A name is an address. That is the whole idea behind a naming convention, and most farms never state it that plainly, which is why so many of them end up with names that work fine for the person who invented them and fail for absolutely everyone else. A record is only as useful as the ability of a stranger to find the exact thing it refers to, and a name is the mechanism that does that finding.
This is why "the north field" is not a small stylistic choice — it is a naming convention that has quietly decided its only valid reader is someone who already knows the farm.
A name has to work for a reader who has never been here
"North field" is not wrong, exactly — it means something, and the person who wrote it knows precisely what. The problem is who else has to read it later: a new hire who has not learned the farm's informal geography yet, a buyer's auditor doing a traceability check from an office two states away, an investigator working backward from a residue result with nothing but your records and a map they do not have. None of them can be expected to know which direction is north relative to a barn none of them have stood next to. The name fails the exact reader a record most needs to serve — the one with no prior context at all.
It is worth being specific about who "a stranger" actually is, because the phrase can sound abstract until you name the real people it refers to. It is the seasonal worker hired three years from now who has never heard the farm's informal names for anything. It is the accountant reconciling records at a distance, who has never set foot on the property. It is the family member who inherits the operation and has to make sense of a filing system built by someone else's memory. Every one of these readers exists eventually, on almost every farm, and a naming convention is really a decision about how much of the burden of understanding your records falls on you now versus on them later.
Good names are stable; descriptive names usually are not
The instinct behind "north field" or "the pumpkin patch" is reasonable — name the thing by what it obviously is right now. The flaw is that what a field obviously is changes: this year's pumpkin patch is next year's cover crop, and a name built on this season's crop stops matching reality the moment the rotation turns over. A genuinely durable name is tied to something that does not change — a parcel number, a fixed position in a permanent grid, an ID assigned once and never repurposed. It will feel less immediately meaningful than "the pumpkin patch." That is the correct trade. A name's job is to keep pointing at the same thing for years, not to describe it accurately for one season.
None of this means informal names have to disappear from the farm entirely — "the pumpkin patch" can stay in conversation as long as the record itself carries the durable code alongside it, the way a street address and a nickname for a house can both be true at once. The failure is not using an informal name. It is a record that has only the informal name and nothing else to fall back on once the season, and the memory that made the name make sense, has passed.
Uniqueness matters more than memorability
A second, quieter failure mode is two different things sharing a name — two "back forties" on a farm with more than one back field, two lots both labeled with this week's date because two harvests happened to land on the same day. A name only functions as an address if it resolves to exactly one thing. When it does not, any record that references it becomes ambiguous retroactively, and there is often no way, months later, to tell which of the two it actually meant.
The fix is a convention with enough structure to guarantee uniqueness on its own — a parcel code plus a sequence number, say — rather than relying on everyone remembering, informally, not to reuse a name that feels available.
It is worth writing the convention down somewhere every person entering records can find it — not because the rule itself is complicated, but because a convention nobody can look up drifts the moment the person who invented it is unavailable for a season. A single page describing how a field code is built, how a lot number is generated, and what to do when a new field is added costs almost nothing to write and saves the exact confusion a naming convention exists to prevent in the first place.
A consistent naming convention is also what makes a multi-enterprise split workable at all — a farm running several businesses off shared land needs its field and lot codes to resolve unambiguously to which enterprise they belong to, and a name built on an informal description rarely carries that distinction cleanly.
A hand-typed code is one keystroke from a broken link
Naming becomes a recordkeeping problem, not just a filing problem, the moment names are used to join records to each other — a harvest lot matched to the field it came from, a shipment matched to the lot it contained. That join depends entirely on the two records using the identical string. A code that gets retyped by hand on the second record, rather than copied from the first, is exposed to a single mistyped character breaking the connection with no visible error — the two records simply fail to match, and nothing in most systems flags that failure as a failure. The practical rule is to generate identifying codes once and copy them everywhere they are needed, never retype them from memory.
This is exactly the seam where a traceability packet breaks
Farm40's traceability packet works by joining a harvest lot to the inputs applied to it, through the shared planting record, and forward to the customers it was shipped to — and every one of those joins is a string match on the lot code you entered. The honest limit is direct: a blank or mistyped lot code does not produce an error. It just silently fails to link, and the packet quietly comes up short without telling you which join broke. A disciplined naming convention — codes generated once, copied rather than retyped, unique by construction — is not a nicety in this context. It is the difference between a traceability packet that actually joins and one that looks complete and is not.
A naming convention is one of the cheapest investments in recordkeeping a farm can make, and one of the easiest to keep putting off, because a bad name does not announce itself as a problem until the day someone downstream needs to follow it. Fix it before that day, not after. For where naming sits inside the rest of the system, see farm recordkeeping and, for the specific case of joining records across a supply chain, traceability.