[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Orekit Developers] Releasing 9.1?



Hi Evan,

"Ward, Evan" <Evan.Ward@nrl.navy.mil> a écrit :

Hi Luc,

On Wed, 2017-10-11 at 18:47 +0200, MAISONOBE Luc wrote:

Hi all,

I would like to release Orekit 9.1 before the Orekit day.
There have been a number of fixes and new features that
would be interesting to put public.

The new things already present are mainly related to
precise orbit determination. Displacement of ground
points (read ground stations) due to direct tide
effect and indirect ocean loading effect are now
available. A simpler way to retrieve covariance
matrix is available (but it probably need some works
as it still includes weights multipliers).

Another thing we could add is merging the work done
by two SOCIS students, one in 2015 and one this
summer. They worked on a way to add inertial forces
and hence to be able to propagate trajectories (which
in this case would not be orbits) in non-inertial frames,
and to add a way to propagate trajectories around
Lagrange points. A side effect is that new frames
for L1 and L2 points are also available, where users
select the primary and secondary body at will.


I took a brief look at the changes and I think merging this would
require a new major release because it changes the semantics of some
common APIs. For example, for a SpacecraftState s s.getPVCoordinates()
used to be exactly the same as s.getOrbit().getPVCoordinates() and now
they have different behavior.

You are right. So this merge must de delayed.




If possible, I may also add new things related to
GNSS (RINEX and ANTEX files parsing for example).


This would be a good feature. Are there plans to add pseudorange / single
difference / double difference measurements to the estimation package?

I want to add as many measurements as possible. We are clearly stepping
into the navigation domain. I would also like to add various combination
of phase measurements. However I don't think we will add everything
already in 9.1.



I will ask in the Hipparchus developers list if we
can also release Hipparchus 1.2, as it brings a few
bug fixes that are interesting for us (for example in
the ellipsoid tesselation)

Here are the current open issues (apart from the very
old ones that correspond to features with very low priority):

  351 for partial derivative in non-Cartesian elements,
  358 about tolerances reset when orbit type change
  359 about central attraction estimation
  361 about ITRF2008 and CCSDS
  362 about propagator converter step size

I think 351 would be nive to solve.

I won't have time to work on this one before Orekit Day. There are two existing workarounds.

Fine.



I think 358 should be rejected as it is both a very
   specific use case and would require major changes
   in both Orekit and Hipparchus

I agree.


I think 359 may be solved, but we need to merge the
   SOCIS students work before as it is related to this
I think 361 and 362 should be solved and are probably simple

I should be able to fix these in the next few weeks.

Great! Thanks.





Do you have other ideas?

Do you agree with releasing soon?

Yes, I like frequent releases. How soon are you thinking? A month from now?

Yes, something like that.

best regards,
Luc


Best Regards,
Evan


best regards,
Luc