ASTERICS DADI Technology Forum 1 Hack-a-thon page

Traditionally Tech Forum meetings have included a free-form session designed to allow participants - be they software engineers or scientists - to meet and discuss technical and scientific details related to VO work they are currently undertaking. These sessions became known as the “Hack-a-thon” - the clue to their purpose being in the name. Hack-a-thons proved extremely popular and successful and we plan to continue them in the current project; hopefully with the wider remit they will be even more successful.

A white board or log will be available for people to note down areas they wish to discuss and the session itself will provide place and opportunity for informal meetings.

This page to be used to record outcomes and conclusions.

Topics

• Inclusion of the IVOA Registry into EUDAT B2FIND. A prototype is available http://b2find.eudat.eu/group/ivoa
  • status and discussion with interested colleagues, in particular ESAC
  • other information (domain specific, etc.) to include

André Schaaff, Christophe Arviset

• Provenance discussion (…):
  • Provenance for complex data sets exposing their progenitors : CTA Example
  • Provenance for other use-cases : code attached to data
  • other use-cases ?

please register here , just to let us now how big the room should be to accomodate this small group

François Bonnarel, Mireille Louys, Catherine Boisson, Mathieu Servillat, Paul Sladen, 

Notes :

• Reasons for developping provenance
  • provide credit for data producers and then get more producers involved
  • integrity checking
  • example PGP signature
• CTA pipe line context
  • an overview of DL data products
  • various segregations criteria are applied from DL2 to DL3
  • telescope congiguration : how many active telescope for one DL4 or DL3
  • reprocessing
  • aid to debugging

to do :

• explore more the Agent concept (a person, a service at one particular data center)
  • need to sort out the pipeline information to expose to the final user from the archive pipeline description

• MOC and PgSphere
  • Demonstration of CDS' MocServer
    • 15k of MOCs, 0.45 GB
    • matching of user-supplied MOC to all 15k MOCs: few ms
    • implemented in Java
    • all queries via HTTP
  • Maybe indexing of MOCs (as objects of their own right) not necessary in view of the above speed
  • further discussions about indexing
  • pointers to existing Java / C code that approximates polygons as MOCs (and vice versa) exchanged
  • MOC to be put into Pgsphere as sort of “BLOB” data type, just akin to Spolygon (PostgreSQL max size for TOAST data types: 1 GB)

Pierre Fernique, Markus Nullmeier, Gilles Landais, François-Xavier Pineau, Jesús Salgado, ...

• Moving the code to the data - discussion of ideas
  • Use cases - why
  • Use cases - what
  • Technicals - how
  • Standards - ?

Dave Morris, André Schaaff, Marco Molinaro, Thomas Boch

Discussion on how to describe as VO resources computational resources like VM, middleware environments, libraries and applications (in an Everything as a Service scenario). Use cases started from an INAF/CADC collaboration, LSST and GAIA use cases.

Tasks towards such a goal require: a survey of the most used code languages and cloud technologies; a check on whether exist already description models for such resources to be re-used inside a registry system for discovery; a resource extension for computational resources; a protocol to access those resources programmatically.

Main IVOA WG interested seems to be GWS, with connections to DAL, Apps and Registry.

A slightly richer report on this hackathon can be found in this attachment.

• Afternoon: SimDAL

Carlos Rodrigo, Marco Molinaro, François Bonnarel

Discussion about time slots and presentations to be reserved for SimDAL in the Sydney interop DAL sessions. Given the few Theory IG participants in Sydney it could be that only a review talk for SimDAL will be provided.

• TAP example
  • Can someone work with me (Mark Taylor) to come up with a short but compelling example of how to use TOPCAT's TAP client to do something scientifically useful? I understand the technology but I don't have enough astronomy background to prepare a good example.
  • Result: Ada showed me the basis for an example and will mail me a summary later

Mark Taylor, Ada Nebot, Catherine Boisson

• DOIs for astronomical data: will be discussed during the IVOA Data Curation & Preservation IG meeting in Sydney. Topic of interest for many groups and agencies. Also DataCite meeting before the Paris RDA Plenary on 21 September 2015 https://www.eventbrite.com/e/persistent-identifiers-enabling-services-for-data-intensive-research-tickets-17500184523

Francoise Genova