As that source states it: after each 12-month qualifying period an employee in Brazil is entitled to paid vacation plus a one-third bonus (Constitution art. 7, XVII; CLT art. 134). The length depends on unjustified absences (CLT art. 130): up to 5 absences gives 30 calendar days, 6 to 14 gives 24, 15 to 23 gives 18, 24 to 32 gives 12, and above 32 gives none. Since Lei 13.467/2017 the vacation may be split into at most three periods, if the employee agrees; one period must be at least 14 calendar days and each of the others at least 5. The bonus is extra pay, not extra days, so it does not enter the counts.
Count ordered splits, since the order of the periods in the year matters to when a person travels. A vacation taken whole counts as one way and is not subject to the 14-day minimum, which applies when splitting.
The three-period count has a closed form. With each part at least 5, subtract 5 from each and distribute 15 over three parts: 136 ways. Those with every part at most 13 (so no part reaches 14) number 52 by inclusion and exclusion, which leaves 136 − 52 = 84. The script confirms 84 by listing every split.
| entitlement (days) | unsplit | two periods | three periods | total |
|---|---|---|---|---|
| 30 | 1 | 21 | 84 | 106 |
| 24 | 1 | 12 | 3 (14+5+5 in some order) | 16 |
| 18 | 1 | 0 | 0 | 1 |
| 12 | 1 | 0 | 0 | 1 |
So the rule leaves a worker with 30 days 106 legal ways to take it, and a worker who has missed enough days to fall to 18 or 12 only one: all in a single block.
A covered worker with a 30-day entitlement produces 30 person-days of vacation a year. If the vacations of covered workers were spread evenly over a 365-day year, then on any one day 30/365 = 0.0822 of them would be on vacation, about one in twelve. If instead every vacation fell inside one 30-day month, the share on vacation during that month would be 1, a factor of 365/30 = 12.17 over the even spread. That is an extreme case chosen for the arithmetic, not a forecast.
The law fixes the total and the shape of the pieces; it does not fix the dates. The dates decide the peak, and the peak is the demand d that the network of chapter 1 has to carry. This is where the book’s models meet: the count of covered workers and the dates they choose are inputs from outside the chapter.
python3 book15/ch03-verify.py uses only the standard library, prints every number on this page and exits non-zero on any failure. The Lean file for these counts is not written.