I Don't Need to Know Everything. I Just Need the Fucking Pointer.
I don't know everything. I can't know everything. And I've never particularly wanted to memorize every implementation of everything I've ever encountered.
What I need is much simpler.
I need to know that a thing is possible. Then I need a pointer.
The rest I can look up when it becomes relevant.
The Index Is Not the Database
I'll be searching for one thing and accidentally find something else.
"Oh shit, that's interesting."
I may not have a use for it. I may not know how it works. Sometimes I don't even remember where I found it.
But something gets filed away.
Years later, a problem comes along and a connection appears:
"Hang on. I've seen something like this. I know it's possible."
Now I can go find the actual details.
I don't need to carry the whole implementation in my head. I just need enough of an index to know where to start looking.
Amazon, 2001: Enter the Fucking Tortoise
At an Amazon interview in 2001, I got a programming question about detecting a loop in a linked list.
I called my friend Chris for help.
Chris is the guy who, in 1998, was so annoyed with Netscape that he started writing his own browser. Not a plugin. A browser.
Chris called his friend Tom, who'd been programming in C++ since he was thirteen.
And there it was: three programmers connected by two phone calls to recover an algorithm.
Floyd's tortoise-and-hare cycle-detection algorithm.
I have looked that fucking thing up repeatedly since then. The name has never reliably stuck.
But the pointer stuck.
"Linked list. Cycle detection. Two moving pointers. Something about a tortoise and a hare."
That's enough to find the rest.
Cycles Are Part of the Requirements
Here's the best part.
People use Floyd's algorithm to detect loops.
Meanwhile, my thinking routinely wanders down one connection, follows another, and eventually circles back to somewhere familiar.
And that's fine.
Cycles are part of the fucking requirements. Floyd can fuck off.
I'm not debugging a corrupted linked list. I'm wandering through a graph of associations.
Sometimes returning to a previous node is exactly what makes the next connection visible.
Sometimes I revisit the same idea with different context and see a different edge.
That isn't a bug I need to eliminate.
It's the point of the exercise.
Pointers Across Domains
A thought about software can remind me of a religious concept. Something in Buddhism can connect to Stoicism. A scene from science fiction can illuminate a question about history.
I am not saying the ideas are equivalent.
I'm saying the pointer takes me somewhere related, and I can inspect the connection when I arrive.
A graph doesn't require every node to contain the same type of object.
Neither does my curiosity.
Rubber Ducking the Brain
Programmers already know this experience.
You explain a problem to a rubber duck, a teammate, or the empty room, and suddenly notice a relationship that was hiding in plain sight.
Maybe the information was present all along. Maybe the act of explaining brought the relevant pieces into contact.
I don't need to settle a theory of cognition to enjoy the result.
I know what it feels like to describe one thought and have it activate another.
Then another.
Then another.
And occasionally the solution appears when I wasn't even looking for one.
The Pointer, Finally
I could have memorized Floyd's name in 2001. I could memorize it now. Maybe I'll forget it again next week.
Who gives a fuck?
The useful part is knowing the technique exists and being able to retrieve it when needed.
And when my own thoughts loop, I won't be running cycle detection to stop them.
I know there's a fucking loop. I put it there.