The Pantry Problem | Cooklang: recipe markup language

I realized something was wrong with how we manage food when I opened my third jar of cumin. Not because I love cumin that much, but because I'd forgotten about the other two.

This is embarrassing for someone who writes software. I've built systems that track millions of data points, but I couldn't track twelve spice jars. The problem wasn't intelligence or organization. It was that I was using the wrong tools for the job.

Most people treat their pantry like a closet - stuff goes in, hopefully comes out before it expires. But a pantry is actually a database. It has inventory levels, expiration dates, consumption patterns. Once you see it this way, the solution becomes obvious: treat it like one.

The Inventory Problem

Here's what most pantry management looks like: You're making dinner. You need chickpeas. You check the pantry. No chickpeas. You add them to a list. You go shopping. You buy chickpeas. You come home and find two cans behind the rice.

This happens because human memory is terrible at inventory. We evolved to remember stories and patterns, not quantity states. You remember making that great chickpea curry last month. You don't remember whether you replaced the chickpeas afterward.

The solution isn't trying harder to remember. It's acknowledging that computers are better at this than we are. Here's what my pantry.conf looks like:

[pantry]
rice = &#34;5%kg&#34;
chickpeas = &#34;3%cans&#34;
olive_oil = &#34;1.5%L&#34;
flour = { quantity = &#34;3%kg&#34;, opened = &#34;2024-12-20&#34; }
&#34;coconut milk&#34; = &#34;4%cans&#34;

[spices]
cumin = { quantity = &#34;1%jar&#34;, bought = &#34;2024-11-01&#34; }
coriander = &#34;50%g&#34;
turmeric = { quantity = &#34;1%jar&#34;, bought = &#34;2024-10-15&#34; }

This isn't about being obsessive. It's about offloading mental overhead. When I check what I can cook, the system knows I have chickpeas. When I generate a shopping list, it knows I don't need cumin.

The Freshness Game

The second revelation came from tracking expiration dates. Not the printed ones - those are mostly fiction - but the real ones. When did I actually open that jar of tahini? How long has that leftover soup been there?

[fridge]
milk = { expire = &#34;2025-01-10&#34;, quantity = &#34;1%L&#34; }
&#34;leftovers curry&#34; = { made = &#34;2025-01-04&#34;, quantity = &#34;3%portions&#34; }
tahini = { opened = &#34;2024-12-15&#34;, quantity = &#34;200%g&#34; }
spinach = { bought = &#34;2025-01-05&#34;, expire = &#34;2025-01-08&#34; }

The unexpected benefit: meal planning becomes automatic. The system suggests recipes based on what needs using first. That spinach becomes saag paneer on day three, not forgotten slime on day eight.

The Shopping List Revolution

Here's where it gets interesting. Once your pantry is a database, shopping lists generate themselves. Not just "what's missing" but "what will be missing when I cook these three meals."

I plan to make pasta arrabiata, chickpea curry, and banana bread this week. The system knows:

Arrabiata needs tomatoes, garlic, chili, pasta
I have pasta and chili
Curry needs chickpeas, coconut milk, rice, spices
I have everything except fresh ginger
Banana bread needs flour, bananas, eggs, butter
I have flour, need everything else

The shopping list writes itself. But more importantly, it's comprehensive. No forgotten ingredients, no duplicate purchases.

The Freezer State Machine

The freezer is where time stops, but not really. Frozen food is in a state of suspended animation, not immortality. Track it wrong and you get freezer archaeology - mysterious foil packages that might be fish or might be cake.

[freezer]
chicken_stock = { quantity = &#34;4%cups&#34;, frozen = &#34;2024-12-20&#34; }
ground_beef = { quantity = &#34;1%kg&#34;, frozen = &#34;2024-12-15&#34; }
pizza_dough = { quantity = &#34;2%balls&#34;, frozen = &#34;2024-12-28&#34; }
pesto = { quantity = &#34;200%g&#34;, frozen = &#34;2024-11-30&#34; }

The dates matter. First in, first out stops being a suggestion and becomes policy. That pesto from November gets used before making new. The system suggests recipes that use what's been frozen longest.

The Paradox of Structure

People resist this level of tracking because it seems constraining. But structure creates freedom. When you know exactly what you have, you stop worrying about what you might have. When shopping is automated, you stop thinking about shopping. When freshness is tracked, you stop playing refrigerator roulette.

The goal isn't to turn cooking into accounting. It's to remove the administrative burden so cooking can be creative. You don't track your pantry because you love spreadsheets. You track it so you never have to think about tracking.

Implementation

Start small. Pick one shelf, one category. Maybe just track your spices for a month. The initial inventory takes an hour. Updates take seconds. Most importantly, you need to update at the right moment - when you're putting groceries away, not when you're trying to cook dinner.

The tools matter less than the practice. You can use Cooklang's pantry.conf, a spreadsheet, or a notebook. The key is making it a database - structured, searchable, updated.

The return on investment is immediate. The first prevented duplicate purchase pays for the time. The first rescued leftover justifies the system. The first perfectly stocked shopping trip validates the approach.

That's not obsessive. That's engineering.