Introduction

• Aim of the working group
• 2006 priorities (use cases : image and spectrum)
• Other use cases
• Next IVOA Interop meeting
• ...

Presentations

• Image use case status, Eric Slézak
  - observation data -> image available for a treatment (orientation on the Sky, distortion, ...)
  - sampling, ...
  - a part of the process is done manually, must be resolved before workflow in production phase
  - workflow quality depends on upstream quality
  - noise correlation
  - do not consider that the instrumental characteristics are not variable
  - first trimester : workflow finished for the China Interop meeting
  - Terapix tools (p.f. estimation, deconvolution, ...)
  - correct propagation of uncertainties
  - noise correlation
  Philippe : it is a problem of Data Model, not taken into account by the DM
  - visualisation tools
  Eric would like that the LSIIT tools available on line (especially Mathieu Petremand work)
  - new instruments : radio = image producer
  - workflow available for the China interop meeting
  - better quality of the workflow processing : backward, etc.
  - see the steps, the loops, ... -> classification of the sources
  - visualisation tools very important (date cubes, ...)
  - cf. slides
  

• Spectrum Use Case, Philippe Prugniel
  - many things are not taken into account by IVOA : for example the noise
  - simple bricks to use quickly
  - cf. last meeting in Paris of a group of spectrum producers and users
  - common bricks in image and spectrum domains
  - many people are ready to develop tools and services
  - interconnection spectrum/tools -> near from the workflow aspects
  - implementation to make science : it is possible to use simple workflows which are not necessary using VO tools
  - so, basic bricks to define : sampling, cross corelation, ...
  Eric : example of which could be pertinent to transmit between image and spectrum
  -> image of a galaxy with a star over
  many programs are available but it would be nice to show workflows involving them
  - cf. slides
  

• Research of objects in archives use case, Jérôme Berthier
  - work in progress : DENIS (index of color, astrometry), DFBS (spectrum)
  - ongoing work : ESO-R, POSS
  - DENIS represents 880740 images for ~90 days processing
  - workflow : (input/output) object list, aim -> create boxes with the correlators to use them in the workflows
  - error management, process duration
  - interactions, master/slave relation, input/output
  - langage and workflow bricks
  - cf. slides
  

New Use cases in other domains

• Research of objects in archives use case, Franck Le Petit
  - connecting software codes together is not easy : output data format -> data model problem, UCD
  - quantity definition not the same for the different codes
  - problems due to binary format : results are written is a big binary file (possible problem like little/big endian) -> a solution could be to read the binary on the server where it has been created but it is often stored on an another server, put a header in the file to describe the format, ...
  - conclusion : code and theoretical databases coupling, asynchronous computing, change of code to study physical results, observation comparison
  - cf. slides
  

• ... proposed by Loic Lechevallier
  - stop point important during the execution
  - full description will be available soon
  

Overview of different languages :

• cf. André's presentation

Short resume of the final discussion :

• Meeting in April to prepare the May China Interop

• Use cases to explain in the frame of the IVOA Theory IG
  - Image workflow
  - link with patterns
  - YAWL patterns ?
  - ESO Reflex to explore (meeting in March with R. Hook)
  

• IVOA Theory :
  - research results : it is important to store the history, it must be possible for a referee to controle the workflow described in an article -> so it must be easy to replay a workflow (user friendly tool, portable workflow defintion, etc. )
  

• IVOA Application WG
  - tools to show
  

• IVOA Data Model WG : interfacing with this group
  

• common bricks between the different use caes : probably not for the next Interop
  - but could be discussed (cf. SNAP) on the OV Theory mailing list
  - example : resampling and coordinate extraction, not the same algorithm but common semantic
  

• definition of a desirable list of bricks needed in the future (including one concerning the noise)
  

• lot of algorithms could be presented at IVOA (in image domain for example)
  

• mailing list