[ImageDM] Derived object

[Douglas Tody]

On Tue, 19 Nov 2013, CresitelloDittmar, Mark wrote:

> Arnold,
> 
> Yes.. a bit,
> There is currently no actual connection between the Image and Spectral
> models other than their use of/extension of the Observation class.
> So you are getting a little ahead of me.
> 
> With the discussion of the relation of Observation and Dataset, I have
> been looking into the separation of these components.
> 
> The printout I was using seems to have dropped several pages, so missed
> the write-up on Derived in the document
> (Thanks Doug for the excerpt.. it made me double-check).
> 
> If Derived holds information about the Target which is derived from the
> Dataset, then it seems that any object associated with a Target could
> potentially have Derived info about that Target.

That is how this is currently used, however in principle Derived could
hold other derived quantities not necessarily associated with a Target.

 	- Doug


> Mark
> 
> 
> On Tue, Nov 19, 2013 at 3:57 PM, Arnold Rots <arots at cfa.harvard.edu>
> wrote:
>       It seems to me that we are running into chicken-and-egg
>       problems here.
> What is the toplevel model?
> If ImageDM describes the generalized data hypercube,
> then Spectral is a subset, with all axes, except the spectral one,
> collapsed.
> If that is not the case, then there ought to be a generalized
> Hypercube
> model from which both ImageDM and Spectral are derived.
> 
> No matter what, all three objects you list should be included in
> the root model,
> because they have meaning for all models derived from it.
> However, I am not quite sure what SNR refers to (I know, I should
> look it up).
> I assumed it is some form of noise and sensitivity measure .
>
>   - Arnold
> 
> 
>