On the naïveness of time

[Enet4]

As far as I understand it, the naive module bears the semantic of lack of time-zone awareness. That is, types also bear a specific time zone unless it's a Naive variant. However, it has been suggested that this may bring some confusion. Namely:

• There is a NaiveTime type, but no Time. I would invite the community to think whether this value with a timezone is useful enough to earn a dedicated type.
• There is a Date type with a time zone, which does not make sense without the rest of the time instance, but it is useful as an intermediate result for the current builder pattern:

let dt = Utc.ymd(2014, 11, 28) // Date<Utc>
    .and_hms_milli(7, 8, 9, 10); // DateTime<Utc>

However, this intermediate type can be avoided if we start from a NaiveDate and add a (currently unexistent) with_timezone method. The deprecation of Date<Tz> has been hinted here too.

let dt = NaiveDate::from_ymd(2014, 11, 28) // NaiveDate
    .and_hms_milli(7, 8, 9, 10) // NaiveDateTime
    .with_timezone(Utc); // DateTime<Utc>

This makes me think that "naiveness" in this context could only be applied to a complete time instant: date plus time. And if so, the meaning of a Naive prefix in either a Date or a Time might just become void. I would like to use this issue to discuss and establish a clearer meaning and purpose of the naive module.

1. Should NaiveTime become just Time, thus discarding the possiblity of time instances with time zone awareness? I wrote a backwards-compatible proof of concept in this branch.
2. Should Date stay as it is in 1.0, or should it become a type without a time zone? Are there any benefits of this type other than enabling building a DateTime using the time zone type as the starting point?