Re: The Prize #6

[Tom Droege <tdroege2@earthlink.net>]

My experience so far is that selection of data helps very little.  So the 
prize might even be awarded for a way to select "good" data from the 
available data.   (Assuming there is something left)

One of the first things I am going to do after I get mike loaded with images 
is to get a big image set, sort it into "good" and "bad" images and see if 
there is a (big) difference in the quality of the photometry.  My bet is that 
there is little difference. 

We shall see.

Tom Droege

On Friday 08 October 2004 11:10 pm, Michael Koppelman wrote:
> Yup, I missed or forgot about this, Tom. My apologies.
>
> I still think the contest should have everyone using the same data.
> Otherwise people can simply exclude noisier data. I think you should
> say that it must include all data between date X and date Y so we can
> compare apples to apples.
>
> Cheers,
> Michael Koppelman
> http://www.lolife.com/astronomy/
>
> > 1) There should be a significant improvement in the tass survey data.
> > Something like that indicated in TN100 would be satisfactory to me.  I
> > will
> > work this out in more detail later if some of you don't come up with a
> > good
> > definition of success.
> > 2) No improvement, no prize.  This is not a "best effort" prize.  You
> > have to
> > achieve something.
> > 3) Collaborations are allowed.  You enter as a single person and make
> > your own
> > deals as to the division of the prize money.
> > 4) I reserve the right to drop the contest up until the time the money
> > is put
> > in escrow.  After that the contest is on.  I expect do this by the
> > first of
> > the year.  
> > 5) The probable period of the contest is Jan 1 to Dec 31 2005.
> > 6) The improvement must be real.  We Will figure out a way to test the
> > result.  
> > In other words, no hidden code that makes the result look good.
> > 7) A description of the algorithm used is required.  Fully documented
> > code is
> > not required.  This is supposed to be fun.  But documentation is
> > encouraged.  
> > 8) A submission consists of the code, the result, and a procedure that
> > allows
> > the entire tass survey to be processed in a reasonable time.  
> > 9) We hope you will put your code under the gnu license but it is not
> > required.  You just have to give permission to process the tass survey
> > with
> > it.  
> > 10) It has to be your code or stitched together public domain code.  
> > 11) I will set up a way that you can verify that the prize is really
> > in escrow
> > before you start work.  
> > 12) I will send out data sets as needed.  The .cal files are widely
> > available.  
> > The images will be made available as requested.  But within reason.  
> > 13) The method can use either the .cal files (preferred because it is
> > easier)
> > or the dark subtracted and flat fielded images.
> > 14) The winner will have unrestricted bragging rights for their
> > result.  While
> > it would be nice to put the result under the gnu license, if you
> > develop
> > something that has commercial profit, you are welcome to it provided
> > tass
> > gets the use the result.
> > 15) I will set a deadline to enter.  Say 1 Dec. 2004.
> > 16) Entry does not obligate you to do anything, but once you submit a
> > solution
> > and win, tass has the right to use it.
> > 17) The contest is open only to those who convince me that I want them
> > to be
> > in it.  It is not open to the general public.