B-V observation source plugin creates observations with no names

[dbenn]

This prevents AAVSO download format files containing just B-V from loading. Take name from B or V ob.

Also, as a fall-back, have AAVSO download format sink plugin set name to Unknown if null or "".

[dbenn]

https://www.aavso.org/comment/161889#comment-161889, the forum comment that indirectly led to this issue being created, raised the question of whether other fields should be added to B-V observations, such as Transformed, comparison star fields (CMag, KMag) charts etc. Clearly comparison star fields, CMag, KMag are problematic because they are specific to the original B and V observations. The time value is the mean of the B and V observation times, for example. The magnitude is of course B-V and no uncertainty is recorded. I suppose that mean uncertainty could be used, but what does this really mean? These are after all, synthetic observations.

[mpyat2]

the resulted uncertainty could be = sqrt(unsert_B^2 + uncertA^2), if I correctly recall the rules for uncertainties of combined values. This must be checked though.

[dbenn]

Okay, yes, probably better than mean error @mpyat2 :) Do you recall the CHOICE course, uncertainty about uncertainty, several years ago. Barbara Harris ran it but I don't think it ran more than once or twice. Perhaps we could ask Brad Walter, or consult CCD/DSLR manuals?

[dbenn]

This paper Uncertainty-2005SASS_24107K.pdf by Michael Koppelmann takes the approach you are suggesting @mpyat2. See Equation 7. I'm going to make this change to the plugin. Thanks!

[mpyat2]

Unfortunately, I have not participated in that CHOICE course, it is a residual memory from my first year at the University, laboratory work :) Thank you for the article!

[dbenn]

Yes, I recalled having seen that elsewhere when you mentioned it, other than AAVSO although I suspect I have spoken with Brad Walter about this as well. :)

Taylor (1982), p77 cited by Michael K in that paper has the equation.