The original intention of this website was to document what I've been working on. I do this not for others, but purely for myself, as a means of summarising what I've learned and to reflect on the outcome of my work / life / whatever really, as it unfolds. It's also a useful way to hold myself accountable for the ideas that I've started. The (big) caveat here is that articulating your thoughts out in the open can be intimidating, even if I end up being the only one reading them. It provokes a sense of vulnerability and I'm not sure why. It's probably of no surprise to any of us then, that the initial intentions of this project have not yet been realised, and it's ended up serving as another form of procrastination rather than a means for expressing myself...
Anyways, this is the third re-write of this thing—each time a little more pared back than the attempt before it—the first iteration used react and typescript to provide a whole load of unnecessary features (self-served fonts, copious amounts of css and questionable design). As you can imagine, this attempt was bloated, completely over-engineered and honestly, not an accurate reflection of myself (hindsight can be cruel 😂). The second attempt was a complete re-write of the original functionality using web-components (Svelte) and Tailwind—custom fonts now removed, brotli compression instead of gzip—the weight-loss was in full-swing. It was by this point that I was becoming increasingly aware of the performance (or lack of) in modern software and so, with this attempt I managed to trim 90% from the website's original footprint (400KB → 35KB compressed). For a simple text portfolio, I would still call this bloated, but on reflection, I learned quite a lot about the Typescript and Svelte ecosystem by using them in a (simple) production environment along with maintenance and improvements, if there were any. Svelte was such a positive experience in fact, that I went on to build a much larger project[1], something I didn't end up deploying but learned a lot from.
We're now ~2 years after inception (and with very little written), I have just spent 1-2 days throwing together this iteration after work. New experiences since my last attempt have materialised as a change in perspective and my priorities have shifted. I feel more strongly about simplicity than ever before (both in work and life) and I continue to whittle away at performance oriented design in my projects. I have gone back to basics with this attempt, pure HTML, sparing use of CSS[2]. and most notably, no javascript. The aim was to improve time-to-first-byte by keeping each page well below 14KB (~2-3KB raw) and embedding the HTML and CSS into a single file (this avoids chunked transfer-encoding and ensures that the entire page is delivered in the first packet of HTML)[3].
What stood out to me most here, is that you quickly realise just how much work these frameworks do for you and with my goal of keeping each file to a single page, I found HTML duplication to be a problem. In order to avoid this (repeating the same CSS and headers per page), I decided to throw together a quick and dirty HTML template-engine, aptly named Temple[4]. . This was an attempt to build a proof-of-concept that implemented the absolute minimum templating syntax required to get this project up and running. A quick look at the code will reveal that I have committed the cardinal sin of template-engines, parsing HTML using regular expressions (forgive me! 🙏🏼), however, I have a lofty goal to re-write this project using tokens and a simple lexer ;). I find coming up with new ideas quite difficult, and what I've liked about this project is that it all kind of happened organically. I'm excited to explore lexers and I'm excited to write some meaningful C | Rust. Wrapping it all up, I'm quite happy with the outcome and I learned more than I anticipated. I no longer have an excuse to avoid dumping my thoughts onto the internet, and the messages will get delivered a little bit quicker than before, whether that's a good or bad thing remains to be seen...