Re: [threeten-develop] Instant should implement DateTime

[Roger R. <Roger.Riggs@Oracle.com>]

I would keep Instant as a simple point in time but accessible  via DateTime
with EPOCH_DAY, and NANO_OF_DAY, NANO_OF_SECOND fields plus 
extract(ZoneOffset)
returning a UTC offset so it is precisely defined at runtime.

There is no field  EPOCH_SECOND which is prominent in the methods of 
Instant.
Might be worth considering.

See https://github.com/RogerRiggs/threeten branch 'instant-rr'  for the 
spec, impl and a couple of tests.

Roger


On 8/7/2012 6:49 PM, Stephen Colebourne wrote:
> On 7 August 2012 18:40, roger riggs<rog...@or...>  wrote:
>> Some of the conversions would be simpler and more consistent if Instant
>> could be treated as supporting the EPOCH_DAY and EPOCH_SECOND fields.
>> There isn't enough precision to support nanoseconds so some conversions
>> to time might be loose the nanoseconds.
>> A reasonable alternative would be to supporting only EPOCH_DAY.
> If Instant implemented DateTime (or AdjustableDateTime) then it could
> support the full range of fields using the UTC conversion to a local
> date/time. Or it could, as you suggest, support a subset.
>
>> Am I missing something subtle?
> One point is with the definition of UTC vs local time. EPOCH_DAY is
> actually LOCAL_EPOCH_DAY, and the same with EPOCH_SECOND, something
> that is quite subtle in the API right now. (This BTW might argue for
> two separate fields for these concepts, or for ODT to take into
> account the offset).
>
> The real issue with Instant is one of intent. The class is intended to
> be a simple point in time, a time-stamp that should only really be
> used to record the instant and compare it earlier/later than another
> one. However, us humans still tend to want to print it out, which
> causes problems.
>
> Thus the underlying question is whether Instant should become just the
> UTC version of ODT, or therefore potentially removed entirely. For me,
> I find modelling with Instant useful because it gets people to
> separate the storage of a totally neutral instant in time from the
> representation given to it by humans.
>
> Perhaps implementing DateTIme would be OK though, it just feels like
> the thin end of the wedge,
>
> Stephen
>
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats.http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
> _______________________________________________
> threeten-develop mailing list
> thr...@li...
> https://lists.sourceforge.net/lists/listinfo/threeten-develop

-- 
Thanks, Roger

Oracle Java Platform Group

Green Oracle <http://www.oracle.com/commitment> Oracle is committed to 
developing practices and products that help protect the environment