Here I am, on my admin panel, writing these words; but how did it come to be ?
(This post does not go into technical depth and is more of a surface/narrative one, decided to make it a devlog still because it's related to this website itself, which is obviously closely related to myself. Hope you'll like it :) )
Context
Some historical context: as a young guy imagining myself as a creator with an online presence, I always felt the need for an unified platform where I could share publicly my work, my thoughts, my projects, my aspirations and many other things that I surely forgot about since then.
So this was awkwardly established through various means over time: a small Wix website (that lasted 1 week before I got bored of it), then a discord server, then I thought about simply using Twitter, then another discord server etc... You get the idea.
The thing was that most of these solutions weren't personal enough, I couldn't get the feeling of "Yeah, this is my place", kind of like having a public online hangout place, see what I mean ? Thinking about it now, I guess the first Wix blog approach was the best out of these for what I wanted, but at the time managing this was a bit outside of my skillset.
But now, things are different; I've been building my dev skills over the years, have been working more on web stuff especially in these recent two years, so when I decided to try doing some light freelancing and realized I would surely need a website, some lightbulb went wild in my head and this (since then kinda buried) desire came back to mind.
"I can just make my own platform now!"
The planning
Okay, I have the idea, this is a good first step. But now I have a few other thousands-ish steps to take. I guess the second one is knowing where I should direct this stream of steps.
What do I need ? First obvious thing is a projects list, I want to be able to easily show to people what I'm working on and what I've already accomplished. Okay great, should be easy enough. Let's do a blog side too, so I can have an expressive outlet, maybe it could interest a couple people looking for context surrounding the project themselves. I also wanna be able to do more simple logs, kinda like tweets, short thoughts that do not need an entire blog post to themselves. Also giving the ability for visitors to interact with the content would be great. Wait, this means rate-limiting, (somewhat) enforcing unique actions, so this implies auth ? Okay I need to reduce scope already.
Turns out planning something from scratch is pretty different compared to tinkering something from emptiness to something more fleshed out, imagining it as I build it in parallel. I wanted to force myself to follow another workflow than the usual, but had some trouble with it, and in the midst of all this I came to the realization that this is why people decouple their projects in incremental versions.
I have a confession: I am a perfectionnist. Yeah, like most people looking for a job, I know... but I'd actually say this is a weakness, and a HUGE one at that. Having trouble saying something is finished because there is still area for improvement is a poison for a creator. The reality is that most things are left "unfinished", they're just "finished enough", which I knew but had trouble applying to myself. And as days went on and I kept planning and prototyping this very website you're on, I felt the compounding waste of time looming over me, growing as time went on while I was micro-refining the ideas and prototypes.
And once this feeling of waste had grown enough, I knew I HAD to do something about it.
I thought for some time, and my mind wandered off thinking about emergence, and how constraint shape behaviors. May sound dumb, but these kinds of parallels often work wonder for my mind, and so I gave myself a fixed set of days for finishing the prototype, from it making a complete plan, and then actually doing the work. And it worked ! All I missed were self imposed constraints.
I knew it would come with frustration. I'd see all the small weaknesses and issues I would have spent hours on fixing until it was good enough to me. But if it was the price to pay for actually having something shipped, it was well worth it.
Putting the "dev" in "devlog"
Okay yeah this is a great story time, congrats to me amazing woohoo. As this is the first post on this site, I thought some surrounding Lore™ would be interesting, both about me and the site itself, but now I also wanna get a tiny bit more technical. I'll slide in a few anecdotes too.
The setup
I'm a "do-to-learn" type guy, I like hearing about the theory, I like to reason, but I know the best way I'll ever crystallize knowledge in my brain is by actually doing stuff, and mostly doing stuff wrong. So I wanted to avoid using things for simplifying my life too much, at least on things I wasn't confident in; I wanted the full SSH, git, actions .sh scripts experience, not a panelled one. Or at least not one externally managed.
I decided to go with a VPS setup for the site, it wasn't my first time (I made 10alect before this website as an experiment) so it was relatively intuitive to integrate with my plans. I used vercel for the domain, had to tweak the DNS to route to a new VPS that could host both websites. Not the best practice, I know but, hey, it's cheaper that way, and that's one of my main non-self-imposed constraints, so one VPS it was.
Migrating 10alect was another story that I won't detail here, but it came with setuping nginx for routing, with one global rerouter for http->https and one specifically for the subdomain. Spoiler: this would bite me in the ass later when I'd implement the mailing; a pretty small bite in hindsight, but still. Anyways once this was in place and everything was correctly linked up, I could finally work on the site itself !
The building
I had an idea already of what I wanted to use, I know svelte/sveltekit relatively well as that was my go-to web framework for a while, I had already used SQLite in numerous small experiments, so these two were considered already from the planning phase.
The prototyping proved real useful here, I was actually not that used to prototyping things as heavily as I did for this, but as I had a really defined vision for the style it was kind of needed. Now I know that I'll keep this in my usual workflow, I just need to give myself a time limit for conceptual exploration as said earlier.
The core of the website itself came together pretty quickly (thank you Claude Opus 4.5 for your service 🫡), all the data schemes and overall specs were already written and not a lot of repair nor adapting was needed to make this work. Remember that "planning this stream of steps" thing ? That's here that I realized how powerful it really was. As said earlier too, I'm a to-do-learn guy, and do-wrong for efficiency, but for that to work I need to realize I'm doing something wrong, and this was yet another such moment.
Okay cool, progress ! Both personnal and project-wise. But there are still a few features missing for the site (didn't plan the content managing side that much before, only the overall idea) and I have to neuroplasticitify(?) this newfound knowledge, so I went on to make myself an admin panel.
Lessons learned
This time was gonna go swiftly. At least that's what I wanted.
Planning phase: structures, schemas, layers, functions, helpers; boom, plan ticked off the list.
Short prototyping: overall layout, visuals (I like my work environment looking good weirdly enough), functionalities, plan adjusting based on all these; bam, satisfying preview.
Porting: Thank you unnamed agentic model for your service 🫡
Review and refinement phase: enforced security on the whole admin layer, verified VPS-side users, permissions, processes and routing, added a few light QoL features for myself.
I had this small iteration loop going on now for working on features, which at times was feeding a positive feedback loop where the more I build, the easier it is, which was a neat feeling.
I did encounter some issues with the way mailing worked though; first time trying this out, nginx was catching things it shouldnt even though it was configured, well it was because of this http->https auto rerouting that was preventing the certificate validation. One more lesson learned.
In the end, I had a working private layer that would let me manage my content. I had added a few convenience tools for working on the website itself, some for managing the content, and that liberty feeling was great. This was exactly what I wanted from my minecraft's command blocks days. And for the first time in a while, I had a place to share this excitement, so I opened my panel's "Posts" page, and I started writing about it.