4) What Must Survive
The Design Brief

The easiest way to design the wrong calendar is to start by designing a calendar.

Choose twelve months or twenty-four. Decide whether weeks should run continuously. Pick an astronomical point for New Year. Find an elegant leap rule. Give everything names. By the time the structure looks satisfyingly complete, it is remarkably easy to discover that the system has been optimised for properties nobody actually asked for.

A better place to begin is with the brief. What kind of Mars are we designing for? What does its civil calendar need to accomplish? Which properties matter enough to protect when they conflict, and which familiar features are merely preferences? Until those questions are explicit, arguments about the correct number of months are mostly arguments about unstated assumptions.

Marian starts from a particular scenario rather than pretending to solve calendrics for every conceivable future Mars. It assumes an early permanent civilisation: no longer merely an expedition, increasingly complex in its own right, technologically dependent on coordination across many institutions, and still strongly connected to Earth.

A Mars that had been independently inhabited for five hundred years might choose differently. So might a small scientific outpost, a community with strong requirements around an uninterrupted weekly cycle, or a settlement determined to minimise every trace of terrestrial inheritance. That is not a weakness in the brief. It is what makes it a brief.

For Marian, the question is narrower:

What should a civil calendar do for a permanent, institutionally complex, Earth-connected human society on Mars?

The answer can be reduced to eight requirements.

These are design requirements, not astronomical laws. They define what Marian is trying to achieve before we decide how it achieves it.

The year must belong to Mars

The least negotiable requirement is that a Martian civil calendar should actually be Martian. One could continue using the Gregorian calendar indefinitely, treating Mars as something like an unusually inconvenient terrestrial time zone. One could also build a perfectly regular numerical system whose annual cycle slowly wandered through the Martian seasons. Either could tell people when a meeting was due, but neither would provide a genuinely Martian annual calendar.

The civil year therefore needs a reproducible relationship to the planet's own astronomical cycle. Someone centuries later should be able to reconstruct what the calendar means from its definition rather than relying on a chain of historical almanacs whose original assumptions have disappeared.

That does not mean every civil date must encode Mars's exact orbital position. Chapter 3 already separated astronomical reality from civil representation. Scientists, engineers and environmental systems can continue using Ls, ephemerides and other precise measures when exact seasonal position matters.

The important distinction is between anchoring and following. A calendar can be anchored to the Martian year without making every month correspond to an equal angle of orbital movement. The annual structure can remain physically grounded while its internal periods are designed primarily for civil use.

The requirement is not to make the calendar as astronomical as possible. It is to make sure the year remains Martian.

Regular enough to operate

Nature is under no obligation to produce convenient integers. The Martian annual cycle does not contain exactly 668 sols or exactly 669, so any practical calendar has to absorb a fraction somewhere. The design question is not whether irregularity exists, but where to put it.

Marian therefore gives substantial weight to structural regularity. Major civil periods should vary as little as practical, exceptional cases should be predictable, and unavoidable irregularity should appear in places users and software can understand in advance. Predictability matters more than mathematical purity.

Earth provides a useful example. February having 28 or 29 days is irregular, but the irregularity is familiar and deterministic. The Gregorian leap-year rule has some complexity, yet software can calculate it indefinitely and almost nobody is surprised when February ends. A calendar whose month lengths had to be recalculated every year from fresh astronomical observation might be beautifully faithful to the sky and unnecessarily tedious to run a civilisation with.

Dates will appear in contracts, maintenance plans, school schedules, financial systems, medical records, transport systems and software. Predictable structure reduces the number of exceptional cases all of those systems must carry. That does not mean every interval must be identical. It means differences should earn their keep.

Regularity also has a second dimension: recurrence. On the Gregorian calendar, a birthday on 12 May stays on 12 May while moving through the weekdays. Most civil life accepts this easily, while organisations that care strongly about weekly recurrence often create separate week-based schedules underneath the calendar.

A new Martian calendar need not inherit that trade-off automatically. It can reasonably ask whether recurring valid dates should retain predictable positions within the week. If a statutory deadline is fixed to one date, is there value in knowing its weekday without recalculating it every year? Could a monthly schedule be reused rather than regenerated?

Marian's brief gives real weight to that property because stable date–weekday relationships reduce variation. A date can carry more information, repeated schedules can remain structurally unchanged and annual templates become easier to reuse. The benefit is not free: fixed month starts and an uninterrupted seven-position weekly sequence can conflict once the arithmetic of the year is imposed.

At the requirements stage, however, the statement is deliberately narrower:

where practical, recurring dates should preserve stable weekly relationships.

The phrase where practical is doing useful work. Calendar design contains many perfectly sensible requirements that become ridiculous when the qualifier is removed.

More than one human timescale

The previous chapters established the most important thing Mars does to Earth's civil-time architecture: it separates functions that our own planetary ratios allow us to bundle together. The design brief should not put them straight back into one unit.

A permanent society will need useful intervals at several scales. There should be a short recurring rhythm comparable to the social role of the week, something around the scale humans commonly use for monthly planning, larger periods for medium-range coordination, and enough annual subdivision for people to understand where they are in a year lasting roughly ninety-five seven-sol weeks.

The design work therefore looks for useful scales around seven sols, roughly four weeks, roughly eight weeks, roughly sixteen weeks and meaningful fractions of the full Martian year. These numbers are not yet the Marian Calendar. They describe the jobs that need doing.

A technician may care about the next shift. A maintenance organisation may work on weekly or four-week cycles. A construction programme may need a longer review horizon. Governments and companies may need budgets, reporting periods and fractions of an annual cycle. There is no reason to force those activities into one unit merely because terrestrial language sometimes does.

This moves the brief decisively away from asking for the Martian equivalent of an Earth month. There may not be one. The better question is which useful temporal functions need homes.

The calendar must also remain legible at the annual scale. It is not enough to provide a collection of convenient short intervals while leaving people with little intuitive idea of where they stand within the year. On Earth, February feels early, June somewhere near the middle and October late. Q1 through Q4 provide another coarse map of annual progress. We use those positions without calculation.

A Martian year is almost twice as long, which makes that kind of orientation more useful rather than less. A perfectly efficient short-cycle operating calendar could still leave the annual cycle feeling like one enormous undifferentiated stretch. Conversely, a beautiful astronomical subdivision could tell users where they stood in the orbit while giving institutions awkwardly long gaps between useful operating markers.

The brief asks for both: people should be able to locate themselves intuitively within the large Martian annual cycle while institutions have useful intermediate periods inside it.

The requirement precedes the vocabulary. First decide what the tool needs to do. Name it later.

The machinery should be deterministic

A civil calendar can be grounded in astronomy without asking ordinary users to practise astronomy. Once the annual reference and epoch are established, routine date arithmetic should follow fixed rules. Software should be able to determine whether a year takes its ordinary or exceptional form. A court calculating a deadline ten Martian years ahead, a database validating a date and a historian reading an old record should all receive the same unambiguous answer.

That is what computational simplicity means in the design brief. It does not mean every part of the calendar must be mentally calculable by a five-year-old. Modern civilisation already relies happily on complicated machinery beneath simple interfaces. The useful test is whether the machinery is deterministic.

The sky should define the system. It should not have to be consulted every Tuesday.

This principle also matters for the relationship with Earth. A genuinely Martian calendar should not become an elaborate way of preserving terrestrial time, but an early Martian civilisation will not exist in isolation. People, organisations, spacecraft, communications, legal arrangements, finance and personal relationships will cross constantly between two civil worlds.

That creates a design requirement no terrestrial calendar has previously needed in quite the same way. The Martian system should avoid unnecessary ambiguity when its units interact with Earth ones. If a Martian month lasts substantially longer than an Earth month, its terminology and interfaces should not encourage people to assume the two are equivalent.

There is also a more interesting possibility. Earth and Mars move around the Sun at different rates, so their relative position and communications delay vary continuously. Precise operations will always rely on proper ephemerides and calculated light time; a civil calendar should not pretend to replace them. But ordinary people use coarse mental models all the time.

If a sensible calendar choice can make the broad Earth–Mars relationship somewhat easier to understand without weakening the calendar's primary Martian functions, the brief treats that as a genuine usability benefit. Not at any price, and not by turning a civil calendar into a navigation instrument, but Earth will remain too important during early settlement to be treated as somebody else's problem.

The structure should survive its names

The final requirement concerns something much more visible than orbital mechanics: what everything is called.

Calendar names attract attention because they are the part people encounter directly. Inventing them is also considerably more entertaining than defining machine identifiers, which can create the unfortunate impression that nomenclature is calendar design.

It is not.

A civil system intended to last should have stable structural identities beneath whatever language people use to discuss them. Computers, archives and different linguistic communities need to identify the same period even if they call it different things. A database should not need to know whether a future settlement uses an English, Mandarin, Arabic, Hindi or entirely new Martian name before it can determine which civil interval is meant.

The same principle matters culturally. Future residents of Mars may quite reasonably decide that some vocabulary proposed by twenty-first-century Earth designers is dreadful. They should be able to change it without changing the calendar.

The requirement is therefore stable structural identity beneath localisable language: canonical identities for the machinery, useful shorthand where appropriate, and human-facing names that can evolve. That is partly an internationalisation requirement. It is also a small act of humility.

Requirements collide

Taken separately, all eight requirements sound fairly harmless. Together, they become a calendar problem.

A continuously running seven-sol weekday sequence can conflict with perfectly stable month starts. Strong correspondence between civil periods and equal orbital angles conflicts with equal elapsed-time periods because Mars changes speed around the Sun. Preserving something close to the duration of an Earth month pushes the year towards roughly twenty-four such periods; preserving twelve annual months makes each one almost eight weeks long.

A perfectly identical set of civil periods cannot make the fractional Martian year disappear. Heavy optimisation for terrestrial familiarity risks producing a calendar that is insufficiently Martian, while optimising only for local mathematical elegance risks making the system unnecessarily unfamiliar to the first people expected to use it.

These are not bugs waiting for a clever enough designer to remove them. They are competing objectives.

That is why Martian calendars should not be compared merely by counting their oddities. One design may look more regular because it gives up a property another strongly protects. Another may track seasonal position more closely because it tolerates greater variation in elapsed periods. Another may preserve more familiar terminology because reducing learning cost is one of its explicit objectives.

The useful question is not which calendar has fewer strange features? It is which requirements was the calendar designed to satisfy, and where does it place the unavoidable cost?

Marian's requirements make its priorities explicit. They do not make those priorities universal. There is no experiment available today that can tell us exactly how a mature Martian society will want to organise its civil time. Astronomy can be measured, arithmetic checked, structural regularity quantified, Earth–Mars geometry modelled and software tested. Whether people would prefer an uninterrupted weekday sequence to fixed recurring dates, whether a new medium-range unit would become intuitive, or how much terrestrial familiarity helps rather than irritates are different kinds of questions.

Those are human-factors questions about a society that does not yet exist.

The Marian brief is therefore best understood as a transparent set of design judgements: the annual system should belong to Mars; irregularity should be predictable; recurring structures deserve weight; useful functions may need several temporal scales; the long Martian year should remain legible; ordinary date arithmetic should be deterministic; early Mars should remain intelligible in relation to Earth; and cultural vocabulary should be able to evolve without breaking the machinery.

Reasonable people can weight those requirements differently. That is precisely why the brief matters.

A calendar cannot avoid trade-offs. It can at least be explicit about which ones it intends to make.