After our Head Of Software has terminated.

I started to take control over our development crew. And in this year I did more then the old head in the last 6 years.
- Swapped from plain old SVN to Gitlab.
- Build a complete autonomous deployment with Gitlab.
- Introduced code reviews.
- Started to refactor the legacy product with 500.000 lines of code...
- learned how to use confluent apache kafka and kubernetes to split the legacy project in many small and maintainable one.(not done yet)
- Last 3 weeks I learned how to use elasticstack with kibana and co. That we aren't blind anymore. Big dashboards are now shown in the middle of the room :) and maybe convinced my coworker that we use unity3d for our business application cause of support for all devices and same design on them. And offline capabilities. (Don't know if this was my best idea)
When I look back, I'm proud to did that much in one year alone. And my coworkers are happy too that they have less work with deployments and everything.
But I can't decide what's the title for this. System or Software Architect cause I litterallity did both :/

I finally ended my first side project ever. I challenged myself to write a Tetris game in vanilla JS (with the less possible lines of code), some algorithm was tougher than I expected (2d array hell) but I made it ! 500 lines of JS code, I feel I could refactor some stuffs now...

It definitely helps to not be proud, especially to talk to people.

I was about to refactor something someone did, without understanding their intentions and because I thought it was incorrect.

Thank God I got off my high horse to talk to them first!

If you are shy or bad at approaching people (like me), just force yourself and try.. You will eventually learn to be less hesitant.

"I'm almost done, I'll just need to add tests!"

Booom! You did it, that was a nuke going off in my head.

No, you shouldn't just need to add tests. The tests should have been written from the get go! You most likely won't cover all the cases. You won't know if adding the tests will break your feature, as you had none, as you refactor your untested mess in order to make your code testable.

When reading your mess of a test case and the painful mocking process you went through, I silently cry out into the void: "Why oh why!? All of this suffering could have been avoided!"

Since most of the time, your mocking pain boils down to not understanding what your "unit" in your "unit test" should be.

So let it be said:

- If you want to build a parser for an XML file, then just write a function / class whose *only* purpose is: parse the XML file, return a value object. That's it. Nothing more, nothing less.
- If you want to build a parser for an XML file, it MUST NOT: download a zip, extract that zip, merge all those files to one big file, parse that big file, talk to some other random APIs as a side-effect, and then return a value object.

Because then you suddenly have to mock away a http service and deal with zip files in your test cases.

The http util of your programming language will most likely work. Your unzip library will most likely work. So just assume it working. There are valid use cases where you want to make sure you acutally send a request and get a response, yet I am talking unit test here only.

In the scope of a class, keep the public methods to a reasonable minimum. As for each public method you shall at least create one test case. If you ever have the feeling "I want to test that private method" replace that statement in your head with: "I should extract that functionality to a new class where that method public. I then can create a unit test case a for that." That new service then becomes a dependency in your current service. Problem solved.

Also, mocking away dependencies should a simple process. If your mocking process fills half the screen, your test setup is overly complicated and your class is doing too much.

That's why I currently dig functional programming so much. When you build pure functions without side effects, unit tests are easy to write. Yet you can apply pure functions to OOP as well (to a degree). Embrace immutability.

Sidenote:

It's really not helpful that a lot of developers don't understand the difference between unit, functional acceptance, integration testing. Then they wonder why they can't test something easily, write overly complex test cases, until someone points out to them: No, in the scope of unit tests, we don't need to test our persistance layer. We just assume that it works. We should only test our businsess logic. You know: "Assuming that I get that response from the database, I expect that to happen." You don't need a test db, make a real query against that, in order to test that. (That still is a valid thing to do. Yet not in the scope of unit tests.)

I once made a .less files of 400 lines and today I redid it completely,

The file is now only 40 lines long and readable again

My co-worker had to add some small feature, and while at it, he thought it’s a good idea to “refactor” the entire repo.

Now the code is over-modularized and full of disconnected 100 line files, just for the sake of modularity.

Sometimes less is more

Spent the whole day trying to refactor a program that consumes an API and didn’t work out. Now we have one less day to try to improve the product.

I think the fault was that we were aiming too high by looking at other libraries and we ended up over-complicating it.

What's some valuable things you've learned this year?

I'll start.
Always ask myself, if I can elaborate on any given statement, otherwise just keep shut.

Always ask yourself before writing any code, if it's gonna be easy to modify later on, if not, take a step back and evaluate the bigger picture.

Don't bother with writing things clean in the first go.
It'll be much easier to refactor later on and take less total time.

ColdFusion and all ColdFusion devs should be executed. Its a god-awful software from the 90's and if you still use it you're either braindead or ignorant.

Shut up about legacy CF code too! No one cares whether or not your embeddable calendar would be hard to make in JS; fucking figure it out.

I realise that CF may make things easier in the short run, but in the long run you'll have introduced so much technical debt that you'll run crying back to JS anyways; CF is so hard to refactor and even to make flexible that you would spend less total time over an application lifecycle learning JS.

So I've been hired as a senior software developer with all the tags included (mentoring, innovating, pushing forward changes) for a company that is trying to move away from waterfall development (yup, it's 2019 and this exists) to a more iterative workflow.
I was initially hired and sent out to do some "field work" abroad for 3 months and then worked "remotely" from the local office with our field partners.
During all this time it seemed that my ideas go through smoothly, there was a lot of chatter about how things are moving forward, how new projects, innovations and new methodologies are implemented.
And yet, after my "remote" work has finished and I have to do things locally more, all of the skeletons fell out. It's just talk, nothing seems to be changing at all and yet any attempts to talk with the brass is like hitting a brick wall.
Not only that, I've been handed a 12 year old project with no possibility to refactor, no technical documentation, very few comments and in a terrible style.
The atmosphere in the company is odd as hell. People are either not very initiative, nor they seem to really care about all of the "changes" that "should be happening".
It almost feels that I've arrived in a company that still lives in 2007 more or less.
Should I quit, or perhaps it's a little "too soon" (have spent 7 months in the place already)? What I don't want is to get in the same train again (work for a company for 8 - 12 months, feel burned out because of the divergence between actual things done and "plans" and then change the job).

Took a SWAG at stand-up this morning, saying I thought I could resolve a circular dependency introduced in a junior's branch by noon. As the morning dragged on, I became less and less certain of it, but lo! and behold I managed to refactor my way out of hell with 2 minutes to spare!

This calls for lunch beer! Which management has no way of stopping me from doing on a regular day due to remote working, but at least today I've convinced myself I earned it.

So i wasted last 24 hours trying to satisfy my ego over a shitty interview and revisiting my old job's codebase and realising that i still don't like that shit. just i am 25 and have no clue where am i heading at. i am just restless, my most of the decisions in 2023 have given very bad outcomes and i am just trying doing things to feel hopeful.

context for the interview story-----
my previous job was at a b2b marketing company whose sdk was used by various startups to send notifications to their users, track analytics etc. i understood most of it and don't find it to be any major engineering marvel, but that interviewer was very interested in asking me to design a system around it.

in my 1.2 years of job there, i found the codebase to be extremely and unnecessarily verbose ( java 7) with questionable fallbacks and resistance towards change from the managers. they were always like "we can't change it otherwise a lot of our client won't use our sdk". i still wrote a lot of testcases and tried to understand the working of major features.

BTW, before you guys go on a declare me an embarrassment of an engineer who doesn't know the product's code base, let me tell you that we are talking SDKs (plural) and a service based company here. their was just one SDK with interesting, heavy lifting stuff and 9 more SDKs which were mostly wrappers and less advanced libraries. i got tasks in all of them, and 70% of my time went into maintaining those and debugging client side bugs instead of exploring the "already-stable-dont-change" code base.

so based on my vague understanding and my even more vague memory from 1 year ago, i tried to explain an overall architecture to that interviewer guy. His face was screaming the word "pathetic" from his expressions, so i thought that today i will try to decode the codebase in 12-15 hours, publish a cool article and be proud of how much i know a so called martech system design. their codebase is open sourced, so it wasn't difficult to check it out once more.
but boy oh boy i got so bored. unnecessary clases , unnecessary callbacks static calls , oof. i tried to refactor a few classes, but even after removing 70% of codebase, i was still left with 100+ classes , most of them being 3000-4000 files long. and this is your plain old java library adding just 800kb to your project.

boring , boring stuff. i would probably need 2-3 more days to get an understanding of complete project, although by then i would be again questioning my life choices , that was this a good use of my 36 hours?

what IS a correct usage of my time? i am currently super dissatisfied with my job, so want to switch. i have been here for 6 months, so probably i wouldn't be going unless i get insane money or an irresistible company offer. For this i had devised a 2 part plan to either become good at modern hot buzz stuff in my domain( the one being currently popularized by dev influenzas) or become good at dsa/leetcode/cp. i suck bad at ds/algo stuff, nor am i much motivated. so went with that hot buzz stuff.

but then this interview expected me to be a mature dev with system design knowledge... agh fuck. its festive season going on and am unable to buy any cool shirts since i am so much limited with my money from my mediocre salary and loans. and mom wants to buy a home too... yeah kill me

I'm slowly realizing how much goofy code I put in my branch and overlooked. This code review is going to be interesting...

Some examples:

import plots as lel
<h4 id="title">Crunchatize Me, Captain! </h4>
go.Scattergeo(name="cheese", ...)
webster = { ... }

The commit messages are even worse.

- 'horizontalize' link list
- very messily hack in <feature>
- partially refactor some of the awful code from previous
- Remove one annoying space
- make background color less annoying
- remove seemingly useless property
- minor fix
- Apparently it's possible to center a DIV. Who knew?
- Made some cool bar graphs
And then there's just a bunch of reverts.

I need to go buy a rubber duck so it looks less like I'm talking to myself. Trying to pull out and refactor some shit functionality in a WordPress theme because the client NEEDS it. Frankly all it is doing is creating a custom post type, but they're used to the way they've been doing it and I'm stuck with dealing with it. I generally like this part of my job (my face in the code) but trying to read this huge mess of code with no standards is driving me insane.

"What in the hell are you doing here?" "Why do we have variables for $thedata, $the_data, and $theData?"
"Why are your brackets on the wrong line sometimes?"
"Why is each line in this function enclosed in it's own PHP tags rather than around the function?"

At least if I had a duck I could say I'm talking to him.

So apparently my code went to prod more or less all right. Phew, the deadline was this week-end. This project have been sitting there for month, they gave me the technical requirement and never bothered to ask the stakeholders about it. When the contract went in, they started to freak out it wasn't usable.

The thing is, this project had way more moving part and trying to threat video in the frontend is not the easiest. But now is REFACTOR TIME.

I dream of getting rid of the browser video api (too flaky), download the bitmap directly and render it in requestAnimationFrame. I call it just-in-time rendering. I think i'd need to put a decoder in aws, I did it already with ffmpeg. I could not manage to put it in streaming mode though, so it was still a bit slow, but i could decode, write and re-encode faster than the video player speed.

What do you lads think? Doable or not? I at least need to general tidy up (this codebase have grown organically without any fucking direction from above, like this project took all my time on the technical side, I did not have time to run after people to get specs), centralizing state, improve monorepo and tooling, perfs,...

Hopefully they understood i cant keep adding whatever feature they want today.

2 weeks+ ago I made a PR into our codebase containing sample refactor that streamlined a significant portion of code. Also, I did refactor only on two handler packages (for MVC folks, that's Controller) as proof of concept, to figure out how convinient / logical the part would be for everyone.

We have rule of 2 approvals for merge (for 5 team members)

While writing refactor, it obviously blown up a lot of unit tests, but still coverage was fairly poor (that stuff was rushed, there was back than no time for unit tests). After my refactor I spent couple of days writing tests that hit fairly sweet (comparatively) coverage. (I managed to bump coverage from low 20s to high 80s, and have less code for tests)

I got first approve pretty much immidietely, other team member was on vacations, and 2 of them forgot.

We generally try to close PRs fairly quickly (usually same day kind of deal), but that one was just.. hanging in there. So I pinged everyone to re-check it to greenlight it but of course, loo and behold, merge conflicts arised. I ended up fixing actual logic (just some method signatures changed, not a big deal) and ran the units.

So, one of that handlers got quite a few of edits, and guess who is pretty much rewriting unit tests for second time now...

Dude, sometimes I question why tf I even bother with these tests... Feels like sabotaging my productivity, especially with bullshit like that