From: Michael Wood-Vasey (wmwood-vasey@lbl.gov)
Date: Thu Jan 22 2004 - 07:41:34 PST
Dear Serena,
Here are my scientific-presentation level comments.
1)
You mention calculating I_1 and I_2 in terms of flux and quoting them
in magnitude. While magnitude is useful to the reader, I am surprised
that the errors you quote on the magnitudes are symmetric to the
degree of precision you quote. Generally many errors are symmetric in
flux space and thus not symmetric in magnitude space. For example, a
flux = 1.0+-0.1
would translate to a
magnitude = 0 +0.114394-0.103482
This is a standard astronomical bad habit, but there's no need to fall into it.
Is there some subtle reason that I'm missing that explains why the magnitudes should have symmetric errors in magnitude space?
2)
As Alex mentioned, add a parenthetical note or footnote defining the
dispersion you are quoting (e.g. RMS or RMS weighted by the inverse
square of uncertainties, etc.) Similarly, you're going to have to
figure out what Hamuy (end of Section 3) and others did when comparing
your dispersion to others. Otherwise, the comparison isn't as
meaningful as currently implied in the paper.
I also believe that if you're going to fold the dispersion back into
the Hubble-diagram plot, then it should definitely done with the
error-weighted RMS.
3)
I think the notation t_sec is confusing. Many people will have to
keep reminding themselves that t_sec is not time in seconds.
I found myself having to do this on my first read through the paper.
Could you call it t_bump, or perhaps something else?
[If you do change the t_sec subscript then of course change the I_sec subscript as well.]
I recognize that this may just be me, but since this is my first
reading through the paper, my reactions may be more like the general
scientific audience.
4)
Fig. 8
I don't understand why all of the t_sec values for the supernovae line
up with the axis markings. I understand that the uncertainty on t_sec
may be on the order of a day, but rounding t_sec is unnecessary and
adds an odd pattern to the plot.
5)
Fig. 11
What do you mean amplified by a factor of 4. As there's no scale for
the SN spectrum, an overall amplification doesn't affect anything.
Did you do a smooth fit to the spectrum and then amplify features
around that smooth line?
- Michael
This archive was generated by hypermail 2.1.4 : Thu Jan 22 2004 - 07:41:34 PST