Arithmetic, Arithmetical, Monotonic Year

[sffc]

Let's settle on a name for our integer-year-without-era concept. It is spec-internal, but the name sets precedent that libraries like ICU4X will follow, so it is still important to consider.

What we have on the table:

Name	Pros	Cons
Arithmetic Year	Status quo. Indicates what this type of year should be used for.	Different pronunciations are confusing. "Arithmetic" is more often a noun than an adjective.
Arithmetical Year	Descriptive. Grammatically unambiguous.	2 extra characters than Arithmetic.
Monotonic Year	Describes how it behaves over what it does.	People might not know what "monotonic" means.

Opinions? I'm adding this to an upcoming Temporal Champions call.

@ptomato @justingrant @Manishearth @robertbastian @nekevss

[Manishearth]

I like Arithmetic or Arithmetical. I don't particularly find "noun vs adjective" to be compelling; most such nouns can be used as noun adjuncts.

I find monotonic to be a specific property amongst many, so it feels strange to highlight it.

[ben-allen]

In the 20^th and 21^st centuries the adjective form of "arithmetic" appears to be more commonly used than "arithmetical." See the following google ngram search: arithmetical, arithmetic_ADJ

[justingrant]

In today's Temporal meeting, I suggested "Calendar Epoch Year" or "Epoch Year" to add to the list.

• Pros: Clarifies exactly how year differs from eraYear: the former is relative to a fixed per-calendar epoch, the latter to an era.
• Cons: Potential confusion with 1970 Unix epoch, although perhaps less confusion if we called it "Calendar Epoch Year".

Another related suggestion would be to define a term "calendar epoch" for the reference point of the year property.

[robertbastian]

Epoch year doesn't really work for cyclic calendars, where arithmetic year is the related ISO year, which doesn't have much to do with the calendar's epoch (if that is even defined).

[justingrant]

Epoch year doesn't really work for cyclic calendars, where arithmetic year is the related ISO year, which doesn't have much to do with the calendar's epoch (if that is even defined).

Are you assuming that "calendar epoch" (as we're discussing here in this issue) must have meaning in that calendar outside of its use in computing? If so, then I agree with you.

But I'm unsure about that assumption. My understanding of the word "epoch" when used in computing means something like "a fixed moment in time that events are measured in relation to". But I don't think a "calendar epoch" (in computing) must necessarily have inherent meaning in that calendar outside of computing. It could be arbitrary, like how the 1970-01-01 Unix epoch has no inherent meaning outside computing. But it's still an "epoch"; it's just an arbitrarily-defined one for the convenience of programmers.

This seems to be pretty close to the concept we're discussing in this issue: an arbitrary point in time, defined for the convenience of programmers, that is used as a reference point for measuring other events in that calendar.

BTW, another possible term could be "reference year". I don't think reference year has as clear connotations as epoch year, but I thought I'd volunteer it too. Here's what Google has to say:

[robertbastian]

I think reference year is a better term, but the API would need to be something like year-since-reference-year.

This seems to be pretty close to the concept we're discussing in this issue: an arbitrary point in time, defined for the convenience of programmers

It's not actually arbitrary though.

I think the issue with epoch is that it's not the UNIX epoch, which is the usual computing epoch. Also epoch is a term that is used more on the physical time side than the civil time side of things.

[Manishearth]

We already use referenceYear for MonthDay. We'd need to pick something else then.