Behind the builder / 28 September 2026
One hundred puzzles. Still a few things between a manuscript and a book.
Downbeat & Across reached a complete hundred-puzzle manuscript. Print layout, checking and KDP economics revealed the next gates between creation and publication.
The manuscript was assembled
Downbeat & Across Volume One had reached an assembled manuscript containing one hundred standard puzzles and a thirty-by-thirty bonus crossword, The Long Play. The draft ran to 368 pages. It was a tangible piece of work: something to open, review and prepare for a reader.
Completing the assembly changed the question. I no longer needed to ask whether enough puzzles existed to make the volume. I needed to ask whether the volume worked as a book. That includes the clues, the grids, the answers, the navigation and the experience of solving without me nearby to explain a mistake.
A finished draft is a milestone with a useful next stage attached to it.
The facing pages were part of the product
The intended layout put the puzzle grid on the left-hand page and all the clues on the facing right-hand page. It is an ordinary requirement that becomes less ordinary with a thirty-by-thirty grid. Readability, space and page pairing all have to agree.
The book still needed final layout, fact-checking and a blind solve before publication. Those are not cosmetic tasks. A clue that works only because the author knows the intended answer has not passed the reader’s test. A generous grid does not help if the clue text becomes uncomfortable to read.
This was a reminder that generating content and designing a usable artifact are connected but separate jobs.
The economics had to remain a model
A separate workbook examined KDP print costs, royalties, pricing, break-even points and sales-volume sensitivities. That made the commercial questions inspectable. It did not make the volume assumptions into a forecast of customer demand.
The distinction matters when a spreadsheet starts looking persuasive. Changing the units sold can produce an attractive monthly outcome in seconds. Establishing that readers want the book requires publication, discovery, purchase and feedback.
The model was useful for making choices under different scenarios. The next market evidence would have to come from outside the model.
A second product was teaching the same lesson
Cantera’s P016 remained installed and was verified within its bounded QA scope. The recorded campaign contained 203 tests: 202 passed and one was skipped. Backup coverage had improved to 95.6% of inventoried source bytes, with a remaining gap explicitly recorded.
The book and the software had very different materials. Both needed the same discipline around readiness. A manuscript needed reader-facing checks. An installed application needed the remaining audio and public-release gates. Neither became commercially validated because an internal artifact looked complete.
The useful move was to freeze a release checklist and put the next test where it could change the outcome. Today produced a book-shaped milestone. The reader-shaped test was still ahead.