The furor over Nova Del 2013 caused a 2010 AJ paper on novae classification across my screen. The following passage really caught my eye:
Some might think that a disadvantage of the AAVSO data
is that the magnitudes are estimated by eye, and they would
worry about reliability. This is wrong. The reliability of the
AAVSO magnitudes is about the same as for observations in
the professional literature. Indeed, with the AAVSO having
many independent observers providing measures at effectively
the same time, the rejection of systematic errors can be more
reliably done with AAVSO data than with professional data
which are generally sparse. An alternative worry is that the
visual magnitude estimates have a 1σ uncertainty that is so large
as to be unusable. Indeed, from extensive experience and control
studies, we find a typical 1σ error for single visual magnitude
estimates to be 0.15 mag. For professional astronomers using
photoelectric or CCD magnitudes with real uncertainties of
0.015 mag for a single measure even for good photon statistics
(Landolt 1992, 2009), the visual accuracy sounds poor. But this
worry is naive for two reasons. First, the AAVSO light curves
generally have many observations, and the usual measurement
errors improve by a factor of the square root of that number, so
the typical real uncertainty is more like 0.05 mag. This is not so
far from the 0.015 mag attainable with modern instrumentation.
Second, the required accuracy for nova light curves is much
poorer than 0.15 mag, simply due to the features being sought
having a much larger amplitude than 0.15 mag. That is, for
a light curve that fades by, say, 10 mag, even half-magnitude
accuracy is adequate for defining the shape and features of a
light curve, while the usual amplitude of flares, oscillations,
and dips is always much larger than the uncertainty in even one
visual estimate. Itwould be “nice” to haveCCDmagnitudes over
visual magnitudes, but the improved accuracy does not translate
into anything useable or any improved science. For the purposes
at hand, the naive worries are not realized, and the AAVSO light
curves are comparable in quality to those from professional
observatories (while having vastly better time coverage).
-Strope, Schaefer, and Henden, "Catalog of 93 Nova Light Curves: Classification and Properties", AJ 140:34–62, 2010
From time to time the great CCD vs. visual debate breaks out. This answers the eternal question most eloquently: there is a place for both, each is the right tool for certain classes of stars at certain apparent brightnesses. In the case of Nova Del 2013, at least in its very bright phase ongoing as I write this, a densely observed visual light curve is as good as a CCD light curve for the purpose at hand, and there is a lot lower overhead in terms of time and equipment. A very bright object like this is routine for binocular observers, but presents unusual challenges for CCD observers (see Henden's forum post on this: Bright nova photometry).
Each is a good tool, and each has its place. Whatever your method, observe wisely and well, because "Theories crumble, but good observations never fade." -Harlow Shapley
Thanks to Mike Simonsen for providing the quote.
Aloha,
Jim