How I Resurrected My Blog!
Ten years later: how I recovered all my old articles from Kirby CMS, what happened in the meantime between backend and architecture, and why GWT was right after all
If you’re reading these lines, the miracle worked: this blog is back from the dead after almost ten years of silence. Same domain, same articles, same author — with a bit less hair and, I hope, a few clearer ideas.
Software archaeology
The old blog ran on Kirby CMS: PHP, a hosting plan to look after, and that constant feeling that sooner or later something would break. And indeed it did. Luckily I had done one thing right: I had kept everything in a Git repository, and Kirby — bless it — stored articles as text files with the body already written in markdown.
So the recovery was archaeology more than magic: a script converted Kirby’s fields (Title, Date, Tags separated by ----) into standard YAML frontmatter, fixed proprietary tags like (link: ...) and (image: ...), and rescued images and code snippets. Out came all my articles from 2015 onwards, including the English versions and even three drafts I had never published.
The result is a static site built with Astro: I write a markdown file, run git push, and a minute later it’s live on GitHub Pages. No database, no PHP, no hosting to babysit. Faithful to my motto: complicating is easy, simplifying is hard.
Ten years in one paragraph
I haven’t been standing still in the meantime. My career started on the backend — Java, Spring, enterprise systems — and over the years it shifted more and more towards architecture: Domain-Driven Design, microservices, cloud architectures. Today my job is as much about drawing the boundaries of a system as it is about writing its code, and honestly I couldn’t give up either one anymore.
And the frontend? Here comes the fun part.
GWT was right (told you so)
Those who read me back in 2015 will remember articles like Javascript Is Dead? and the series on scalable Javascript projects (in Italian). At the time I was developing with GWT and defending a thesis that earned me the antipathy of half the web: Javascript alone was not enough to build serious applications. You needed types, modules, tooling, a compiler lending you a hand before runtime.
Over these ten years I’ve worked extensively with Angular first and React later, on projects of every size. And what amuses me most is seeing that the ecosystem landed exactly where I hoped it would: TypeScript brought the type system I was calling for, ES modules and bundlers brought modularization, frameworks brought structure and patterns. What GWT did by compiling Java to Javascript, TypeScript does today by compiling… to Javascript. The direction was right; the tool was simply ahead of its time.
It’s not a personal vindication — it’s confirmation that certain software engineering needs are universal: sooner or later, every platform that grows has to face them.
What comes next
The plan is to get back to writing regularly, in a more agile format: bite-sized articles, short and concrete. There’s no shortage of material: I attended reactjsday in Verona in 2025 and came home with insights that deserve to be explored one by one — from rendering patterns to the new directions of the React ecosystem.
If the old blog was a developer’s diary, this one wants to be the notebook of an architect who never stopped writing code.
See you soon. And this time for real :)