[Haskell-cafe] threadDelay delays less time than expected (Windows)

[Bardur Arantsson]

On 2016-12-30 11:09, Joachim Durchholz wrote:
> Am 30.12.2016 um 01:14 schrieb Bardur Arantsson:
>> Beause the leap second is entirely artifical, invented to
>> keep the *calendar* ("Earth" time) in sync with *physical* time.
> 
> Serving a practical purpose isn't "artificial" in my book.
> 

I'm emphatically *not* saying that there's no such thing as "calendar
time". I'm saying we need *separation* between "calendar" and "physical"
time. (Which I thin we agree on?)

> I haven't looked at Java 8 date/time yet, I hear it's heavily influenced
> by Jodatime so it's pretty likely that you can draw the same ideas from
> that.

Same principal developer :).

> I'd still start with Jodatime, just to have a better chance of
> interoperability with needs for alternate calendar systems. Mostly to
> glean a list of potentially useful data types, and which data fields are
> relevant to each.

The developer of both Joda and the jdk8 time.* system decided against,
so there's that. (And I think it's pretty much undisputed that he *is*
an expert and can actually make these trade-offs.)

Regards,