloud as infrastructure is ALWAYS the right choice for cost and availability for example... Which in my experience is actually backwards, because legacy software and bad developers fail to understand the limitations of this platforms, that it's untrustworthy by design, and outages insue.
In these scenarios understanding how the code interacts with actual hardware (network, server and storage or their IaaS counterparts) is like black magic to most devs... They don't get why their designs are going to fall over and sink into the swamp because of their nievete. It works fine on their laptop, but when you deploy to prod and let customer traffic in it becomes a smoking hole.
Devops is also usually more dev than ops, and it shows in the availability numbers.
One of my managers supported the idea that it's better to come to devOps from a Dev perspective than an Ops perspective; just because the coding is cleaner and easier to manage with that strong coding background. Also, it's how devops started: we were doing rudimentary devops - or what Ansible is capable of - back in 2002, long before they even fabricated the term 'devops'.
I know we're just having fun, but in the future consider adding the word "some" to statements about groups. It's just one word, but it adds a lot of nuance and doesn't make the joke less funny.
That 90's brand of humor of "X group does Y" has led many in my generation to think in absolutes and to get polarized as a result. I'd really appreciate your help to work against that for future generations.
In my experience a lot of IT people are unaware of anything outside of their bubble. It's a big problem in a lot of technical industries with people who went to school to learn a trade.
The biggest problem with the bubble that IT insulates themselves into is that if you don't users will never submit tickets and will keep coming to you personally, then get mad when their high priority concern sits for a few days because you were out of office but the rest of the team got no tickets because the user decided they were better than a ticket.
If people only know how to summon you through the ancient ritual of ticket opening with sufficient information they'll follow that ritual religiously to summon you when needed. If they know "oh just hit up Rob on teams, he'll get you sorted" the mysticism is ruined and order is lost
Honestly I say this all partially jokingly. We do try to insulate ourselves because we know some users will try to bypass tickets if given any opportunity to do so, but there is very much value in balancing that need with accessability and visibility. So the safe option is to hide in your basement office and avoid mingling, but thats also the option that limits your ability to improve yourself and your organization
No, we can't get gigabit fiber everywhere. No, I don't care if your program needs it. Yes, the laws of physics are laws for a reason. Write more robust code.
Ouch yeah that windows endpoint stuff is really rattling though. I get you just can't whitelist some folder without compromising security, but when the "eNdPoInt pRoTeCtIon" just removes dlls and exes you are compiling (and makes your PC crawl) you really hate that shit.
Right click? 40 seconds plz (maybe any of the possible contextual right clicks might be on a virus so lets just check them all once again).
At home I have an old linux pc, and it blows those corpo super pcs out the window.
Rant off :-D
Ah yeah, IT people are chill, always be cool with them is also a good idea, not their fault all this crap exists.
Hahaha! We've an "architect" who insists he needs to be the owner on the gitlab. My colleague has been telling him to fuck off for the entire week. It reached the point that fool actually complained to our common boss...
The guy is so used to working as a start-up and has no fucking clue about proper procedures. It's terrifying that he could be in charge of anything, really.
In a rapidly churning startup phase, where new releases can and do come out constantly to meet production requirements, this one size fits all mentality is impractical.
If you refuse to whitelist the deployment directory, you will be taking 2am calls to whitelist the emergency releases.
No it can't wait until Monday at 9am, no there will not be a staged roll out and multiple rounds of testing.
I am more than willing to have a chat; you, me and the CEO.
No it can't wait until Monday at 9am, no there will not be a staged roll out and multiple rounds of testing.
I hope you're doing internal product development. Otherwise, name and shame so I can stay the hell away from your product. This is a post-Crowdstrike world.
In my experience it’s been IT people telling me you can’t use a certain tool or have more control over your computer cause of their rules.
The expression is appropriate but the meme assumes that im doubting the IT person’s expertise. I’m not, I’m just not liking the rules that get in the way of my work. Some rules do make sense though.
Edit: just wanted to point out, yes I agree, you need the rules, they are still annoying tho.
Their rules have stopped me from being able to do my job. Like the time the AV software quarantined executables as I was creating them so I literally could not run my code. When security enforcement prevents me from working, something needs to change.
As an IT guy, I'd love to give software devs full admin rights to their computer to troubleshoot and install anything as they see fit, it would save me a lot of time out of my day. But I can't trust everyone in the organization not to click suspicious links or open obvious phishing emails that invite ransomware into the organization that can sink a company overnight.
Fair points but as someone who works in cybersecurity. Phishing emails can happen without admin access. I haven’t heard of any randsomware that is triggered by just clicking on a link.
I think there should be some restrictions but highly technical people should slowly be given more and more control as they gain more trust/experience.
And the more corporate the organisation the more rules, at least the places I have worked trusts developers enough to give local admin, that takes the edge off many tasks.
I think you probably don't realise you hate standards and certifications. No IT person wants yet another system generating more calls and complexity. but here is iso, or a cyber insurance policy, or NIST, or acsc asking minimums with checklists and a cyber review answering them with controls.
Crazy that there's so little understanding about why it's there, that you just think it's the "IT guy" wanting those.
So you don't trust me, but you trust McAfee to give it full control over the system. Yet my software doesn't work because something is blocked and nothing is showing up in the logs. But when we take off Mafee, it works. So clearly McAfee is not logging everything. And you trust Mcafee but not me? /s kinda.
I worked in software certification under Common Criteria, and while I do know that it creates a lot of work, there were cases where security has been improved measurably - in the hardware department, it even happened that a developer / manufacturer had a breach that affected almost the whole company really badly (design files etc stolen by a probably state sponsored attacker), but not the CC certified part because the attackers used a vector of attack that was caught there and rectified.
It seemingly was not fixed everywhere for whatever reason... but it's not that CC certification is just some academic exercise that gives you nothing but a lot of work.
Is it the right approach for every product? Probably not because of the huge overhead power certified version. But for important pillars of a security model, it makes sense in my opinion.
Though it needs to be said that the scheme under which I certified is very thorough and strict, so YMMV.
I think the meme is more about perspectives and listening to the way someone thinks about operating IT is very different from the way someone things about architecting IT
I think it's on a case by case basis but having help desk ppl help you out and opening powershell and noodling without any concept of problem solving made me make this face once.
It probably goes both ways, I'm a dev and I assembled computers at 12 yo so I believe I have a lot of experience and knowledge when it comes to hardware. I've also written code for embedded platforms.
IT people in my pov can really come across as enthusiast consumers when it comes to their hardware knowledge.
"did you guys hear Nvidia has the new [marketing term] wow!" . Have you ever thought about what [marketing term] actually does past just reading the marketing announcement?
At the same time I swear to God devs who use macs have no idea how computers work at all and I mean EXCLUDING their skill as a dev. I've had them screen share to see what I imagine is a baby's first day on a computer.
Interesting comment on the Mac. At my workplace we can choose between Mac or Windows (no Linux option unfortunately, my personal computer runs Debian). Pretty much all the principle and senior devs go for Mac, install vim, and live in the command line, and I do the same. All the windows people seem over reliant on VSCode, AI apps, and a bunch of other apps Unix people just have cli aliases for and vim shortcuts. I had to get a loaner laptop from work for a week and it was windows. Tried using powershell and installing some other CLI tools and after the first day just shut the laptop and didn’t work until I got back from travel and started using my Mac again.
If you don't have access to Linux, MacOS is the closest commercially available option so it makes sense.
Also please take what I said lightly, I by no means want to bash Mac users and generalize them. It just has been my experience. I'm sure there are thousands of highly competent technical users who prefer Mac.
I have to use windows for work. Installed vim through winget and set a powershell alias, allowing me to use it similarly to linux.
Windows ist still just ass though.
Lmao, devs who insist on using VIM and the terminal over better graphical alternatives just to seem hardcore are the worst devs who write the worst code.
"Let me name all my variables with a single letter and abbreviations cause I can't be bothered to learn how to setup a professional dev environment with intellisense and autocomplete."
Agreed. I have colleagues that I write scripts for (I don't do that any more, I stopped and shit stopped working, so they solve things manually now), they don't know shit about scripting... and still don't.
On the other hand, I've had the pleasure of working with a dev that was just this very positive, very friendly person and was also very knowledgeable when it came to hardware, so we were on the same page most of the time. He also accepted most of my code changes and the ones that he didn't, gave him an idea of how to solve it more efficiently. We were a great team to be honest. We're still friends. Don't see him as frequently, but we keep in touch.
They do exist and some of them swear Mac has better workflows (than windows because most of the time your options are Windows or Mac). I would call them loonies but I've seen some smart people use Macs.
Meh it's usually for shitty companies that expect their devs to write real software, ssh into things, access databases, but put the same hurdles in front of them as joeblow from sales who can't use an ipad to buy a sandwich without clicking a phishing link. So every new project is slowed down cause it takes weeks of emails and teams conversations to get a damn db sandbox and it's annoying.
On the other hand IT doesn't know you and has millions of issues to attend to
IT guy here. If we give one user special rights, that login will get passed around like a blunt at a festival to "save time".
Users are dumb and lazy, and that includes devs.
I took it as software engineers tend to build for scalability. And yep, IT often isn't prepared for that or sees it as wasted resources.
Which isn't a bad thing. IT isnt seeing the demands the manager/customer wants.
I'm glad you've done both because yeah, it's a seesaw.
If IT provisions just enough hardware, we'll hit bottlenecks and crashes when there's a surprise influx of customers. If software teams don't build for scale, same scenario, but worse.
From the engineer perspective, it's always better to scale with physical hardware. Where IT is screaming, "We dont have the funds!"
I spent a weekend helping my buddy who graduated magna cum laude with an Electrical and Computer Engineering degree build a PC. Given a breadboard and some schematics, he could probably have created working prototypes of half of the components, but figuring out where to put the screw risers under the motherboard? Forget about it.
I'm almost done with my CS degree, I started learning programming at age 10, low-level software development like drivers and embedded really interests me and that's the direction I want to go in for a career, but I had to ask my friend who was studying with me to help me build my PC. Hardware just scares me. I'm worried ill bork something :3
That's how I look at 90% of the shit "systems" I'm forced to interact with (xiaomi's MIUI, banking apps, govt apps, apps that should've been fucking websites, websites that "gently nudge" you to use the app, electron apps that are windows only)
MiUI is not that bad IMO. The ad services and the integrated apps are horrible (even without the ads), but apart from that, the UI is fairly usable. They really haven't changed that much from what Android comes with by default.
I definitely have moments like this too. I have been reflecting more lately and trying to decide if the feeling is temporary or permanent. I have been pondering what else I would do. Are you considering a career change, and if so, what would you do instead? I don't know if I could transition to something else without going back to school, and it would kill me a bit inside to take out more student loans.
Infrastructure maintenance is management, security and day to day business, while software engineering is mostly concerned with itself. They use distinct tools and generally have nothing to do with each other (except maybe integration).
We need new terms, IT means "works with computers, but more than Word and Excel" for too many people. In Switzerland they split the apprenticeship names to 'platform engineer' and 'application engineer', which I think is fitting.
IT is an administrative function and is really part of operations.
Software development is generally a creative position and is a profit center. If you work somewhere where you develop internal apps, you may have a different perspective.
My current workplace organizes both development and infrastructure within IT which itself is a sub department of finance. I'm not saying this is the best approach because honestly it only took 1.5 layers of apathetic management to make long term planning a nonstarter
That's the face I've made just yesterday when my friend told me she's now eligible for a subsidized IT mortgage. That thing was one of Russia's last ditch attempts at stopping skilled workers from fucking off to different countries. The problem is, she's a web designer. I guess that counts as IT nowadays, so good for her. But it's bitter to hear as sr. backend tech who never hit the criteria...
That's pretty much how the Russian economy works right now, in a nut shell. To stop emigration caused by the expensive war, they're giving away a ton of expensive handouts.
The interest rate is at 19% and counting. Very cool, very sustainable. I have a feeling "the last laugh" will be yours, OP, even if they win in Ukraine.
Yep, I know the face... made it a few times with colleagues that don't know basic Windows scripting, but somehow got bonuses... but I don't kiss upper management ass, so I never do. That's life I guess...
When trying to request a firewall change IT told me "ports between 1 and 1024 are reserved and can't be used for anything else" so I couldn't be using it for a pure TCP connection, and besides, there would have to be a protocol on top of TCP, just TCP as protocol is obviously wrong. I was using port 20 because it was already open...
as a full stack dev, everything you said has offended me.
port 20 is used for FTP, unless you were using FTP, then go right ahead. Guessing that since you didn't know the protocol you were not using FTP.
port usage reservations are incredibly important to ensure that the system is running within spec and secure. imagine each interface like a party telephone line and the ports are time slots.
your neighborhood has reserved specific times (ports) for everyone to call their relatives. if you use the phone not in your slot (port) your neighbors might get pissed off enough to interrupt your slot. and then it's just chaos from there.
As IT/network/security, using a well known port for something that's not what is supposed to run on that port, is inviting all kinds of problems.
Especially the very well known ones, like ftp, ssh, SMTP, http, HTTPS, etc (to name a few). People make it their mission to find and exploit open FTP systems. I opened up FTP on a system once to the internet as kind of a honeypot, and within a week or so, there was someone uploading data to it.
No bueno. Don't use well known ports for things unless the thing that well known port is known for, is what you want to do.
This is the same between many different software development disciplines, fpga devs (or hardware devs for that matter) vs. driver devs, driver devs vs. backend dev, backend devs vs frontend devs, integrators vs everyone.
I'm both, and while I do hate myself, I don't think it's related, so I'm not sure I get it.
(I hate computers more, though, except when they're turned off — no bugs when they're off —, but they're the only thing I'm good enough at to make a living off of.)
It goes both ways. At my old job, they took away local admin. But for some reason they configured visual studio to run as admin. So, I just wrote a little program that opens the shell. Whenever I needed admin, I just ran that program from Visual Studio.
Fair enough. Local admin is generally not something that I would want to restrict from people, especially those that are, or at least, should be, more knowledge than most.
I'll fight for that right for people most of the time.
Some users I would say should not have it, but generally developers are not those people. You know the ones.