The 23rd Annual Astronomical Data Analysis and Software Systems (ADASS) is being held this week in Wailkoloa, Hawaii, and I will be reporting on the action. The conference includes oral presentations, posters, demos and Birds of A Feather sessions. This year’s themes are:
- Ground Based Observatories – from operations to archive
- Planetary Systems – here and elsewhere
- Cosmology Challenges – now and in the future
- Emerging Technologies
- User Interfaces
Many of Monday’s sessions were on the topic “Ground Based Observatories – from operations to archive,” and I’ll summarize some of the talks. Jean-Charles Cuillandre described an innovative low-cost software system in operations at the CFHT that uses a dynamic approach to assessing exposure times for queued observing, and it allows the observations to achieve the SNR needed by the astronomer – in effect, the system optimizes shutter time.
Tom Vermeulen described Remote Operation of the Canada-France-Hawaii telescope. The observatory redesigned its software systems to manage fully remote observing, with no-one at the summit. This has been in operation since 2011 and has proven successful, with less down time than previously. Danuta Sosnowska spoke on Using heuristic algorithms to optimize observing target sequences. She described how such algorithms, which can run in 10 seconds, are in use HARPS at the La Silla observatory, HARPS North at the La Palma observatory and SOPHIE at the Observatoire de Haute-Provence.
James Bauer (Quickly buy WISEly: The WISE Moving Object Pipeline Sub-System) described how this sub-system has so far detected 158,000 asteroids detected, including 34,000 new discoveries.
There were a number of talks about managing big data sets. Brian Glendenning talked about data management at NRAO in Data Management at the National Radio Astronomy Observatory, and Adriaan Renting talked about managing the 10 PB-scale data sets of LOFAR. Jesus Salgado talked about Gaia technological challenges for data archiving and emphasized the problems with using HADOOP, with its steep learning curve and low-level API.