Software Development for ALMA Robert LUCAS IRAM Grenoble

Software Development for ALMA Robert LUCAS IRAM Grenoble France Hunt for Molecules, Paris, 2005 -Sep-20

Science Requirements • Science Software Requirements Committee work – Started in 1999, relatively early in the project – About 15 scientists from Europe, US, Japan – General requirements ready in 2002 – More detailed requirements by subsystem – Now one scientist attached to each subsystem to follow up the development work, answer questions, participate in testing. Hunt for Molecules, Paris, 2005 -Sep-20 2

Main Goals • Motivated by Science Requirements • Deal with a large variety of projects – Observing should be relatively easy at mm wavelengths, site is good – Much more difficult in sub-mm range • High degree of automation needed to guarantee observing efficiency at high data rate • Deliver data of controlled quality • Use should be easy for the non-specialist (unique instrument). Hunt for Molecules, Paris, 2005 -Sep-20 3

Proposal/Project Preparation • Uses an Observing Tool to prepare proposal submission and actual project editing • This tools knows the standard observer modes • Has novice and expert modes • Translates astronomical information into observing parameters • Provides reasonable defaults • Has components for observation setup, spectral setup, sensitivity calculation, … • It prepares the scheduling blocks, and informs the user of specific resources needed (data rate, data processing). Hunt for Molecules, Paris, 2005 -Sep-20 4

Scheduling • Dynamic, flexible scheduling is an essential feature to allow optimized use of variable site conditions • Use both stringency and scientific rating stringency = ratio of total available time to time during which a given programme could be scheduled • Project split in scheduling blocks – They cannot be split – They may be logically linked for complex projects Hunt for Molecules, Paris, 2005 -Sep-20 5

Observations • Control command language – simple, easily editable, programmable – control all basic hardware components – provides a way to develop the basic Observer modes and the Observatory modes • Operator interface – Used to control the actual operation of hardware and of the software subsystems. • Staff Astronomer Interface – Used to perform quality control Hunt for Molecules, Paris, 2005 -Sep-20 6

Data Acquisition and Archiving • A Science Data Model has been developed to handle the science data and associated observing parameters, and auxiliary data • Should be well suited for archiving and archival use of science data • Data written in real time into this model. • Mainly developed here at Paris Observatory. (f. Viallefond, M. Caillat, …) Hunt for Molecules, Paris, 2005 -Sep-20 7

On-Line Calibration • Needed in quasi-real time to ensure required data quality: – pointing, focus, delay measurements, … – Perform atmosphere calibration using state-of-the-art atmosphere model, optimize use of radiometric phase correction – Monitor atmosphere phase quality on phase calibrators and amplitude calibrators – Feedback to observing process and to dynamic scheduling • Developped at IRAM (Grenoble) Hunt for Molecules, Paris, 2005 -Sep-20 8

Quick Look • Display observing parameters and monitor online calibration • Provide feedback to operators and staff astronomers on the site on the quality of observations, of calibration, of data • Visual check of the data acquired itself (e. g. display spectra, raw image) Hunt for Molecules, Paris, 2005 -Sep-20 9

Science Pipeline • Run at the end of the observing • Perform data reduction in a quasi-automated way for most projects • An essential input for quality control by staff astronomers • Will need to be tested by the staff during the first years of science operation • Seen as an essential element to make ALMA well open to a wide community Hunt for Molecules, Paris, 2005 -Sep-20 10

Off-Line Data Reduction • Based on AIPS++ which is now essentially focused on ALMA and EVLA. • Development is regularly tested by would-be users and benchmarked against data processing requirements. Hunt for Molecules, Paris, 2005 -Sep-20 11

Development Process • Use of ALMA Common Software (ACS) – ‘Middle-ware’ layer that hides the underlying hardware / operating system – Ensures distributed software deployment: (CORBA based, for the specialists) • Common rules of software engineering • Regular releases (6 -month); but frequent integrations (1 -month) • Enforced testing policy Hunt for Molecules, Paris, 2005 -Sep-20 12

Organisation • Computing IPT led by B. Glendenning, G. Raffi, K. Tatematsu • Subsystems: – Software subsystems: Prop. Prep. (UK), Scheduling(NRAO), Control(NRAO), Correlator(NRAO), Telescope Calibration(IRAM), Pipeline(NRAO), Off-Line(NRAO), Archive(ESO), Executive(ESO), ACS(ESO), Operations(ESO), ACA(Japan) – Other: SSR (IRAM), High Level Arch. (ESO), Software Eng. (ESO), Int. and Testing(ESO) • Many phone meetings and travel required … Hunt for Molecules, Paris, 2005 -Sep-20 13

Development Status • In parallel: support for antenna evaluation, laboratory integration with prototype system (control/correlator) • Design, prototyping of the final end-to-end system • Integration of the whole system very early (R 1, R 2) but with limited functionality • Release R 3 for 2005 Oct 01, emphasis of simulation of hardware • Support ATF (VLA test site) operation early during 2006. • On-site commissioning and science verification in 2007/2008 Hunt for Molecules, Paris, 2005 -Sep-20 14