2022 Tattered

It was kind of busy year, (Pretty sure 2023 will be much more busy but anyway) but somehow managed to get through. Also it was my third year at current job, which I had to prove some career development, better performance with good result and personal achievements:

What I learned

AWS Solutions Architect: Professional

I had to renew my old AWS SA Associate with newer one until July, so I took this opportunity to challenge with professional certificate. Quite challenging, though fortunately got passed with good score.

Vite & Vitest

Vite was one of options that out team had in mind to introduce in development, and review if it is suitable for production level, but I thought it’s speed wasn’t a critical factor. It was Vitest which made us shift to new bundler.

Vite is really fast, and there are plenty of plugins (react, css injectetion) which allow us to migrate from other bundler easily. Vitest configuration file is simple enough to understand. These two were critical factors to make a decision.

Domain Driven Design

Control panel manipulates A LOT OF terminology, and has thousands of buttons with similar(but different) functions. In order to organize ubiquotous language and terms, Out team implemented Domain-driven design in newer architecture of next version of Control panel.

To be honest, a definition of domain itself is quite vague, and it depends on each one’s interpretation. Every consultation using these design patterns, a team should take their own way to digest and make practical way to put it on right way…

Few regrets

Code productivity

Type assertion is a great tool, but I found very tired to follow a whole maze made of Typescript’s operators. Even typescript quizzes didn’t helped me to organize a whole assertion. Harder assertion makes test code longer, duplicating test cases with creative errors.

I’m not a huge fan of concrete application; because I know every front-end application has a short lifecyle, not because of advanced languages but also the business models’ alteration affects mostly on front-end first. New designs and updates are demanded frequently. That is why we recently focused on refactoring core-level only. Even though the presentation layer are demanded to change, we wish we can recycle our core-layer sources to implement new functions and designs with no problem. Sincerely, I’m not pretty sure how strict we have to be on type-guarding things and which will cost a lot of time on refactoring…

Exciting Discoveries

AI, mostly ChatGPT

Whole globe was chanting AI like crazy. I remember exactly a decade ago I was trying to build a translation dictionary (which was a kind of data farm) implementing Elasticsearch, to create a translate-recommendation service which allows people to preview a similar translation occured recently. When I saw ChatGPT, I was astonished to an inmense size of big-data they’re operating. That was what I wanted to create 10 years ago.

By far I know its limitations. Both legal and lucrative problem are not easy dilemma to deal with. Most importantly, data-sources that they had will be reclaimed since their propaganda just made them acknowledged their sources’ opyright are not revisited. But somehow, I still believe that AI will serve a great role in our productivity sooner or later.

lima/docker

Since Docker Desktop has released the enterprised version, I was looking for alternative option to replace desktop application to open source. Lima creates a virtual machine with broad access to hard drive without mounting directories and etc; I just followed the instructions to create docker environment, so I didn’t experienced a whole lima magic yet. Since I will get a new M2 Macbook next year, I will leave that fun to next year.