DS5 Planning Stage 01

The DS5 Stage01 Planning meeting has been held on Mon 21st Feb 2005 in Leicester.

Participants

Alasdair Allan, Elizabeth Auden, Thomas Boch, Sebastien Derriere, Markus Dolensky, Paul Harrison, Marco Leoni, Tony Linde, Keith Noddle, Andrea Preite-Martinez, Anita Richards, Jonathan Tedds

Agenda

10:00	Coffee
10:30	Introduction/Scope	SebastienDerriere	.pdf
11:00	Planning for Stage01	All
-	Suggested use cases	SebastienDerriere	.pdf
-	Convergence of registries, Data Models and protocols	AnitaRichards	.ppt, .sxi
-	Discussion, Short Term Planning	All
12:30	Lunch
13:30	Scope & Outlook to Iterations 2-6	All
15:00	Summarize → Work plan	SebastienDerriere+All
15:30	Tea & Coffee, Adjourn

Notes

Sebastien's Intro

• 72 pm CDS, 36 pm U.Edin, U.Leics, INAF, ESO ==> 216 pm
• Scope:
  • feasibility study
  • component designs and trial implementations
  • technology watch
• wiki, forum, rss feed
• Kickoff meeting
  • technologies: ontologies, semantic web, (xml, daml/oil, OWL)
  • support: W3C, commercial companies
  • tools: Protege, OilEd
  • astronomy:
    • VORegistry: map to RDF?
    • VOTable
    • UCD & IAU thesaurus
  • applications:
    • core ontology: IAU thesaurus, UCD, FactGuru (astronomy ontology, converted to DAML+OIL in 2003 - astronomyFactGuru7.daml)
    • specific task ontologies
    • registry exploration
    • generic queries
    • compute measurements
    • query against virtual columns
      • Starlink AST library does unit/freq conversion
    • broker agents as discoverers: from resource & broker agents
    • resource investigation language?
    • workflow-building investigators
    • sky-oriented resource discovery (PFO presentation at Kickoff)
      • lots of discussion

Ontology discussion

• will need to build our own: not much available
• FactGuru not really oriented to professional astronomy
  • Elizabeth has converted this to DAML/OIL
  • could import into Protege & convert to OWL for first-pass ontology to play with
• CDS working with groups in Toulouse & Nancy
  • mostly textual mining
  • have looked to develop ontology from IAU thesaurus
• AstroGrid has links to bio-informatics people who've done a lot
• issue of:
  • differences of opinion over classification
  • vagueness of specification (eg V-magnitude)

Anita's talk

• SIAP, SSAP, Registries, Data Models: the need for Convergence
• dynamic data discovery
• SIAP problems
• can't do everything with headers
• RadioNet members' archives
• IVOA Registry: quality
  • what is bad data - can it be defined? not for VO to judge?
  • key issue is badly described data
• AstroGrid portal use of Registries
  • Curation and column metadata incl. UCDs
  • Help with query construction
  • Descriptions of tools
  • Use in building workflows
• UCDs very important
• use of ontologies: automatic harvesting, learning
• Vizier tool for automated creation of UCDs
  • would be useful if available on VObs
• data models
  • characterisation most relevant
  • basis for extracting virtual data
  • work with data providers
  • what does ALMA need?
  • INT-WFS - observing log info saves a lot f trouble
• Registry is backbone for metadata

Sebastien: CDS proposals for Stage01

• Finalize UCD1+
  • access to Vizier implementation
  • tools: validation, explanation, translation
    • available via WS
• Ontology of object types
  • astro object types (star, B-star, ...)
  • reuse: SIMBAD (100 categories, no multiple links), IAU thesaurus
    • eg low-mass xray binaries
  • apps:
    • classification, eg Registry searching app
    • connection to UCD for associated quantities
• SED construction
  • identify resources
  • automatically retrieve/compute
  • possibility of automating B. Vollmer tool
• Compute quantities
  • based on UCD ontology
  • description of possible measurements
• Planned support
  • contract not signed

Discussion

• EA: VOEvent people
  • AA: Current proposals have some problems, also duplicates work done on RTML
    • meeting coming up in Caltech in April
    • meeting coming up in Exeter in July

Partner assignments

Work

• see Sebastien talk for CDS
• INAF: APM DS5 time working with CDS
• AstroGrid: mainly evaluation of existing tools & capabilities
  • have ECA 50% time on DS5 + 100% new resource to be hired at Leics
  • other resources tbd
• ESO:
  • work on SAADA tool: reads FITS headers and populates database
    • useful for VObs

Personnel

• CDS: (contact is Sebastien)
  • Sebastien: 50%
  • Thomas: 25%
  • Bernd Vollmer: 10%
  • waiting for new hires: 1 to work with Bernd
• AstroGrid: (contact is Elizabeth)
  • Elizabeth: 50%
  • new hire at Leics: 100%
  • various people: science advisory: 50% in total
• ESO: (contact is Marco)
  • Marco: 50%
  • new hire: 50%
• INAF: (contact is Andrea)
  • Andrea: 50%
  • new hire: 100%

Assignments

• AstroGrid:
  • EA report on:
    • technologies
    • what other fields (eg bioinformatics) do / other projects
    • how this relates to real world
      • eg astronomer language interface
      • updating ontologies
      • how to reflect different terminologies, diffs of opinion
    • ontology visualisation tools
  • AA report on:
    • what agent infrastructure requires from DS5 (& DS3)
• ESO:
  • evaluate tool for creating metadata content : channel metadata to VObs
    • from FITS headers
      • using SAADA: Laurent
    • incoming data from PIs: external surveys
  • create report on experience
• CDS & INAF:
  • UCD handling tools (all WS)
    • UCD validation tool: from UCD or VOTable
    • from description, get UCD
    • translation from UCD1 to UCD1+
    • input from RadioNet on extraction of data - see RadioNet archives & VOs forum
  • ontology of astro object types (with INAF )
    • has relationship 'has-a-ucd'
    • app for querying registry based on object types (eg equivalents)
  • begin SED construction
    • plan for future work
    • includes ideas for quantity computation

Stage02 ideas

• build on objects ontology (CDS / INAF)
  • how many relationships do we need?
• SED construction tool (CDS)
• create events ontology (AstroGrid)
  • investigate need for rules & reasoning
• creation of catalogues from ontology
  • dynamic based on queries?
• science scenarios / use cases (ESO)
• use of ontologies in workflow (AstroGrid)
  • links with DS3
  • Apply relationship or rule tool to results of queries

Interactions with other DS areas

• DS3:
  • workflow
  • deployment of:
    • tools for creating / validating UCDs
    • tools for creating metadata content
• DS4:
  • mining of registry
• DS6:
  • visualisation ontologies

Interactions with IVOA workgroups

One person from this group will track activities in each IVOA workgroup which is related to DS5 and report back to the DS5 group on work which might be done, possible collaborations etc.

• Registry: Tony
• Data Models: Anita
• UCDs: Andrea
• VOEvents: Alasdair
• VOQL: Marco

Discussion

• tech report should investigate language libraries
  • may mandate specific language libs later
• no mandatory programming language
  • but services should be Web Services
• must use Protege as ontology editor
• ontologies must be produced in OWL

DS5 Meeting 01 Actions

Number	Person	Action	Due Date
ds5.m01.01	Tony	Agree license for software releases with Board	30-June
ds5.m01.02	Sebastien	Produce Stage01 plan for TAP meeting	04-March

DS5 Stage 01 Assignments

Number	Person	Partner	Assignment	Deliverable	Due Date
ds5.s01.01	SebastienDerriere	CDS	Progress report for end of stage	Report	before next planning mtg