[AstroPy] Convert GPS time to UTC time

[Aldcroft, Thomas]

Based on the discussion here I added an astropy issue to consider adding a
GPS time scale:

https://github.com/astropy/astropy/issues/1879

Cheers,
Tom


On Mon, Dec 9, 2013 at 3:58 PM, Paul Hirst <phirst at gemini.edu> wrote:

>  On 9/12/13, 09:48 , "Aldcroft, Thomas" <aldcroft at head.cfa.harvard.edu>
> wrote:
>
>
>  On Mon, Dec 9, 2013 at 12:22 PM, David Berry <d.berry at jach.hawaii.edu>wrote:
>
>> On 9 December 2013 16:57, Aldcroft, Thomas
>> <aldcroft at head.cfa.harvard.edu> wrote:
>> > This basically creates a Time object with the GPS format
>> representation.  By
>> > default this object is in TAI scale, so you first convert to UTC and
>> then
>> > ask for the ISO format representation.  As I understand GPS is not a
>> time
>> > scale, but a simple difference of time in the TAI scale from the given
>> > epoch.
>>
>>  Doesn't that make it a timescale? After all TT (Terrestrial Time) is
>> also just a constant offset from TAI.
>>
>
>  This is a good question and one for which I'm definitely interested in
> input.
>
>  From what I can see there are two ways to view "GPS time":
>
>  1.  As a time scale which is behind TAI by 19 seconds.
> 2.  As a time format which is the floating point number giving the number
> of TAI seconds since UTC 1980-01-06 00:00:00.  This is roughly similar to
> the "unix" format or "cxcsec" format.
>
>
>
>  I would have thought (1) was actually a more appropriate representation.
> I'm not sure of the formal definition, but I always considered GPS time to
> be another timescale, which just happens to have a nice and simple
> transformation to and from TAI - I understand it's by definition just the
> simple and exact 19.0 second offset.
>
>  Depending on the level of accuracy required, I don't think that's true
> for TT as suggested above, the TT = TAI - 32.184s is an approximation; a
> more accurate transformation is given at eg
> ftp://tai.bipm.org/TFG/TT(BIPM)/TTBIPM.12
>
>  Now, I'm not suggesting that this is a priority for implementation right
> now, but in general I'm thinking that having things as separate time scales
> would make it easier in the future for someone to contribute a more
> advanced transformation - as opposed to figuring out how to make it a
> separate timescale first.
>
>  Cheers,
>  Paul.
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/astropy/attachments/20131209/842a35ec/attachment.html>