Oral Abstract

Oral Contribution (O3.2) Gijs Verdoes Kleijn (Kapteyn Astronomical Institute, University of Groningen)

Theme: Evolution of software development and management

Data model as agile basis for evolving calibration software

We design the data calibration and processing for the imager of MICADO, the first light near-IR instrument of the Extremely Large Telescope. In this process we have hit the limit of what can be achieved with software design that is primarily captured in pdf/word documents. Unfortunately, collaboration through print-formatted documents does not scale:

1) Collaborations as large as MICADO and telescopes as complex as the ELT, result in numerous long documents. Maintaining the intra- and inter-consistency between the software and hardware design documents is nigh impossible.
2) Precise trade-offs are required to meet stringent science requirements. This requires a lower turn-around time between the software and hardware groups than can be achieved with documents only.
3) More (prototype) software needs to be developed in the early phases of the project to validate such trade-offs: simulators, archives, pipelines. A machine readable version of the design is therefore essential.

We turn the process around. We construct a fully digital detailed design that is both machine and human readable. From this, the design documentation, prototype pipelines and data archives are generated automatically. This digital-native design ensures that:

1) The entire design is internally consistent, making it easier to adapt to inevitable change.
2) Collaboration is quick and frictionless.
3) All prototype software meets the design, and can be a stepping stone to the final software.

We present a comprehensive approach of such a digital design for the ELT MICADO imager based on lessons learned in earlier projects (e.g. OmegaCAM, MUSE, Euclid).