55,041,902 Lines of Code

I did some exploration on KDE's code base. It's amazing what you can find when you have almost 30 years of public history in git.

In doing so I ran some statistics on KDE's core software. That core is relatively well-defined, even over the years. It's the libraries (Frameworks), the desktop (Plasma), and the standard applications shipped together on a regular release schedule (Gear). Of course there are other fine applications in Extragear and elsewhere, but I didn't look at them for now.

In 2009 I did an analysis of KDE's sources and found 4,273,291 lines of code. So I was curious to see where we are today. The numbers are not 100% comparable, because the shape of KDE's core software has changed a bit, but when you look at what is in git (omitting graphics, translations, and other non-code files) you find 8,173,148 lines of code.

But wait, didn't the title say 55,041,902 lines of code?

Yes, when you look at the history, there is more. There is churn, code gets added, code gets modified, code gets deleted. It evolves. So behind every line of code which is shipped today, there are about 7 lines of code which have been written, changed, deleted. This is what you find in git.

So 55 million lines of code have been written to arrive at 8 million lines of code today. They represent progress, learning, and adaptation to a changing world. You see life there.

The most amazing part for me is actually not the code itself, but that there are thousands of people behind this code who have worked together for decades to continuously improve KDE's software. That's the real story.

P.S. Take the numbers with a grain of salt. They are based on a couple of assumptions and influenced by quirks of tools and history. This is not science, it's the report of a hobby code historian.

Invest in your identity

I have 30 years of documented history on the web and in my personal recordings. That defines very well who I am, what I do, how I see the world, and how people see me. I worked on that. Sometimes consciously, sometimes as a side effect of my job, my side projects, my community work. Now that AI agents make it easy to use this kind of material, I have a base to anchor them, to build on what I did before and accelerate what I do, still staying me.

If you are starting now, you won't have this body of material to anchor your agents. So do spend some time building this corpus of what is genuinely you. Don't let an AI generate what you are. Write yourself, publish, think through your thoughts, give presentations. Small things are fine. They will accumulate over time.

Of course, tools will shape part of your identity. I used to do my presentations with xfig, printed on overhead projector slides. This was painful, but it shaped quite a bit how I worked and how the result looked. So it is part of my identity. The technical constraints did influence how I spoke, how I presented. It also shaped what I presented, because there was a bias toward what I could show with the tools available to me.

This won't be different with AI. It will shape who you are. But be aware, and make sure that there is a signal from the human in there. It's ok if it's imperfect, if it's a bit weird. It's ok if it's different. But make sure it's yours.

Shape that signal. That's you. That's your identity.