The year in review 2025
Daniel Opitz
12 Dec 2025
2025 wasn’t a year of big launches or loud announcements.
It was a year of slowing down, refactoring, and choosing quality over speed.
Looking back, that was exactly what I needed.
Slim 5
I began active development with Slim 5 and made a few architectural decisions that felt necessary.
- The DI container is now mandatory, followed by consistent refactoring.
- The middleware stack was changed from LIFO to FIFO, making behavior more predictable.
These changes weren’t about features.
I wanted Slim to behave more like frameworks such as ASP.NET, especially in terms of middleware order, to make things clearer and easier to maintain in the long run.
Statistics
My GitHub activity dropped noticeably this year:
- ~221 contributions in 2025 (432 in 2024)
- Most work focused on maintaining existing packages
- Several packages were prepared for PHP 8.5 compatibility
Less visible work, but more responsibility.
Maintenance is not glamorous, but it matters.
- Only two blog posts in 2025 (including this one)
- No new YouTube uploads
This wasn’t neglect - it was a conscious decision to reduce output and increase focus.
Development Environment
With XAMPP being discontinued, I explored modern alternatives - mainly Docker / WSL2.
While powerful, Docker didn’t work well for me as a daily PHP development environment:
- complex setup
- performance issues
- frustrating debugging with Xdebug and PHPStorm
So I built my own solution instead:
a local setup script that installs and configures everything automatically - fast, simple, and debugger-friendly.
Windows 11: Progress in the wrong direction?
This year, I switched to Windows 11 - and honestly, it feels like one of the worst Windows versions so far when it comes to usability.
What really surprised me is Microsoft’s current direction: adding AI-powered features everywhere. Notepad now comes with a rich text editor. Yes - Notepad. Hahaha.
At the same time, Windows keeps consuming more and more memory. A big reason seems to be the ongoing shift toward web-based UI technologies inside the operating system itself. More abstraction, more overhead - and not necessarily a better user experience.
Sometimes it feels like progress for the sake of progress, rather than improvement for the user.
Learning
Instead of conferences, I invested time in Udemy courses.
Why this worked better:
- targeted topics
- flexible learning pace
- easier integration into everyday life
Less travel, more learning.
Observations on PHP
Is PHP losing momentum?
It feels like overall PHP interest has declined:
- less activity in the Slim framework forum
- fewer discussions than in previous years
- the job market for PHP in my area has nearly disappeared
That said, this might be misleading.
With AI tools, many developers simply need less community interaction to move forward.
AI is a Tool - Responsibility is still ours
One of the biggest topics in 2025.
AI has become an extremely valuable tool in our daily work. It helps us move faster, find solutions more easily, and reduce repetitive tasks. Used well, it can significantly improve productivity.
However, faster produced code does not automatically mean better software.
We are still responsible for what we build. For its quality, its security, and its long-term impact. If we allow convenience to replace understanding and care, this industry risks losing the trust of the people who rely on our systems every day.
AI should support our thinking, not replace it. Craftsmanship, experience, and ethical responsibility remain essential - especially now, when producing software has never been easier.
A job market in transition
The job market changed dramatically this year - and not in subtle ways.
What once felt like a stable, growing environment suddenly became more uncertain, more competitive, and noticeably quieter.
Open positions disappeared faster, hiring processes slowed down, and expectations increased at the same time. Companies seem more cautious, more selective, and less willing to take risks. For many, this shift felt abrupt and uncomfortable.
AI, economic pressure, and years of overhiring are clearly reshaping the landscape.
The result is a market where experience, adaptability, and real skills matter more than titles or buzzwords. It’s no longer enough to “know the right tools” - understanding fundamentals and being able to think critically has become essential again.
This change is unsettling, but it might also be a necessary correction. One that reminds us that long-term value is built through substance, not hype.
Where I see things going (2026)
Looking ahead, I expect a few shifts:
- Stronger separation between frontend and backend
- server-side React has introduced serious security concerns
- A return to backend technologies built for backend work
- .NET, PHP , Java, and so on…
- JavaScript does not belong everywhere
- Fewer microservices by default
- more modular monoliths
- selectively extended by a few satellite microservices
- AI will drive more industries, not just IT
- Writing code will be faster, but writing good code still require experience, responsibility, and critical thinking
Personal Notes
One unexpected highlight of 2025: I started baking bread.
BTW: The software industry is slowly losing its sense of real quality. It reminds me of bakeries that switched to pre-mixed flours to produce faster, cheaper bread that isn’t even good for our health anymore. Bread (and software) needs time, passion and experience to become good.
Lessions learned
If I had to summarize 2025 in one sentence:
Quality beats speed. Most of the time.
Rushing leads to stress, poor quality, and burnout.
Thinking things through and deciding clearly is healthier - and more effective in the long run.
Looking toward 2026
In 2026, I plan to:
- write more about other topics, such as C#, .NET, and ASP.NET
- continue focusing on architecture and maintainability
- publish less, maybe shorter blogs, but with more intent
2025 reminded me that quiet progress is still progress.
I wish you all a Merry Christmas, good health, and a Happy New Year 2026 🎉🥂✌️🍀.
Daniel