Debugging

Visual InterDev 6.0's HTML editor. Screenshot via irt.org, used for historical commentary; Visual InterDev is a trademark of Microsoft.

Further mutterings on finding joy in development.

Debugging was the first thing that really hooked me on development. I didn't come into the software world through a traditional path. I was self-taught, and for the most part I viewed coding as a fancy cousin to data entry. I read a requirement, typed it up in a formal language, pressed F5, and watched it do its thing.

Two things shifted my perception of development: Visual InterDev and its integrated debugger, which could attach right to Internet Explorer 6 to step through client-side script. They gave me the ability to see code in action. To visualize code execution flow in ways that had eluded me in my compile->run->check logs era. With these platforms in hand I fell deep into the thrilling churn of triaging critical issues.

I'm surprisingly good at debugging. I think there's a few factors in play but at least some of it comes from being slightly ignorant. I was the type of developer that had to turn to AltaVista for the correct ADODB.RecordSet creation syntax. I wasn't a "real" programmer and that put me in the position to accept that there could be an error anywhere. From there I could systematically narrow my field of search.

Later in my career I'd stumble across the Old New Thing by Raymond Chen. I would always enjoy his Psychic debugging blog entries. The premise being that there are key thought patterns you can use to improve your ability to find the haystacked-needle. This tidy example encapsulates the concepts nicely, a lot of debugging occurs before diving into the code. It's a mindset, especially in the sense that you need to push urgency and panic to the side and think about what is possible and probable.

One of the first soul crushing moments in my AI journey was realizing that I could paste an error message into a prompt, point an agent to a code base and a few sips of coffee later I'd have a very likely culprit. Worse still for web issues I could just flip on the Chrome extension and wait. Sure there'd be some flailing and a lot of wasted tokens but I'd eventually get a plausible answer.

My special talent is that I can look at a system, accept that there's an issue, reason about the most likely culprits and mentally step through the execution. That allows me to not only solve problems but brings me a significant level of delight; but now I've been replaced by a school of statistical minnows.

Just yesterday I was annoyed by a jittery website; where if you paused the scrolling, it would enter a flickering state scrolling up and down frenetically. It was in Safari so no extension. I only had to ask Claude if it could help, and in its overly pleasant response it mentioned "position: sticky". I was annoyed because it was right without even trying.

I keep telling myself that this is a win - I don't need to mind weeds, agents will happily come and pull them for me. I now have infinite time to focus on BIG IDEAS. Except I know that with everything I do, a tremendous amount of joy comes from the littlest things. When I play guitar I'm happy if I can play a song, but I'm thrilled when I notice that there's no string buzz where there was some yesterday.

What's a logic-flow data entry specialist supposed to do when those small sparks are tamped down? There's still a chance that someone might want Artisanal Software - hand woven for that human feel. There are the outskirts of technology where there isn't a large enough sample set to allow effective agentry.

In reality I think the joy is going to have to come from people. I spent some time as a dishwasher, and the work sucked. Every day I went home covered in grease and smelling like hollandaise sauce. Yet I enjoyed working because of the wild and wacky players who slogged alongside of me.

I was never special because I could find that needle, I was special because I could backstop my teammates. The skills I brought: calmness under pressure, the ability to visualize systems and flows, the ability to set ego aside; allowed me to fill a crucial role on a lot of teams. Those skills still matter and I'll need to find another way to apply them, with another team that needs completing.

The thing I'm working on accepting is that the joy was never about a specific act, it was about finding the small subtle value that only I could bring to the table. It's going to be a journey but I'm an irrational optimist who thinks I'll find new ways to win, and a new team to do it for.