Selling aquarium puzzle books as downloads: the tanks your buyer fills level by level, trusting every count is true

An aquarium book punishes a bad file in a way a sample puzzle never shows. Your buyer judges puzzle forty-one, when the water they have poured into a bent tank stands at two different heights in its two arms, or a row that should hold exactly five filled cells holds six because one tank's level was keyed one row too high, and the solver who trusted the rules concludes the fault is theirs. Aquarium looks gentle: a rectangular grid divided by bold borders into irregular tanks, a number outside the grid at the end of every row and column, and one instruction — fill the tanks with water. Underneath it is a contract in five parts, each part exact. First, the tank rule: the grid is divided by bold lines into tanks, and each tank is filled to exactly one water level — one flat surface for the whole tank, however wide or bent it is. Second, the gravity rule: inside a tank, if a cell holds water, every cell in that same tank at the same height or below it holds water too, and if a cell is empty, every cell above it in that same tank is empty — water never floats, and it never stands uneven. Third, the level rule in full: that single level applies across the tank's full width, so pouring into one arm of a bent tank raises the far arm to the same height, cell for cell. Fourth, the count rule: the number outside each row and each column is exactly the number of filled cells in that row or column, not nearly, not at most. Fifth, tying it shut: exactly one assignment of tank levels, reachable by deduction alone, with no guessing. A puzzle that breaks any part is not slightly harder. It is false as stated. A good aquarium book is a verified board: every puzzle re-checked tank by tank and row by row from its own key, puzzle by puzzle.

Two warnings belong at the front, because they are the mistakes this genre makes most often, and both start with puzzles aquarium is not. Aquarium is not a nonogram: the outside numbers here are single totals per row and column, not run-length clues, and the cells are not filled one by one — they are filled in whole horizontal layers inside each tank, so one decision about a level fills or empties many cells at once. It is also not thermometers, the filling puzzle buyers reach for next: a thermometer is one narrow shape filled from its bulb, while an aquarium tank can be wide, tall, or bent around a corner, and its level couples cells in distant rows and columns that share nothing except the tank they belong to. And the gravity rule needs saying precisely, because sloppy rules pages ruin this puzzle in one specific way: the level is per tank, not per column and not per connected patch of blue in the picture. A rules page that says "fill some cells in each tank" or forgets that an empty cell forces everything above it in that tank to be empty teaches a weaker puzzle with many more solutions, and a setter who generates under the weaker reading ships boards whose keys are one answer among several. The second warning is about bent tanks, the clause short summaries most often blur. Two patches of water at the same height may or may not be the same tank; the bold borders are the truth, and the colouring is only the consequence. Print the gravity rule, in full, on the rules page, and say what one level means in words a first-time buyer cannot misread: choose a height for each tank, fill everything in that tank at or below that height, and leave everything above it dry.

The book is the whole board, not the sample puzzle

Buyers do not buy a border style. They buy the promise that every puzzle in the file obeys the rules printed at the front, at the difficulty the cover claims, in the order the book presents them. A free sample almost never breaks that promise, because a sample is the one puzzle the setter checked three times. The breakage lives at puzzle forty-one, in the middle of the book, where a generator exported the tanks from one board and the key from the next, or a hand-setter moved one bold border to fix a corner, quietly changed which tank a column of cells belongs to, and never re-ran a single check. That is why the unit of quality in a puzzle book is the whole set, verified as a set: the same checks, run the same mechanical way, on every single puzzle, with the results kept. If you generated the puzzles, keep the generator's settings per puzzle, including its grid size, tank map, and the level chosen for every tank in the key. If you set them by hand, keep your solving notes. When a buyer writes in about one puzzle — and in this genre they write in with the page number and the tank whose level they cannot make agree with both its row and its column counts — you want to answer about that puzzle, not go hunting through a file you can no longer reconstruct.

The rules page is part of the product, not decoration. State all five parts of the contract in plain language, with one small worked example showing a short board being solved layer by layer: one tank whose level is forced because a tight row count leaves it only one honest height, one tall tank where a column count rules out its top two levels at once, one bent tank where pouring into the near arm visibly raises the far arm, and the board finishing with every row and column counted aloud against its outside number. That worked example is the one idea a new buyer must carry away, because aquarium is solved against tanks and counts together, not against single cells, and it is the idea a short rules page most often drops in favour of a pretty picture of blue tanks with no level ever justified. Most one-star reviews in this genre are not "too hard" or "too easy". They are "the finished board breaks the rules" — and that mismatch is almost always a setter who held the real board in their head, or in the generator, and printed tanks that were never re-checked against their own key.

The contract, part by part

The tank rule is the frame everything else hangs on, so check it like one. Every cell belongs to exactly one tank, the tanks together cover the whole grid with no gaps and no overlaps, and each tank in the key carries exactly one level: a single height, stated or countable, from which the tank's entire filling follows. Audit it mechanically: for each tank, name its level, then confirm that the filled cells of that tank are precisely the cells of that tank at or below that level — no filled cell above the line, no dry cell below it. A key in which one tank holds water at two heights fails here however well its row counts add up, and it is a failure a surprising number of generators produce when a border edit splits one tank into two without anyone updating the key.

The gravity rule is the heart of the puzzle. Inside any tank, water cannot float: a filled cell with a dry cell directly or indirectly below it in the same tank is impossible, and an empty cell with water above it in the same tank is equally impossible. This check is purely structural — it can be run on the key cell by cell without solving anything — and it catches the classic transcription error where a tank drawn one cell too short leaves a filled cell stranded above a gap that the setter's eye slides past because the blue still looks like a plausible pool. Check it in both directions, filled-implies-below and empty-implies-above, because each direction catches errors the other can miss on bent tanks.

The level rule is where wide and bent tanks earn their keep, and where keys most quietly go wrong. A tank's level is flat across its full width: if the tank spreads across five columns at its base and narrows to one at its top, the water surface is still one horizontal line, and every arm, pocket, and step of the tank obeys that same line. For each tank, draw its level line across the whole tank map and confirm no part of the tank holds water above that line in one arm while sitting below it in another. This is the check that separates aquarium from a per-column filling puzzle, and a key built column by column without it can look perfect in every column while breaking the puzzle's actual logic in the tank that bends.

The count rule is exact arithmetic, row by row and column by column. For each row, count the filled cells in the key: the count must equal the printed number outside that row, exactly. For each column, the same: the count must equal the printed number outside that column, exactly. Record the count for every row and every column, one line per line, no sampling. A single row off by one does not make a slightly different puzzle; it usually makes a puzzle with no solution, or with a different one, while the printed page still promises the original. Column counts deserve the same care as row counts — setters who audit only rows ship books whose column errors are found by buyers within minutes, because buyers check the line they are stuck on, which is as often a column as a row.

Uniqueness is the part of the contract a plausible board cannot promise. A set of tanks and outside numbers can usually admit more than one assignment of levels that respects gravity and the counts if the setter only checks that an assignment exists — large tanks are where second solutions hide, because a big tank can often sit one level higher while a neighbour compensates a row count and no single line betrays the swap — and a puzzle that solves two ways tells one of its two correct solvers they are wrong. The fifth part therefore has to be proven, not assumed: an independent solve, or a solver run, that counts assignments and returns exactly one, with that one reachable by deduction rather than by trying a tank at two levels until one branch survives.

Verification: count every tank, then count every line

Verification is a separate pass from solving, done mechanically, ideally by a different method than the one that made the puzzle. Solve each puzzle once as a buyer would, by deduction only, and keep the solve honest: if you guessed a tank's level, and the guess worked, that puzzle has not been shown to have one solution, and it goes back. Then run the structural audit on the key, in this order. One: inventory audit — every tank border and every printed outside number on the puzzle page appears on the key page, in the same position, unchanged. Two: tank audit — every cell belongs to exactly one tank, and the tanks together cover the grid with no gaps and no overlaps. Three: level audit — every tank has exactly one flat level across its full width, arms and pockets included. Four: gravity audit — no filled cell has a dry cell below it in its tank, and no empty cell has water above it in its tank. Five: count audit — every row's filled-cell count equals its outside number exactly, and every column's filled-cell count equals its outside number exactly. Six: givens audit — no outside number was moved, added, or dropped between the puzzle page and the key page, and no border shifted by a cell. Seven: uniqueness audit — a second, independent solve, or a solver run, that counts level assignments and returns exactly one. Eight: consistency — puzzle page, key page and any thumbnail or marketing sample of that puzzle are the same puzzle, with the same tank map, the same outside numbers, and the same levels in the key.

Keep the audit results with the book's files, puzzle by puzzle. When a corrected edition ships, that record is how you know which puzzles actually changed and which were only suspected. It is also how you answer the buyer who found a real error: not with an apology and a shrug, but with the corrected puzzle, the audit lines that now pass, and the corrected file sent to everyone who bought the book — which is where the delivery side of this business, covered below, stops being administration and becomes the difference between a one-star review and a loyal repeat buyer.

Difficulty is tank shape and level pressure, not just grid size

Grid size is the weakest difficulty lever in aquarium, and the one covers lean on hardest. A large grid whose tanks are narrow and whose counts are tight, with levels forced tank by tank, falls layer by layer in a satisfying sweep, while a small grid with one huge bent tank and counts placed to tease can stop an experienced solver cold. The real levers are these. Level pressure: how many tanks have their level pinned early by a tight row or column count — a tank that can only sit at one height hands the solver its whole filling at once, while a tank with four honest heights defers many cells at once, and a book can be graded by how much of the board the counts pin down before any deduction starts. Tank shape: tall narrow tanks couple many rows at once and force early decisions, wide shallow tanks couple many columns, and bent tanks couple distant parts of the board that look unrelated on the page; a setter who mixes shapes without grading them misjudges every board. Count tightness: rows and columns whose numbers sit near their floors or ceilings — near empty or near full — bite early, while middling counts stay silent until levels elsewhere are known, and that silence is this puzzle's signature late difficulty. Compensation room: pairs of tanks that can trade a filled cell between rows without breaking a count are where second-guessing lives; more compensation room, placed further apart, makes a harder board at any size.

Grade a book by testing solves, not by measuring grids. Time honest solves across the set, note where each puzzle's deductions stall and restart, and order the book so difficulty genuinely climbs — early puzzles with many tight counts, narrow tanks, and levels that hand the solver the board's skeleton, later puzzles with sparse tight counts, wide bent tanks, and compensation pairs where level logic does the heavy lifting near the end. Print the grading honestly: "easy" should mean the counts mostly pour the tanks for you, not merely that the grid is small. A buyer who trusts your grading buys the next volume; a buyer who was ambushed by a bent tank on page three of the easy book leaves a review about that instead.

Print honesty: the page is the interface

An aquarium page is read in pencil first and pen last, and the print has to survive both — starting with its borders, because bold tank borders against thin cell lines are this genre's whole interface. A tank border must read as unmistakably bold after a photocopy, at arm's length, in greyscale: if a border prints at nearly the same weight as a cell line, the buyer is solving a different tank map from the one the setter checked, and every count in the book becomes an argument. Outside numbers must align unambiguously with their rows and columns, sized to be read without zoom and placed so no number can be mistaken for belonging to the line next to it — a number printed halfway between two rows is a misprint waiting for its first careful buyer. The water a buyer shades needs room too: cells must be large enough for a pencilled fill to read as filled at a glance, with space for the small dry marks solvers make above a level before committing. Your key must print the finished tanks at the same size relation to the puzzle page, with the tank borders and level lines still visible over or under the shading, so a buyer can count any row's filled cells in the key for themselves instead of taking the picture on faith — in this genre the key that hides its borders cannot be checked, and an uncheckable key reads as an untrusted one.

Proof the final PDF, not the working file. Export at the exact trim size, print one puzzle and its key on a home printer at 100 percent, and solve the printed puzzle with the printed rules page as your only reference. Check the three things screens hide: whether tank borders still read as borders, not cell lines, after the page has been photocopied once — classrooms and care settings will photocopy it — whether every outside number is unambiguously aligned with its row or column on that same copy, and whether there is room in each cell for a pencilled fill plus dry marks. If a puzzle needs a screen's zoom to tell where one tank ends and the next begins, it is not ready to be a download.

Licensing, provenance, and the download itself

State the licence where the buyer decides, not buried after purchase: who may solve the puzzles (one person, a household, a classroom, a club, a care setting), how many copies they may print, whether a teacher may hand a page to thirty students, whether a club may run one puzzle as a group activity, and whether the file itself may be shared, resold or republished — and say plainly that sharing the file is not the same as sharing a printed page. Puzzle books are bought heavily by teachers, club organisers and carers, and those buyers are trying to comply; a licence they cannot find reads to them as a licence that says no. Say how the book was made, too: hand-set, software-generated and hand-checked, or software-generated and mechanically audited — and state your conventions in the book and in the listing, that every tank takes exactly one flat level and that water never floats, because a buyer raised on a looser filling puzzle will otherwise judge your correct boards as wrong ones. Provenance is not an apology — in this genre it is a quality claim, as long as it is specific. "Generated, then every tank levelled and every row and column counted from its own key" is a claim a buyer can weigh; "high quality puzzles" is not.

Delivery is the last part of the contract. The buyer should get the file immediately, from a link that works the first time, on the device they bought it on, with re-downloads available when — not if — they lose it. Send a corrected edition to every past buyer when one ships; do not wait for them to discover the errata the hard way. I run my own download delivery through FileFreight — flat monthly pricing instead of a cut of every sale, expiring links, and updated files reaching past buyers without me re-emailing attachments by hand — but whatever you use, test it as a buyer: pay, receive, download, lose the file, recover it, and receive an update, before you ask strangers to trust it with money.

The launch checklist

1. Every puzzle solved by deduction alone — no guessed tank level in any solve, start to finish.
2. Tank audit passed: every cell in exactly one tank, tanks covering the grid with no gaps and no overlaps.
3. Level and gravity audits passed: every tank at exactly one flat level across its full width, no floating water and no water above an empty cell in any tank.
4. Count audit passed: every row's and every column's filled-cell count equals its printed outside number exactly.
5. Uniqueness audit passed: an independent solve or solver run returns exactly one assignment of levels for every puzzle.
6. Print proof passed: tank borders unmistakable from cell lines after one photocopy, every outside number unambiguously aligned, on a home-printed page at final size.
7. The errata rule honoured in advance: any second solution, any uneven level, any floating cell or any row or column count off by one means a corrected file goes to every buyer, not just the one who noticed.