[IPTA Clock] Kickstarting the clock project

[George.Hobbs at csiro.au]

Hi all,

  I hope that you all had a great holiday over the new year (unless you're Chinese and the new year is still to come).  It is my intention to finish a paper on "developing a pulsar-based timescale using the IPTA data sets" this year and so I plan to get started now.  The status so far is that:

1. KJ, Mike and I have had some initial chats about the project
2. The PPTA clock paper has been accepted and has now been published by MNRAS
3. All three PTAs have made their data available online
4. Joris is working on putting the datasets together to make an "official" IPTA data set.
5. Quite a few of us have informally put the three data sets together for a first-go at looking at the IPTA data set as a whole.
6. the IAU has agreed to form a Working Group on pulsar time scales.  I am currently leading that working group, but we haven't got started yet.  I'll keep the people on this email list updated with the terms of that working group and our status.  My intention is that the members of that working group join this email distribution list.  Note, as stated originally, just being part of this email list does not guarantee authorship on any resulting paper.

The current project plan is as follows and I'd welcome comments and suggestions:

1. I'll put together an informal IPTA data set quickly.  I hope to have this complete within 1 week.  This will be made available to people on this distribution list.  Note to Joris -- I know that putting together all the data sets very carefully is a long and difficult job.  My plan here is to have some data that we can use to test the algorithms. Of course, before publication we'll make use of the properly-formed data sets.
2. I'll work with Mike to simulate a set of data sets that look like the actual data, but include a large clock error.
3. These simulations will be made available to anybody on this list that wishes to try their algorithm on the data.  Note that the results will also be made available with the data.
4. If the algorithms produce inconsistent results then we'll start a discussion on what could be wrong
5. If/when the algorithms agree then we'll decide whether our paper will contain the results from all the algorithms or whether we'll just choose one.  (At the moment, I don't see why we wouldn't include all the algorithms.)
6. During this time, I'll be pushing on the various PTAs to make available as much of the early data as possible (i.e., so that we have data sets longer than ~6 years).
7. We'll then run the algorithm(s) on a final IPTA data set and obtain the differences between TT(IPTA) and TT(BIPM).
8. This will allow us to start writing the paper and interacting with the atomic time community to understand the implications of our result
9. We can then write the paper.

My suggestion is that I try and get the initial data sets and simulations ready by the end of Jan.  We can all get our algorithms running during Feb. and get going on discussing why we all get different answers.  That should then allow us to make more realistic predictions for how long it will take us to start writing the paper.

Best wishes,
 George








George Hobbs
Research Scientist
Astronomy and Space Science
CSIRO

Phone: +61 2 9372 4652
george.hobbs at csiro.au<mailto:george.hobbs at csiro.au> | www.csiro.au<http://www.csiro.au/>
Address: PO Box 76, Epping NSW 1710
Australia (Cnr Vimiera & Pembroke Roads, Marsfield NSW, 2122)