Reconsider numbering of C-type associations

[stscijgbot-jp]

Issue JP-3870 was created on JIRA by David Law:

Between https://jira.stsci.edu/browse/JSOCINT-859 removed group-type associations, and https://jira.stsci.edu/browse/JP-3116 no longer creating L2 and L3 Obs-type associations when they are part of background Candidate-type associations, it may be worth looking back at how Candidate-type associations are named.

This ticket is to start that discussion; is there a better way to do this numbering?  And is it worth making any such changes?

Namely, c-type associations aren't (I think) easy to map to a specific observation number.  Often that won't matter too much, as the target name is enough.  However, that's not always the case.  As an example, consider PID 4499 which observes the same source each month.  I can't tell what c-type observation corresponds to which month, so I usually looked for the o-type observations instead when I want to get the raw data for a specific Observation number.  In this case the Observation numbers are 80-101 (mix of science and background observations), while the Candidate numbers are 1022-1032.

Is there some way that we can make the Candidate numbers more obviously linked to the observation numbers that they're composed of?

Complications pointed out by Tyler Pauly :

• Background-type candidate associations may be possible to link to the corresponding science observation number in some way, but what about mosaic and coron-type associations where there may be many science observations?
• c1NNN identifiers are set within APT, not by the pipeline.  Is there a way to make this numbering clearer in APT?