Huy Nguyen

My history with a "database project"

As you can guess from the title, this is not my first time thinking about making a database. I've tried ever since summer of 2024. Then again in summer of 2025. And, like an yearly occurrence, I give a shot again now, in summer of 2026.

It's kind of sad I never kept a log of what I do for the attempts in 2024 and 2025. The attempts "failed" (again, you could've already guessed), but that doesn't mean I didn't learn anything. So this time, I want to keep a log of what I do. I like writing things down on paper, sure, but maybe by writing in this form, someone might find inspiration from what I do?

First time trying

It was summer of 2024. I'm an unemployed first-year college student. Not having a job to do in the summer means having a ton of time. Even though I'd spent the 2 semester prior learning (and learning quite well, if you ever trust grades reflect your actual abilities), I felt like I couldn't get enough of coding. And I want to code more. As it turned out, this wasn't good for me...

I "recruited" three of my friends, forming a little group on GitHub. I made the project. At the time, all I really knew was C (or Java, which I really didn't like). I actually knew quite a lot of C for a novice, I must say. I could write Makefiles, I could structure projects, I could turn on all the relevant compiler flags. Think about it, that's almost all of the C you'll need for the vast majority of your toy projects. And for a novice, I was quite good at using Git as well... or at least, good enough that I can guide my other friends with it.

I know it's a bit hard to really put what I just said into perspective. But, it was 2 years ago, and I have no remaining form of memoria for the attempt. Yes, not even the Git repo. My memory is unreliable, too. But, there are certain things that you just can't forget...

Back to topic. I had a small group that wanted to code a bit more. I knew how to use C, and could guide others on some basics of C. I knew how to use Git. Sounded like a smooth sailing trip. What could go wrong?

First time stumbling

I loved shiny new things, especially if the shiny new thing promises a way out of a problem. And in C, you meet the problem very quickly. Oh, you want a dynamic array? Write your own. What'd you say, a generic dynamic array? Well, you could try deal with macros I suppose... If only there were a language, very compatible with C, shares a lot of the same concepts with C, and has generic containers? That's exactly what C++ looked to me.

Now, I'm not here to hate on C++. It's probably the language I use the most. And it's also a very deep hole, which does satisfy my nerd requirement of feeling like I learn something even though that thing is never getting put to use and I would forget about it in a few days. But in this little story, the decision to jump boat from C to C++ proved to be one of the many wrong choices I made in a very short period of time.

So, what was the state of the project when I decided to jump boat to C++? I didn't even have a CLI working.

But, it's just a CLI, right? Even without a nice CLI parser, it's not difficult to write a CLI in C. And if I planned to make a database, I better knew how to make something as "simple" as a CLI...
That project really stopped at the CLI.

Oh, but at the same time, I also switched from Makefile to CMake. And I also added GoogleTest, even though I had nothing to test yet. But it was just delaying the inevitable. That is, I realised all that time I spent resulted in a barely functioning CLI, and nothing else. And with that, my motivation for the project was pretty much dead.
And I was busy with another project I came up with. At least I got that to a somewhat functioning state. It was my regex engine.

I wish I had something to show on here for this failed attempt. Too bad.

Second time's not the charm

The second attempt, I set out to do this on my own. Compared to the first, I had accumulated more experience. I had two projects, both in Rust, to a working state; the regex engine mentioned above, and the capstone project for one of my classes. Wow, I still have this one. I have a terrible habit of nuking any project I felt "disappointed" about from the earth's surface, so it was fortunate this one ended up successful for its purpose.

I considered myself the "leader" of that capstone project, which makes a lot of sense: I came up with the idea, I designed practically every thing, from how the CLI should look like, what the minimally functional feature set is, how it keeps track of files... basically almost all of the details. Looking back, it was a banger project. And it surely did fill myself with confidence that I can do whatever I wish for (in coding, at least) and succeed.
It was another year of learning's worth, so I gained a lot of experience as well. This time I got a CMake template that works quite well, so no need to write much CMake again. And the template had all the belts and whistles I could think of using; unit testing, CI, automatic documentation generation, somewhat comprehensive build instructions, integration with the conan package manager.
I can spend a lot more time designing and coding now, not having to worry about this bureaucracy called the build system.

Too bad, I made some more terrible choices. With too much time spent designing, alongside the knowledge of my last failed attempt, I decided to go big. Think of all the features I could, "design-pattern" the code the moment I begin, and whatnot. This time I did get farther than just the CLI. But not much. I had a lot of separate pieces that couldn't work together. One moment I was coding one of the components, the next I was doing another, just because some thought pop up.

And again came the inevitable. After a while, I realised I'd done nothing worth showing (or so I thought).

Are you really sure third time's the charm?

No. But I do have hope.

By writing about it, it's already different than my last 2 attempts. Even if my code fails, I will still have something tangible to show to other people, or more importantly, to myself.

Really, for a toy project, the most I care about is to have fun coding. And from that second attempt, I know that I can't really theory-craft something intangible (intangible goal, intangible scope, and many more intangible things).
Best way to have fun? YOLO.

Currently, the repository is private. But in the following entries, I will write about the development process.