-- AndreasWicenec - 06 Sep 2006 -- AnitaRichards - 04 Sep 2006

Meeting of VOTech Science Team at CDS 6 September 2006

Start 9.30 am, finish by 12.00 am

See VOtech Science page links to previous minutes and draft Science Framework Document.

Attendance

• Present: Anita, Mark, Andreas, Paolo
• Apologies: Nic, Eduardo, Guiseppe, Matteo, Jonathan Tedds

Review progress towards meeting tech requirements of science cases

General impression that some people are working on tools/applications which are not directly associated to a science requirement. People should outline which science requirement their work is addressing. The distribution of people makes it hard to tie them to the broader VOTech picture. Very interesting progress in several areas, the most obviously useful (to us) were:

• improved interoperability using Plastic
• new features in Aladin and TopCat^?
• SED builder/magnitude flux conversion (under development)
• Ranking of query results, although the integration/usage of this seems to be quite complicated. Useful ranking is only possible if more query options are offered (VOscope) and if registries contain valid and good descriptions of the datasets (e.g. coverage).
• Catalogue homogenization tool
• Applications for time-variable phenomena

Various Remarks

• Plastizised tools are essential and should be pushed forward (e.g. DS9)
• How will 'R' be used in the VO world?
  • Why does the user need it locally? Why not alternatively supply algorithms where the data are?
  • Feedback: Security and server load issues. Richard H and Jonathan T to discuss use cases and possible selected routines for CEA wrapping.
• What is happenning about integration of AstroNeural of in VOTech framework?
• What is the timescale for progress in datamining tools in general?
• Quality control of registry entries is necessary and data providers should be asked to update incomplete entries. This is a DCA topic.
  • This is becoming more vital as the number of resources increases so that it is important to allow selection by waveband, time etc. as well as just position - but in many cases the information is missing.
  • Find good examples e.g. ESA (Madrid).
• Access to Vizier should finally be addressed and resolved.
  • Registry entries can be visible but the data cannot actually be got at by most VO services e.g. AstroScope, workflows, ADQL.
• Authentication: ESOs problem is a good example case for the problems to be solved and they have to be solved soon.
  • Make requirements clear - e.g. users need to be identified individually not just by email domain name.
  • Possible liaison with RadioNet Common Proposal Tool development e.g. rules for sorting out access permissions for multiple partly overlapping groups of PIs/CoIs.
• Best fitting tool: Chi2 minimisation would be a perfect start e.g. applied to spectral template fitting. Eduardo and Anita to supply test data. (Check status of GAVO (PN) project?).
• Workflows: two different types why? Example workflows necessary. SED could be a template workflow which people would modify to their needs.
  • Avoid non-essential choices e.g., for a given data set, default SExtractor unless user elects to choose between multiple locations; default configuration files if available.
  • Under consideration: switching to more general workflow engine e.g. Triana (need to support server-side operation i.e. you can log out but workflow progresses).
• Need to make it clearer what VOSpace will achieve.
  • Feedback: Will enable e.g. remote cross-matching on scales which are presently inefficient or impossible. Also see UcUseDsaViaVoSpace

Report from Euro-VO SAC

Paolo: SAC will recommend to Euro-VO management to raise call to communicty for Science Verification proposals.

• Timescale?

Tighten up descriptions of science requirements as requested e.g. for fitting, SEDs

• Anita to provide short list of GOODS objects to be used as test cases for the SED builder. This should include at least one object with a well known and confirmed SED.
  • Two levels of precision; very high accuracy needed e.g. for photometric redshifts; few percent OK for radio-through-X-ray SEDs.
  • Must provide accuracy estimates/error bars in any case
  • Need IVOA standard model for describing photometric calibration
• Science Team in the future provides more specific, detailed science use cases for maturing tools.

New/extended requirements

• Seems to be useful to provide more local workshops for potential VO users.
  • Large national/international schools probably require Infrastructure project to commence but a handful of VO scientists and developers can organise workshops for one or a few nearby institutes.
• Should provide sample 1-line acknowledgement prominently at many stages of VO tool use
• In view of the progress in Solar and time-variable phemomena and interest in the community, consider adding someone in field to science group.

Future events

• IVOA Moscow, Beijing
• JENAM/VO Science meeting Armenia August 2007

Writing report

Assignment of several paragraphs to authors

-- AndreasWicenec - 06 Sep 2006

-- AnitaRichards - 07 Sep 2006