These are my notes taken from Leslie's demo. Let me know if you have
other points that would be good to have in our memory.
- (A PCF is a piecewise continuous function; these are the
functions computed to create the scheduling windows)
- It would be nice to be able to have the VTT show the times in
the year when a particular RA and DEC would not be in solar
exclusion and how long a "sched 30" orbit is, ie, how many
minutes.
- The user can raise the "sched" value to get short orbits, ie,
less observing time, but which can schedule more during the year.
- But, this is a hard concept to explain to users.
- Currently, Trans may expand orbits beyond a requested sched
that is > 30% to sched 30 orbits.
Transverse will NOT do this.
- Would we like to have a graphical means to specify special
requirements, particularly temporal ones similarly to the USC's
(user specified constraints) in the spike gui?
- The problem with links is that there can be several different
solutions to a scheduling problem, but it is very hard to see
what those solutions are and understand the scientific trade-offs
among them. (Kinda like not being able to see the trees for the
forest)
- It would be good to be able to fix some constraints and then
ask what values can other constraints have to make visits
schedulable.
- We should advocate to users to use methodical approaches to
building proposals by adding constraints one at a time to be able
to better see which constraints raise scheduling issues.
(But, does this go against how PI's think about and "compose"
their proposals?).
- Links and the sched requirement are the hardest things for GO's
to understand and solve.
Maybe we need to come up with a better way of presenting and
explaining the concepts to users.
Maybe turn around the meaning of sched and have sched 100% to
mean the longest orbit they can have (ie a current sched 30) and
lesser percentages mean shorter orbits, but more schedulability.
- It might be good to have PC's and IS's track:
- Which proposals are the thorniest?
- How often particular problems occur?
- What are the solutions?
- How are PI's educated about the problems?
Do we want to ask PC's to log how they deal with user's phase 2
problems for cycle 11?
- But, we are also not learning about the problems that PI's
are solving on their own which we could make easier in APT's
design.