All workEssay 12 / 12

You are inside this experiment

The work changes.
The thread continues.

A portfolio can list what I built. I wanted a place that could also hold what I noticed, what I reconsidered and what I carried into the next project.

A thread, not a list

The most interesting work happens between projects.

A conventional portfolio asks for compression: a title, a screenshot, a stack and an outcome. Those details are useful. They do not always show why one project changed the way I approached the next.

The connection between a small data exercise and a financial replay system is not obvious from their cards. Neither is the connection between a chatbot’s recovery message and a production service’s failure state. Both ask what the person can understand and do when the expected path breaks.

I wanted this site to keep those relationships visible. The articles follow turning points and recurring questions, with room for context, compromises and changes of mind. They are an interpretation of the work, not a complete operational record of every system discussed.

Follow a recurring questionAcross the archive
01A registration form
02A conversation in progress
03A replayable portfolio

Different products, the same question: which state can the person trust?

Illustrative example · not live product data
The page you are reading

Depth should not make the entrance heavier.

The core reading experience is server-rendered. Text, headings and navigation are available before the interactive parts need to become active. Search, preferences and small demonstrations then add useful behavior around that foundation.

Motion can help establish a relationship or reveal a change. It should not be the only way to access information. A diagram needs labels; an interactive example needs an understandable starting state; a reader who prefers reduced motion still needs the complete argument.

Long pages also need orientation. A chapter list and reading progress let someone choose their depth instead of committing blindly to an endless scroll. A link to another story should offer a reason to continue, not simply increase the number of pages visited.

A personal archive can hold a change of mind as carefully as a finished project.
Room to continue

Leave a place for the next question.

A deeper story asks more attention than a project grid. That trade only makes sense when the extra space gives something back: a concrete decision, an explanation or a connection the shorter version could not carry.

Live signals introduce another trade. They bring freshness, but they also depend on sources outside the site. The reading experience should remain coherent when those sources are quiet or unavailable. A personal archive does not need every part of the internet to be working at once.

I do not want this site to become a monument to a fixed version of myself. It should be possible to add another chapter, revise an explanation and admit that a belief changed. That is why the thread matters more to me than a neatly finished collection.

The interactive examples are another part of that editorial choice. A reader can change a purchase price and see the consequence, rather than only read that earlier events matter. They can compare two recovery replies instead of accepting an abstract claim about conversation design. The interaction earns its place when it makes the argument easier to inspect.

Those examples also need honest labels. A simplified illustration is not a product screenshot, a live service or a performance measurement. Keeping that distinction visible lets the site be expressive without asking the reader to infer evidence that is not there.

If a question here connects with something you are building, that is a good place for a conversation to begin.

CONTEXT / LINKS

Related work & references

Selected public entry points and related portfolio pages. A company website or profile is context, not a claim that private source code is available.

Scope, context & further questions

These notes describe this site’s article structure and reader controls. They are not a performance benchmark.

  • Long-form archive: 10 stories. Ten statically generated articles connect the projects, decisions and recurring questions.
  • Reading architecture: Server-first. Article text is rendered on the server; focused interactive examples enhance it in the browser.
  • Reader control: 3 modes. Theme, density, and motion preferences adapt the experience.

Questions I am still working through

How can an archive stay useful instead of becoming a museum?

Let new work link back to earlier decisions, keep “what changed” visible, expose unresolved questions, and allow old conclusions to be revised. The archive should support continuation rather than only preservation.

The long-term challenge is maintaining these relationships without turning personal publishing into full-time taxonomy maintenance.

What is worth preserving from years of work?

Not every artifact. Preserve the decisions, constraints, mistakes, turning points, and explanations that still help someone understand why the work changed direction.

The archive will always require editorial judgment about what can be safely forgotten.

Stories, products & useful placesEsc to close