[Kstars-devel] Ideal time for observing objects?

[Akarsh Simha]

What is the best algorithm to assign observing times to a list of
objects?

Currently, in the Observation Planner, we sort objects by transit
time, using a special sort model that sorts 6 PM before 12 AM.

In practice (tested on the field), the observation order produced by
this sort seems to work well for objects that are close to the
celestial equator.

This is because of the following reason: although the meridian transit
time is indeed the best observation time, at a high northern latitude,
it is advisable to observe the objects along the equator first,
because they set to lower altitudes more quickly. Especially when we
start the "race down the west". However, KStars doesn't take
declination into account.

Instead, I think the time used for sorting should be the time when the
object hits some heuristically defined lowest comfortable observing
altitude.

Ideally, this would vary from object-to-object. For instance, one
could observe M45 or M44 when it is just 10 degrees above the horizon
and get a reasonable view, but for the Corona Borealis cluster, it
better be as close to the zenith as possible.

So for the first trials, I'm thinking of adding a method to SkyObject
that calculates the time at which an object achieves a given altitude
with a western / eastern azimuth depending on a flag. Easy with the
conversion in EquatorialToHorizontal and the template in
SkyObject::transitTimeUT. Then, choose an arbitrary altitude for now,
say 20 degrees.

Really, this is a question of avoiding overcrowding any particular
region of time in the observing schedule. So maybe the better way to
assign observing times to objects is to look at their time of crossing
of this (arbitrary) 20 degree altitude, and try to space the objects
so that there is user-settable average amount of observing time
dedicated to each object. In this race, we could afford to delay the
observation of brighter objects.

I feel that an efficient planner will help the reasonably experienced
observer clock about 20 ~ 40 new observations per night.

Any more suggestions on implementing a good observation scheduler?

Regards
Akarsh