Lessons learnt with Aladin and characterization experience for

Lessons learnt with Aladin and characterization experience for SIA 2. 0 F. Bonnarel, CDS (credit to Aladin developpers, CADC, ECF, ESAC, ESO VO people, DAL and DM IVOA WG)

SIA 1. 0 (2002 !!!) • Pros: – Solved the availabilty issue in a standard way for the first time. – Gave some rather unformal description/characterization of the images • Limitations: – No mapping on datamodel – Ad hoc UCDS and no utypes – No structuration of the query response. – No staging

Query phase: extended ROI • Extend ROI requests to all or a significant subset of stc Astro. Coord. Area – Use case: Draw a contour on top of a Preview or a wide field image and use it as a ROI. – Trick : drawn contours are on the image plane How to translate that in string stc in some given coordinate system….

SIA 2. 0 Query interface: SSA 1. 0 + ? ? ? • Requests by more characterization utypes (sampling in combination with resolution) • More request by ranges (resolution and sampling) • Add query by Observation/characterization utypes using « utype Extended » ADQL as it is proposed for TAP (non redondant with parameter= value for level 3 or 4 queries)?

Query response: extended datamodel usage • Use of Curation, Target, Dataset. ID and Characterization as in SSA. • Extend SSA Characterization package towards full level 2 for resolutions and sampling for spatial and additional sampled axes ( spectral or whatever) • Level 3 and Level 4 maybe available through DAL Extension mechanism (described in SSA): – Level 3 example : footprint description – Level 4 = PSF described as a link to a PSF service (eg TERAPIX qualityfits ) JPEG visualized in Aladin – Level 4 = Filter transmission curve available in Aladin

Query response: Dataset linkage • Provenance: – instrumental (heavy stuff in FITS keywords) – Processing ( calibration --- links to unprocessed data, ancillary data) • Access and views of complex or heterogeneous data (Linkage model) – Linkage to some simple derived datasets or excerpts of the main dataset (heterogeneous DAL protocols, generic dataset discovery) • Parallel development of a provenance or dataset linkage datamodel by DM WG

Staging/Asynchronous access • • Individual images may need heavy statistical processing (Bayesian fusion, Markov multispectral segmentation) – as eg CDS/LSIIT collaboration has shown. • Extremly faint source detection • Image cleaning from cosmics … Nice to get staging facility to access results later from another aladin session replaying the original images sequences for comparison… – Idea: original query response give acref to process follow up page, – replaced by standard DAL (image, spectrum, table) query response when processing is completed – Interface has only to keep track of original acref ….