asymmetric systematic error for flux

[GernotMaier]

Is it possible for a pl2 spectral model (or for any model) to give an upper/limit systematic error?

In the VERITAS paper on B2 1215+30 (2017), the systematic error is given as +4 / -1 x 10^-10 1/cm2/s.

[cdeil]

For spectral points, we do support upper limits, but not systematic errors:
http://gamma-astro-data-formats.readthedocs.io/en/latest/spectra/flux_points/index.html

For spectral models, we do currently have an err_sys field, but no asymmetric systematic errors or upper limits:

/Users/deil/code/gamma-cat/input/data/2005/2005A%26A...432L..25A/tev-000110.yaml
21:      flux: {val: 5.7, err: 0.7, err_sys: 1.2, scale: 1e-12, unit: cm-2 s-1}
22:      index: {val: 2.40, err: 0.11, err_sys: 0.20}

Over time, we can and should make our schema more complex to cover more information, within reason (i.e. someone has to document it, and users have to be able to grok it).

What I suggest is that you add the extra information you would like to put for now in a comment with a TODO marker. Example:

/Users/deil/code/gamma-cat/input/data/2005/2005A%26A...442L..25A/tev-000118.yaml
27:  # stats: {chi2: 9.8, ndof: 8} # TODO: add this to the spec schema
28:  # livetime: 5.4 # TODO: add spectrum livetime to the schema

This allows you to get this in for now, and then we have this reminder issue open and the TODO marker to extend our schema.

If you want to do this now, OK, please make a proposal for extra fields, directly in the example file you're adding now.