The "Right" Keyword

I'm sipping coffee. It's my third cup today. My eyes are fixed on the cherry tree just outside my office window—it's heavy with fruit. The branches bend under the weight of ripe, crimson cherries. A few have already dropped to the ground. Soon, we'll taste them. My mouth waters at the thought.

I glance back at my notebook. The problem stares back at me. Is it really new? Has no one tackled this before? I feel a brief surge of excitement—maybe I’m the first to approach it this way.

No. I reread my notes. That’s too hopeful. A problem this fundamental must have crossed someone’s path. And yet, I’ve been stuck on it for two weeks. I’ve tried every keyword I could think of—“structureless data,” “non-structured data,” and dozens more. Still, no meaningful results in the research databases. Nothing leads me to the insight I need.

A colleague walks in. I hesitate. He's from an entirely different research area. I’ve considered asking him before, but always backed out. Still, maybe a fresh pair of eyes wouldn’t hurt.

I explain the problem to him. He listens patiently, then offers something that pulls me back to the basics—fundamental data structures from our undergraduate years. I nod, half-smiling, though I feel the urge to retreat.

“Sorry,” I interrupt. “I just remembered—I have a meeting with my professor.”

It’s a lie. I have nowhere to go. There’s no meeting, and I don’t want to return to my desk and sit again with this frustration.

I linger near the corridor, unsure of what to do next. Then, by chance—or luck—my professor walks by. I smile, and almost without thinking, ask, “Do you have a minute?”

“I’ll be back in five,” she replies. “Wait in my office if you’d like.”

I nod. I don’t want to go back to my room. I wait.

When she returns, I explain my situation, this time in full. “I have a large amount of data,” I begin, “but I don’t have a defined structure in advance. I can’t create a traditional schema because I don’t know yet what form the data will take. I need to analyze it later—but how?”

I tell her about the hours I’ve spent combing through research papers, how none of the terms I searched seemed to match my needs. I wait, hoping.

She smiles. Her response is gentle, almost amused. “Ah, nothing new,” she says. “People have worked on this. It’s just called something else.”

I feel my hope flicker. My earlier excitement begins to dull. She notices.

“Have you looked into schema-on-read?” she asks. “Sometimes also called schema-on-need. It’s used when data is stored without predefined structure, and the schema is applied at the time of querying. It’s common in big data systems. There’s been quite a bit of attention on it recently.”

She turns to a bookshelf and pulls out a volume. “Check the conference proceedings from last year,” she adds. “This one has a few promising papers.”

I take the book from her hands and thank her. As I step out of her office, I feel something different—less excitement, perhaps, but more grounded. The kind of clarity that comes not from a sudden flash, but from finally finding the right words after searching in the dark.