Six inputs a workday calculator needs before it can answer

There is no calculator on this page. What there is instead is the list of things a calculator has to know before it can answer, because a tool that does not ask for them is not computing your answer. It is computing someone else's and showing it to you.

Six inputs. Miss any one and the result is confident and unverifiable.

What to take away

The six inputs

The two ends of the span, or a start and a count. These are different inputs. A start plus ten open days and a start and end date are not interchangeable, and a tool that accepts one and reports in terms of the other has made a conversion you did not see.

Which end counts. Counting between two dates has two defensible conventions, and they differ by one. Stepping forward by a count has the same ambiguity at zero. Neither is more correct; what matters is that the tool says which it used.

The rest-day set. Saturday and Sunday is a default, not a fact. Where the convention differs the answer moves by roughly one day a week, which over a month is four or five, larger than any closure list. The arithmetic is under how national calendars differ.

Every closure layer that applies. National, regional, local and organizational lists can all bind the same obligation, and they overlap. This is the input a calculator cannot supply, and the one most tools quietly default to empty.

The scope of each layer. A layer that binds courts does not necessarily bind banks. Applying one outside its scope gives a wrong answer with no symptom at all, which is why the registry of publishers matters more than the arithmetic. Who publishes what is under state and regional calendars.

The as-of date of each list. A count computed against a list retrieved eight months ago is not wrong arithmetic. It is a stale input, and without the retrieval date nobody can tell the two apart.

What each missing input does

The two rows marked "no" are the reason this page exists. An arithmetic error usually announces itself somewhere. A missing or misapplied layer does not, because every intermediate value is valid.

What an honest tool shows

The number, and then the working. For a count, the list of days it skipped and why each was skipped. For a step, the dates it passed over. That list is short, it is checkable by eye against a published calendar, and it turns a dispute into a two-minute comparison.

It should also echo the inputs it used, including the ones you did not supply. A tool that silently applied a Saturday and Sunday weekend should say so in the output, so that a reader in a place with a different convention sees the assumption rather than inheriting it.

Doing it without a tool

For a span, count the weekdays first, which is arithmetic with one right answer, then subtract the closures that fall on weekdays. The weekday part is under counting a month for a month and counting a year for a year, and both are short enough to check by hand.

For a step, walk the days. Ten open days is at most fourteen calendar days to check, and walking them is faster than establishing whether a tool used your closure list. The mechanics are under adding business days.

Either way, write down the closure list you used and where it came from. The calculation is reproducible. The inputs are the part that goes missing.

Why the layering is the hard part

Layers overlap, so count them as a set union, not a sum of counts. Two lists naming the same day close one day, not two. That is ordinary idempotence. A tool that adds list lengths instead of uniting date sets will over-deduct on every overlap while looking reasonable.

The state calendars formula keeps each closure layer separate before the union, so an overlap is never double-counted.

Whether a given day is open at all is also not a purely arithmetic question. The business day is a legal and commercial category, defined differently for different obligations, which is exactly why the scope input cannot be inferred from a country or a postcode.

Common questions

Why does this page not just include a calculator?

Because a calculator that does not know your closure layers gives an answer that is wrong for most readers, and one that does know them is really a form asking for six inputs. The inputs are the content.

Is a general-purpose date calculator ever safe?

For weekday counts with no closures, yes, and that is a genuinely useful thing to compute. Treat any working-day figure it gives as a draft until you have supplied the layers yourself.

What is the minimum I should record with a computed deadline?

The two ends, the convention, the rest-day set, the layers with their retrieval dates, and the skipped days. That is enough for someone else to reproduce the number.

Who decides which layers apply?

Whoever set the obligation, usually in the document that created it. Where the document is silent, that silence is a question to raise rather than a gap for a calculator to fill.