Name rhythm in an authored verse

Leave room for the name before polishing the line

A character’s name can fit on the page and still feel crowded in a sung promise, a chant or an invented ceremonial verse. The problem is sometimes a shortage of positions. Sometimes the positions exist, but the author wants a prominent unit where the supplied reading has an unmarked one. Shortening the name may solve the arithmetic while changing the relationship: a public chorus uses an intimate form that the character has never agreed to share.

This workbench compares explicit name readings inside an authored grid. You supply each unit, its binary prominence mark and the surrounding words. You also declare which positions must be occupied, which marks are strict, and whether a small number of surrounding units may sit outside the grid. The program then finds permitted ordered alignments and shows the lowest-cost witnesses under its published editorial costs.

It does not pronounce names, identify Japanese morae, infer accent or certify a real verse tradition. A unit is whatever you explicitly supplied for this worksheet. A mark of one means prominence in your model; zero means its alternative, and null means you do not know. These choices make a limited calculation useful without disguising it as linguistic authority. Use an actual reading source or a performer’s agreed notes when the project needs one, then enter those observations accurately.

Original fictional scene: the chorus changes its promise

The community hall’s closing song had a space for the person who would reopen its garden. On the rehearsal sheet, the words were “we — Iria — stay.” The fictional chorus had agreed to a five-position pattern: zero, one, zero, zero, one. Iria’s authored reading occupied three units, I / ri / a, marked one, zero, zero. With we marked zero and stay marked one, every position had a place.

The arranger brought a second sheet with Iri in the middle. It was shorter, easier to fit into the narrow printed heading and familiar to a few friends. In the five required positions, however, its two supplied name units left a gap. The arranger proposed stretching the last unit without writing that alteration down. Iria asked why the group had changed the name before asking whether the song could change.

Nobody needed to accuse the arranger of cruelty for the scene to matter. They had optimized a visible heading and assumed the sound would follow. The rehearsal exposed an unspoken decision about who could use the short form. Iria chose the full form for the public song, while a friend kept Iri in a quiet exchange after the room emptied. The difference belonged to the relationship, not to a universal rule about affectionate names.

The group also tried making the fourth position optional. That admitted the shorter reading, with an explicitly unused position. It did not create an extra syllable or show that either reading was more authentic. The worksheet made the alternative legible enough for the group to discuss it. They retained the original five occupied positions because the final stay now arrived where their fictional promise required it.

This is an original fictional example. The hall, garden, people, reading units and prominence convention were created for this guide. No established meaning or actual pronunciation of Iria or Iri is asserted. The emotional decision is the author’s interpretation of the scene, not a conclusion produced by the alignment count.

Try the counterfactual: if the character had explicitly chosen Iri for the public event, preserving the longer form could become the exclusion rather than the kindness. You would revise the grid or its surrounding words to accommodate that choice. A tool that merely favored the form with more units would miss the reversal. The task is to expose the production tradeoff so that the story can make its own decision.

Declare readings instead of asking the spelling to confess

Each pronunciation variant needs an ID, a display name and a nonempty units array. The worksheet accepts one to twelve variants, with one to twenty-four units per name. Each unit has exactly text and stress. The text is a literal label of at most forty UTF-16 units; stress is zero, one or null. The display name can contain at most one hundred units, and the ID must be unique within the comparison.

Different IDs can use the same display name. This is useful when a project has two explicitly supplied performances of the same written form. Preserve the difference in their unit arrays instead of adding a misleading spelling change. The comparison does not decide which performance is correct. It reports the consequences of the alternatives you actually entered.

The before and after arrays each accept up to sixteen units using the same text/stress contract. They represent the surrounding line, not disposable padding. Enter enough context to preserve the emphasis that matters to the scene. A name can align neatly while pushing an important final word into an unsuitable position. The witness marks every unit as before, name or after so that this displacement remains visible.

Null is an honest information state. When a position requests zero or one and the corresponding unit is unknown, the alignment remains possible but receives an unknown finding and the disclosed cost. The program does not copy the requested mark into the reading. When a position itself has null stress, it makes no prominence request. That is freedom in the authored grid, rather than newly discovered knowledge about the name.

Separate occupancy from prominence

The pattern contains one to sixty slots. Every slot has exactly stress, required and hard. Stress is zero, one or null. Required determines whether the slot must contain a unit. Hard determines whether a known difference from a requested mark is forbidden or permitted as a soft mismatch. These switches answer different questions and should not be collapsed into one vague strictness setting.

A required slot with null stress must still be occupied. It accepts either supplied prominence mark because you have made no request about that property. An optional slot with hard stress may remain empty, but if it is occupied by a known mark, that mark must match. Making a position optional does not automatically make a conflicting occupied version acceptable.

All placed units preserve their order. The program does not rearrange a name or swap the words around it to obtain a better result. It also does not stretch, split, merge or repeat a unit. If the production allows a special sung realization, author it as another explicit variant and explain its status in the scene notes. Hidden stretching would make the result impossible to reproduce from the declared inputs.

The default Iria form fills five required positions exactly. The shorter Iri form has four total units once we and stay are included, so it has no permitted alignment in that pattern. Make the fourth slot optional and the shorter form has one permitted alignment with one skipped position. This is a concrete change in occupancy behavior, not a random replacement of the displayed name.

Licence a boundary without cutting the person out

Pickup and ending controls each allow zero to two surrounding units outside the grid. A pickup can take only a prefix of the before array. An ending can take only a suffix of the after array. Neither control can remove a unit from the chosen name. The outside material is retained in the witness and private export, so a licence is not confused with deletion.

For a minimal example, enter a name with one marked unit and a single required marked position. Put an unmarked we before it. With no pickup permitted, there are too many units. Allow one pickup and we can sit outside the grid while the name occupies its required position. The witness shows precisely which unit moved and charges the disclosed boundary cost.

This is an authored arrangement convention, not a claim about how every musical pickup works. The software models only the finite before/after licences you declared. If the scene requires a held note, an internal pause or another operation, it is outside this contract. Write a separate explicit reading or change the surrounding text; do not describe an unsupported transformation as something the engine verified.

Boundary choices can create several distinct permitted alignments. The total counts them separately because their placement differs, even when the printed line stays identical. Examine their traces before deciding whether those differences matter to your performers or to the story. An exact count is useful for scope; it is not a prediction of how many alternatives a listener will notice.

Read a minimum cost as a disclosed compromise

The heuristic costs are fixed and public: a known mismatch in a soft position costs three; an unknown unit at a position requesting a mark costs two; an unused optional slot costs one; and each licensed external pickup or ending unit costs one. Hard known mismatches are forbidden. The calculation finds the minimum total among permitted alignments and counts every alignment achieving it.

These numbers express an editing order chosen for this workbench. They are not a quality score for a name, a musical judgment or an empirical estimate of listener effort. A cost of two caused by uncertainty is different from two optional gaps. The witness preserves those causes, and an author may prefer a higher-cost alternative for a clear dramatic reason.

The result reports both all permitted alignments and minimum-cost alignments. The display cap, from one to twelve per variant, limits witnesses of the minimum only. It does not limit the exact counts. A large count with one visible example means that alternatives remain unshown; it does not mean that the program found a unique way to perform the line.

For a useful stress test, twenty-four ordered units can be placed into sixty optional unconstrained slots in 36,052,387,482,172,425 ways. Every arrangement leaves thirty-six slots unused and has the same cost in that particular fixture. Showing one trace cannot summarize their artistic implications. The exact decimal count exists to prevent a sampled display from becoming a false uniqueness claim.

Diagnose failure before replacing a name

A variant with no permitted alignment remains in the feasibility table. First compare its total unit count with the number of required and available slots. Then inspect whether the permitted boundary licences could address the difference. Length and hard-mark conflicts are separate causes, and neither proves that the name is unsuitable for the character.

For some failures, a diagnostic alignment temporarily relaxes hard marks while retaining the original boundaries. It can expose an actual conflicting unit and position. This is explicitly an invalid diagnostic witness, not a repaired result. It is not an exhaustive list of failures or a minimum-edit recommendation. If no such diagnostic exists, inspect shape and boundary choices rather than inventing a phonetic explanation.

When correcting a conflict, change the thing you actually intend to change: an authored reading, the line, a requested prominence, or an occupancy licence. Recalculate and read the full surrounding phrase. Replacing only the display name while leaving its units untouched changes the label, not the performed structure. Preserve that distinction in any production handoff.

Carry the decision into rehearsal

Copy, favorites, comparison and the private audit retain source readings, marks, alignments and scene context. A saved project restores editable parameters. These are working documents that may disclose character choices or later promises, so review them before sending them to collaborators or readers. The audit establishes what was supplied and computed; it does not replace a performer’s agreement.

Use the scene field to record why one arrangement matters. “Keep the full public name; leave stay prominent” gives a later editor something concrete to preserve. “The first option scored best” loses the relationship and the reason for the line. After selecting a version, rehearse it in the medium that will carry it and record any revised reading explicitly before recalculating.

Does this count Japanese morae or infer a name’s accent?

No. The units and zero/one/null prominence marks are supplied by you. The calculation checks your authored grid, not the natural-language correctness of those inputs.

Can a pickup licence shorten the chosen name?

No. Only a prefix of the before array or a suffix of the after array may move outside the grid. All name units remain ordered inside it, and every moved surrounding unit remains in the witness.

Why are all alignments and minimum alignments different numbers?

Some permitted placements include more soft mismatches, unknown requested marks or boundary changes. All are counted; only the lowest disclosed heuristic cost is sampled. The display cap never changes either exact count.

Does the relaxed conflict table repair a failed name?

No. It temporarily permits hard mismatches at unchanged boundaries to expose one diagnostic conflict. That trace is invalid under your real constraints and is not an exhaustive or minimum-edit repair.