Re: Useful Research?

[Tom Droege <tdroege2@earthlink.net>]

Martin,

>Making maximum use of TASS data
>
>Perhaps I should start by saying that nothing I have written here should
>be seen as being in any way critical of individuals. Indeed some the
>points might be obvious to TASS readers - nevertheless I hope I have
>raised issues worth debating.

I appreciate that you are asking these questions.  I am happy to enter into 
a debate with you and anyone else who wants to join in.  Let me try to 
respond to some of your points:

>Let me start with a real attention grabber! Is looking for new variable
>stars a useful project?

Bohden Paczynski thinks that it is.  That is good enough for me.  He has 
been right on a lot of things lately.

>It seems to me that there are four main stages in useful astronomical
>research.
>
>In brief they are:
>1) Deciding on the project
>2) Collecting the data
>3) Processing the data
>4) Publishing the data in a manner acceptable to the wider astronomical
>community

This is the way you do it if you are doing sponsored research.  If you ask 
for money then you have to have a plan and stick to it.  If you are doing 
things on your own, then you can be an opportunistic adventurer and take 
big risks and do things just because you can do them.  One might or might 
not win doing this.  But it is then an individual's choice.

>I have used the phrase "useful astronomical research" quite deliberately
>since much time and effort is wasted by amateur astronomers on projects
>that are rendered pointless when the participants ignore one or more of
>the basic principles of science. "One variable at a time" is a well
>established doctrine that astronomers should ignore at their peril.

Why is this true?  You don't present any argument for this "doctrine".  I 
will make the argument that measurement of anything just for measurement's 
sake is constructive.  I present Tycho Brae as an example.  I am betting 
(my time and effort) that a big catalog of measurements is useful to 
science.  Note that I am attempting uniform coverage of the sky.  I notice 
that most variable star hunters seem to concentrate in the Milky Way where 
there are more stars per square degree.  This introduces a bias, I 
think.  The Mark IV was deliberately designed so that it could only be 
operated as a survey instrument.  It cannot follow the Milky Way any longer 
than any other part of the sky.

>Let me turn to the specifics of variable star research. It seems to me
>that taking images of the same part of the sky at intervals is a
>perfectly acceptable way of searching for stars that appear to vary in
>brightness. The key though is the phrase "that appear to vary in
>brightness". Any experienced user of a CCD will tell you that all sorts
>of factors can influence the readings obtained - the degree of moonlight
>and the presence or absence of thin high cloud are just two factors that
>spring to mind.  I have also carried out work on asteroid light-curves
>that show variations of as much as 0.1 magnitudes in readings taken 5
>minutes apart on nights where the above mentioned variables can be
>ignored. I respectfully suggest that unless we can standardise the
>results taken on different nights much of the value of the data
>collected disappears.

Part of our program is to measure and quantify the error of the 
measurements.  You may notice that all the measurements in DS23 include 
error terms.  We may not have it completely right yet, but that is a major 
part of the program.  Some variable star measurement groups don't seem to 
be concerned with error, and they don't discuss error.   See TN-83 for one 
look at the Mark IV data.  See some of Michael Richmond's TNs also.  I have 
been looking at the errors and am pretty happy with the results so 
far.  One would always like them to be smaller, but the night to night 
variations appear to be within the error expected from the plots like those 
in TN-83.  The point to point variations are smaller, but that is just 
measurement.

>Calibrating every value on every image against the results obtained for
>ten or a dozen known non-variable stars that appear on the same exposure
>might be helpful.

The frames I took last night had 403 calibration stars.  That is 25 per 
square degree.  This is typical for a sparse field.   There are many more 
as we go through the Milky Way.

>I read recently that the vast majority of scientific data is never
>looked at by those responsible for collecting it. Whatever processing
>methodology TASS decides to adopt it needs to involve as many people as
>possible. Too few "data crunchers" can mean that data is coming in
>faster than it can be processed. The backlog then never gets cleared.

As of last night I had caught up on processing all the data taken since the 
engineering run became stable last December.  I had last night's rather 
slim data processed to a star list by 8 AM this morning.  I have now 
completed the analysis of roughly 18000 images.  I have only had a working 
analysis pipeline for a few weeks.  While I expect to reprocess all the 
images again after we make further improvements in the pipeline, the data 
is good enough to make measurements as indicated by TN-83.

>I identified 900 variables in earlier TASS data that needed to be
>examined more carefully. The data has been in the public domain since
>December 1999 but virtually no progress has been made. By the time my
>web site had been created most of the active participants had moved on
>to collecting yet more data. Take a quick look at
>
>http://www.further-education.com/tass/variablefront.htm

I did not know that you were doing this work.  I assume this was done on 
the 10xcat?  A brief look at your web page did not show what data you were 
examining.  I would invite you to be more active on the list if you have 
this interest.  You can influence what is measured by participating.

>I suppose what I am saying is that collecting data without the
>supporting infrastructure to process it is not a sensible policy.

I certainly agree with this.  I have had a plan all along.  I could see 
that when I started this program that it would require a monster 
infrastructure to perform the analysis.  I could also see that by the time 
I got the instruments built, taking data, and the software working, that I 
could do the analysis solo.  The result bears this out.  I can reduce the 
data I collect in real time.  I have been doing so for a week or so.  I 
have also caught up on the old data.

>Moving on - suppose we are satisfied that a star is variable. Without
>accurate information on the position, the period and the amplitude of
>the variation then this information is hardly worthy of note.
>Information on a minimum of two complete cycles should be regarded as
>the qualification for publication. This might take the best part of two
>years of data collection for long period variables.

I think we disagree here.  The position measurement appear to be good to an 
arc second which is good enough to match them to more accurate position 
catalogs like UCAC2.  I think just collecting measurements and making them 
available in an archive is a useful scientific result.  It is not all the 
work, but it is some of the work.  So I am content to make 100,000,000 or 
so photometric measurements on 3,000,000 stars a year from my installation 
and put them in an archive.  I will attempt to collect a variety of 
epochs.  There are some that are hard for me.  This is my primary 
goal.  For amusement on the side I may try to identify and catalog 
variables.  This is not my primary goal.  I figure that once I get enough 
measurements in the archive that it will be actively used.  But that is 
just my bet.

>Once the information is of the required quality and quantity where
>should it be published so that other astronomers can find it? It seems
>to me that although most enthusiasts have heard of the GCVS there is
>still some uncertainty as to exactly how to get information included in
>the catalogue. There are lots of variable results floating around on any
>number of web sites that have clearly never been submitted to any third
>party.

Bohden Paczynski was one of the people who prompted me to do this work.  I 
just sent him a disk with millions of measurements.  Early on he said that 
if I took measurements, he would find a way to archive them.  He is now 
working on it.  Nothing like shoving data at someone to get them away from 
talking about how to do it and into actual doing.

I agree that a there needs to be a way to access all the available 
measurements.  I do not know what form this will take.  One way to cause it 
to happen is to accumulate several large data bases of good 
measurements.  I will do my part by creating one.  If there is good data 
available, then there will be pressure to get at it.  It is not so large a 
problem.  I figure that all the data I can take in a year will fit on one 
of today's DVDs.  No doubt by the time I run for 10 years (I do not expect 
to live to run longer) it will fit on one storage device of the future.  So 
I can just send out copies of my data base to all who ask for it.  I am 
doing that now.  I just sent out 26 copies of the data taken between 6 an 8 
hrs at +7.

Your questions are welcome.  Let us carry on this discussion.

Tom Droege