SED data model v0.92

[Ed Shaya]

Gilles,

   Yes.  The schema should certainly allow for upper and lower limits 
alone (flux.detectionThreshold is fine in spectra and images, and one 
may need another thing for saturation level which results in knowledge 
of only an upper limit).  It is necessary for both legacy data in which 
the value and error are simply lost forever and for experiments in which 
there is no clearcut algorithm to estimate a value.

Our disagreement amounts to the fact that I would encourage the observer 
to use value and error if at all possible, and you seem to encourage use 
of the fluxDetectionThreshold.  In the end, it is not up to us anyway.  
  It would be up to the observer/experimenter to decide which is 
appropriate.  

Ed

Gilles DUVERT wrote:

> Sorry for the long blackout. More pressing matters, not Alzheimer (I 
> think).
> Did you say you agree that the property Flux.DetectionThreshold can be 
> present (and can be present even without [ Flux.value + 
> Flux.StatErrXXX ])?
> Best
> Gilles
>
> Ed Shaya wrote:
>
>> Gilles,
>>    We seem to agree now.  Perhaps we always did.
>> Ed
>>
>>
>
>