IT/Software career thread: Invert binary trees for dollars.

OheoHana said:

And then I realized.. if doing a web-based app, lets say I upload to my app a .CSV of client information that I parse and want to save, I would need some type of backend framework, like Firebase? Man, this is going to be a fun journey trying to sort all this out.

Click to expand...

Right, you'll have some kind of backend API running on a server. You could go the cloud route with Azure or AWS. In fact, that might be a worthy option since you'll learn how the cloud systems work. Most everything is moving in that direction now so definitely shouldn't be ignored. And they aren't really that difficult to understand. They make a lot of stuff easier, actually.

Mist said:

This is just edgelording. No company can operate on 50% of people being incapable of filling even their basic role.

Click to expand...

Phazael said:

Oh yeah they can. I mean the government obviously can and anyone who has ever been to a fast food restaurant, let alone worked in one, knows its possible. If anything, its easier to be useless and get away with it in white collar work.

Click to expand...

Yeaaaah, I was going to say exactly this. I work for a Native American Tribal Government and I can say with absolute certainty those numbers here are something like -

5% of people who wear 6 hats and know how everything works. Half of these people are so critical that their loss sets departments back for years.
20% of people who can do their one job passably well but are completely ignorant of how anything outside of their day-to-day tasks actually works.
75% of people who are just totally fucking useless, it doesn't make any sense how they got hired, let alone how they weren't fired 10 weeks in when it was obvious that they aren't even capable of filling a chair.

I'm not even joking, its fucking insane. We have 400ish people employees and yeah, 5% is around 20 people... and yes, I would probably struggle to list 20 people spread across all departments that I would say really know their job. 80 people who are just ok... sure, that's probably fair for the 20% who keep stuff moving forward to some degree.

75%, 300 people - peoples grandparents, brothers, cousins etc. who are 99% useless. We just hired a new payroll person (There is only 1 person in that department) and I've seen this girl do 6 jobs in 6 other departments over the years, so when I went over to set up her accounts I was like "I didn't know you did payroll stuff, congrats on the job" and she responded with "I don't, I'm super nervous, I'm not really good with numbers, I hope you can train me!" lol. And yeah, our IT Department flat out does many other peoples jobs a ton of the time. It's pretty nuts, and whenever I interview people for a job here I always let them know "You can't walk into this like its any normal job. Think of it more like if your family owned a company and you were their IT guy. If the Chairman calls you and says his mom needs help moving firewood, thats your job now. If a Council member calls you and tells you they just bought 3 laptops at a pawn shop and they want you to set them up for their kids, thats your job now. We deal with shit every single day that would be absurd at any other job. I've had a Director call me at 11pm on a Saturday night because their Netflix kept buffering and they wanted to know if I could fix that. If this sounds like something you would balk at, you probably are not going to fit in here."

It has its pros and cons for sure, on the one hand - typically nothing is important to anyone, so if there is a project that needs to be done.. have a week, have a month.. no one cares. On the other hand, you need something asap from someone? It might take a week, a month, more.

But yeah, can a company work with primarily completely useless employees? Yes, yes it can. (And I get that the Gov isn't exactly a company... but that being said, this Tribe owns its own company that is run the exact same way and its still generating millions per year. It's stupid)

OheoHana said:

Thank you kindly.

Can I ask how React / Svelte fit in to all this, especially with the CSS Frameworks like Bootstrap? And Node.js

Click to expand...

React, Svelte, Angular, and VueJs are all client-side JavaScript frameworks. That's the stuff that runs directly in your browser. Bootstrap (or any other CSS framework) simply supplies you with CSS classes that you can apply to buttons, inputs, etc. to get that slick look. With the client-side framework of your choice you can create components that wrap those styles. For example, you could create a 'text input' component that includes the Bootstrap styling and any other unique functionality you desire, ensuring that all those inputs everywhere share the same behavior, look and feel.

Node.js (or alternatively .NET Core or Java) are the server-side frameworks that communicate with your client side. That's the stuff that queries a database, sends data to the browser, or receives data from the browser to store in a database, etc.

With the Cloud you can go 'serverless' and set up simple function calls that communicate with your client-side. I only have a basic understanding of them, but the point is that you don't have to spin up an entire virtual machine running Windows or Linux. They're just bits of code that handle data received/sent from the browser.

If you want to learn this stuff, I highly HIGHLY recommend https://www.udemy.com/user/academind/

Max has courses of all the popular Javascript frameworks and takes you as a complete novice to having a pretty damn good understanding of the subject. I think the courses are on sale now -- you can usually get them for under 20 bucks.

As for which client-side framework to go with? I'd probably say Vue.js is the easiest to work with. Svelte is really similar but has some limitations that I don't think you should have to deal with.

I am getting interviews for telecomm stuff by the boatload now. My family ended up not selling, so I no longer need 100% remote which means all doors are open. Time to move on and maybe move up.

Neranja said:

Contrarian opinions:

• JavaScript is a horrible language and ecosystem that we have put up with for far too long. The sheer amount of frameworks to make JavaScript even tolerable is mind-boggling. Let's not even talk about complete languages and frameworks that compile into JavaScript, so you don't have to touch it in the first place. I think there are over 50 by now. Hell, even languages like Kotlin that originally targeted the Java VM now have a JavaScript backend.
  
  
• Node.js on the server should also die in a fire. The whole ecosystem is fucked beyond belief, because developers just include whatever they find via npm. Which then may or may not include malicious code, because someone either put it up, or took over a package. Sometimes even one of the developers has a meltie and pulls his package, making a whole house of cards just implode overnight.
  This already happened with a package that ... *drumroll* ... centered a string on a line. I kid you not.
  
  
• Contrary to popular opinion, those Cloud service providers are not your friend. Sure, it's convenient, and there is always the promise of "you can pay a little now and then scale up if you hit it big."
  Trust me, once you "hit it big" (either by being successful, or by trending on Twitter), if you don't have your ducks in a row when you are small (read: architecture), then you won't suddenly scale out. Sure, you can fire up loads of instances to serve HTTP and load-balance them, but at the end of the line the data has to be stored somewhere, and that is in almost all cases some form of database. And you can't scale those out easily without very specific knowledge.
  Here's some homework: Look up the costs for small, medium and big instances on Amazon, Google and Microsoft. Then look up the prices for their dedicated database instances. Also look up how many transactions per second they guarantee.
  
  
• Trusting someone else with your data is always a risk-benefit calculation. This is especially true if you are handling sensitive data, e.g. legal, medical or military.
  
  
• Also, nothing says that those cloud service providers can't suddenly increase pricing once they have you by the balls. The moment you have started to use all those "conveniences" of a specific cloud provider, you are basically vendor locked-in, or else you will have to re-tool your backend.

Click to expand...

My company is doubling down on dedicated/metal/colo hosting and leasing our DC space back out. Been our biggest growth area for 2 years now as people recoil from the cloud space over costs and go back to hosting their own stacks.

Cloud is super cool and useful for a lot of things, don't get me wrong, but we're already seeing a relatively large sprint back to in-house or datacenter space leases after everyone pushed too fast in to AWS/gCloud/Azure etc.

OheoHana said:

I can't tell if the "state" stuff is hacky, or if it was just the video.

Click to expand...

All the frameworks have the concept of "state".

ronne said:

My company is doubling down on dedicated/metal/colo hosting and leasing our DC space back out. Been our biggest growth area for 2 years now as people recoil from the cloud space over costs and go back to hosting their own stacks.

Cloud is super cool and useful for a lot of things, don't get me wrong, but we're already seeing a relatively large sprint back to in-house or datacenter space leases after everyone pushed too fast in to AWS/gCloud/Azure etc.

Click to expand...

Interesting, I'm still seeing the reverse where I work, though our own data centers are really badly managed.

When I started my current job, the entire server stack I manage ran very poorly, constant performance issues, apps crashing for seemingly no reason. My predecessors had tried to solve it by requesting more servers, so they had like 40 RHEL servers running on VMWare doing the work of less than 10. Everything ran like they had linux running on old 486's or something, it was really bizarre and nobody knew why. I suggested that the Vmware cluster might be overloaded, and they responded which graphs showing nope, they're actually quite low in utilization.

Eventually I got access to the VMWare console and I learned that while they had assigned the requested vCPUs to our RHEL servers, they had severe restrictions in place so it was basically only giving us a single core on each VM. Dum dums.

Later I ran into the same problem with RAM, they assigned the right amount but VMware lets you restrict how much a VM can use and they had no idea how to configure that properly so eventually apps would die for seemingly no reason. "servers don't usually require this much RAM so we never set it any higher". More dum dums. They took away our ability to see how our own VMs are configured shortly after this lol.

That's just the tip of the spear. People lie constantly, probably unintentionally but they'll claim something is configured in X way when it's really Y. In AWS, there's never any guess work, but when you have so little visibility into your own devices or the network it can be a little maddening.