The name 'Zoned' feels inconsistent and overly vague

[FeldrinH]

I appreciate that this is kind of a bikesheddy nitpick, but it still annoys me.

Why is the zoned datetime type in Jiff called Zoned and not ZonedDateTime? When you combine a Date and a Time you get a DateTime, but when you combine a DateTime and a TimeZone you get a Zoned. That feels odd and inconsistent.

Also, I suspect that for a user unfamiliar with Jiff the name Zoned is overly vague. It doesn't really communicate what the type stores. Just looking at the name it isn't even clear that it stores a time-related value.

[BurntSushi]

My main concern with ZonedDateTime is that it makes the type that I think most people should reach for much more verbose. That's why I named it Zoned.

I note that the first sentence in the docs of Zoned is:

A time zone aware instant in time.

And the first and second sentence of the Overview at the top of the crate docs are:

The primary type in this crate is Zoned. A Zoned value is a datetime that corresponds to a precise instant in time in a particular geographic region.

So I feel like it should be pretty hard to miss what a Zoned is here.

With that said, I have heard other folks complain about this. I'm open to receiving other feedback about this and possibly making the change for Jiff 1.0.

[tamasfe]

Most applications will be working with time but will not be centered around it, I predict that encountering a Zoned will incur a "a zoned what?" response more often than not.

Zoned is also vague enough to become confusing in applications that work with any kinds of zones other than timezones.

At work we have OffsetDateTimes all over the codebase and no one has mentioned its verbosity so far.

That being said I got used to Zoned just as quickly, It's not something I'd ever complain about if it wasn't about something that looks to become the best (de facto standard?) time library of rust.

[BurntSushi]

I've given this a lot of thought, and I'm inclined to make this change for the next breaking change release (likely 1.0, this summer).

[anton-dutov]

Postgresql has timestamp / timestamtz types, this seems logical and obvious (although it's not timestamp there), maybe we should consider the names DateTIme / DateTimeTz.

[BurntSushi]

Under no circumstances will I look to PostgreSQL for inspiration about anything related to datetime.

Besides, there is already a jiff::civil::DateTime.