Post Meeting Support

There will be a limited amount of support available after the workshop for those participants wishing to further develop the work that they have created during the workshop. Details are posted here.

Feedback from the workgroups:

During the meeting a number of key issues were highlighted - to be addressed:

group A comments:

• good
  • interoperability between data and applications within one environment
  • ability to run a model with data from a variety of sources
  • not having to re-invent the wheel
  • quickly were able to make some more remote data available (e.g. DST from RAL)
• bad
  • error reporting when jobs failed - debug messages
  • interface (user friendliness) to expose applications (scripting) (reduce load on data providers to plug in apps
  • data formats (need to include STP demands > IVOA input for instance in STP)

group B comments:

• good
  • workflow is a great idea
  • ease with which it was possible to integarte a major package (e.g. R)
• bad
  • lack of ability to see myspace of other people (allow groups)
  • portal is difficult to use
  • workbench is not yet fully developed ('will come come back to look in 6 months')
  • handling bulk data flows (it didn't take long to produce GB size files)
  • need to improve reliabilty and speed

group C

• good
  • potential to share workflows with others (bad bit is that can't do that just yet)
  • even now there is a reasonable range of data available: usage will increase with more data being included
  • helper apps such as Topcat very useful - interacting with MySpace
  • the cross match app was very powerful

• bad
  • limits on returns from database queries (10,000 row limit is too small)
  • lack of sensible defaults for applications in the workflow
  • shortcuts to allow 'quick' queries - a one step query

group D

• good
  • actually got to do some science - analysis was quicker - even now with certain science topics astrogrid access to data speeds processing up
  • provide access to 'workflow' libraries and examples of script.

• bad
  • much remains to be done to make the system more 'user friendly'
    • web interface
    • reduce 'number of key strokes'
    • make it more 'ASTRONOMER' friendly - need astro-speak rather than compu-speak
  • track 'history' of what the astronomer is doing

Next steps

Group A Next

• Deploy datasets of UKSSDC via DSA.
• Complete CEA installation at RAL, so that CTIP model can be run as a CEA app there.
• Fix generic movie maker.
• Join all steps together into 2 workflows: Workflow 1: retrieves UKSSDC data, calls plotting program with sliding window, generates mpeg movie. Workflow 2: retrieves UKSSDC data, runs CTIP model, runs CTIPmovie application to generate a movie of output data.

Group B Next

• integrate swarp pre sextractor
• visualise in R results from the redshift maker

Group C Next

• a new service to automate astrometric fitting
• bring in astroscope to allow return of catalogues with data at a positional search
• photometric transformation
• how to i cross correlate small cats against all sky cats?

Group D Next

• user notification of completeion of (large) tasks - by SMS, email?
• need wider data access (2df, spitzer image cats, MACHO, OGLE(2))
• cross correlations to determine radio/ir matches

Format of the meeting and final comments

• the portals remain available
• please add anything else that you think to these pages

• wiki to capture examples of use, workflows etc
• working together in solving deployments - encourage sharing of solutions (e.g. tricks an scripts)

• format of the meeting?
  • tech people at the meeting provided good (instance) support
  • good to mix the tech and sci staff - having a time limit of fairly intense activity, team vs team was productive
  • it was good to mix science backgrounds (see use of tools across sci areas)
  • maybe have 'tool' specific themes
  • but something like this would be useful again next

Any feedback - let me know - naw at ast.cam.ac.uk