ADASS Day #2

These are my opinions and thoughts about the second day of ADASS.   It isn’t necessarily a summary of the day or complete but  just my reflections on the day along with some random commentary or thoughts.  Also these thoughts might not be in chronological order and can also be pretty random.  Plus I need to get out a rant, so if you aren’t in the mood for that, skip to #2.   Also if you are Brian Schmidt, please read through #2 before deciding whether or not I’m an idiot–which I likely am. 

1. At the beginning of his talk, Brian Schmidt commented on how twentieth century astronomy was driven by heroes — individuals or small groups making major break throughs– but today, we are in an era that major discoveries are driven by big surveys and big data.   I realize this fell into the narrative of his talk (and of the conference of entering an era of big data) and helps present a good story (and without a doubt, he likely knows all the history I’ll talk about now), but it bothered me.    At the same time that Hubble was deriving an erroneous value for the expansion of the Universe, teams primarily made up of women were doing the hard work of classify and measuring the properties of millions of starsa.  They developed classifications still used today and produced the huge data sets required to model stellar evolution that predicts the behavior of stars amazingly well.  They did this while often receiving little to no credit.    Today we are in a much better situation where more people are properly credited (and have the opportunity) for their contributions to astronomy.  Yet there are still small groups that are making unexplained discoveries (like Fast Radio bursts as highlighted in Mathew Bailes talk).   So yes, there were amazing heroes of twentieth century astronomy.   And while some of the heroes were discovering something new, others were producing the big data sets that explained the details of our Universe.   The same thing is going on today as well.

2. Brian Schmidt’s talk was awesome and highlighted so many great points.   Besides being the fourth submitter ever to astro-ph (and getting in trouble for having 7 MB of figures), he talked about so many things that astronomy does right and how we benefit from that.  We share our data.  We share our software.  We share our papers.  And those who do so are able to do better/more/faster science even if it might not always seem like it.   Many of the things we take for granted (arXiv, ADS, FITS) are practically unheard of in other fields.    The open source tool from 30 years ago is still being adapted, updated, and used today (and I might say complained about but at least we have the tool to complain about it and it is generally free).   However,  we have a problem of how to employ people so that they have the job security to work on the hard problems that will require years (or decades) to do.

3.  The next generation of big data is here.  SDSS represented the last generation and now projects like GAIA, Euclid, SKA, and others represent the next generation.  We are talking about billions of measurements of objects (Carlos Gabriel, Laurence Chaoul),  using HADOOP for handling distributed, large data sets (Lloyd Harischandra among others),  and the great, big radio data sets that we know are no longer coming, but are here (Mike Wise, Jan David Mol, Jan David Mol).

4.  Day #2 challenges from my perspective and some more on reflection:

  • How do we keep people who code employed in astronomy?
  • Why is there no prize for software in astronomy?   (If not AAS or IAU, why not award at ADASS as long as announced widely?)
  • As much data as we will produce, when we start to include things like likelihood functions, meta data, and everything else,  that is going to even increase it even more.  What’s the best way to represent additional information?

5. Outside of the talks, there has been a good bit of chatter on twitter about Sextractor.   Today it was started by Kyler Kuehn discussion of Sextractor and DES.   It is a great tool that is very fast, user friendly, and powerful.  It is also can be complex and only work well in certain regimes.  There are also some other amazing tools that might be appropriate but some might still be in development while others would need to be adapted from different wavelengths to the optical.

5. Software citations BOF was an interesting conversation.  For details, see the minutes and related documents.  However, since this is my post, I’m going to highlight the points I think are important:

  1. Authors: CITE THE SOFTWARE YOU USE
  2. Editors:  MAKE SURE AUTHORS CITE THE SOFTWARE THEY USE
  3. Referees:  MAKE SURE AUTHORS CITE THE SOFTWARE THEY USE
  4. Developers: MAKE SURE AUTHORS KNOW HOW TO CITE YOUR CODE (bonus points if you include a license)

Many of the software citation aspects are being pushed forward by Alice Allen and a number of other people that aren’t doing under-appreciated work.   The AAS is setting up new policies and does have a reference group giving feedback and people are thinking about better ways to do this, but the process might be a little slow and run into cultural differences in the astronomy community.   Journals like MNRAS are already encouraging users to cite software.

Likewise, Albert Accomazzi describes some of the ways to track the influence of data products in ADS.  The ADS is another example of incredible useful and invaluable tool.    There are a lot of people doing really under appreciated work on very useful tools and if you are reading this, thank you!

6. Squeezed in among a number of radio talks, Nuria Lorente gave a great talk on starbugs–I’d say the best kind of bugs.  These are robotic fiber positions with a soul.  Or at least personality.  And maybe ruby slippers (which is very appropriate for something from Oz).   These bugs can reconfigure your fiber positions quickly and simultaneously to minimize the amount of time between configuring a spectrograph on a new field.

7.  I love solar talks.   So many photons.   The Daniel K. Inouye Solar Telescope will be producing a lot of resolved data of the solar surface as Steven Berukoff told us,  and the public nature and how to deal with the data will challenge the solar community.

8. The Chandra point source talk by Janet Evans showed what was possible when adopting a development framework with planning, mitigating risks, testing, releases, and iterations.   She showed an absolutely gorgeous image of Tycho’s supernova remnant (I wish I could find an online copy and share it)  and although it did take longer  than expected, it has generally been very successful.  It is nice to see the project management in software projects that have worked for others.

Footnotes

a. Here’s a comparison:  SDSS measured the spectra of 500,000 stars by DR3.   So did Annie Jump Cannon in her lifetime.