DS6: Stage 7 Plan

Introduction

For the main part, the DS6 plan for Stage07 centres on continuation of activities started in Stage06 (or before), but it also contains some tasks resulting from the Gap Analysis undertaken by the DS6 team during Stage06.

Stage07 Task List

Number	Title	Who
T01	Completion of testbed science scenario: Kernel Density Estimation	AstroGrid
T02	Server-side STILTS	AstroGrid
T03	Scripted access to TOPCAT graphics	AstroGrid
T04	GridSphere portlet interface to STILTS	AstroGrid
T05	Dissemination of recent standards	CDS
T06	Developing Characterisation 2	CDS
T07	Further development of SAMP/PLASTIC	CDS, AstroGrid
T08	Further development of SIA2 protocol	CDS
T09	Footprint standard	CDS
T10	Characterisation DM applications	CDS
T11	Develop modified version of MAF2.2	INAF
T12	Release VisIVO 1.5	INAF
T13	Development of VisIVOServer	INAF
T14	Improvement of VONeural GRID-launcher	INAF
T15	Deployment of further VONeural services	INAF

Stage 07 Task Descriptions

T01: Completion of testbed science scenario

This is a continuation of the work started in Stage05. The basic workflow for the project is working, but further development and testing is required to validate it against the previous work of Richards et al (2004) and then apply it to an expanded dataset. The ultimate aims are: (i) a paper publishable in an astronomy journal on the science; (ii) a comparison with the similar work being undertaken in Napoli with VO-neural, for inclusion in the DS6 Study Report; and (iii) a report on the problems encountered in running the project, possibly for inclusion in the DS6 Study Report, but definitely for internal consumption by the partners, and for posting on the DS6 wiki page.

T02: Server-side STILTS

Previous DS6 work by MarkTaylor has shown that, for certain types of query, STILTS running on a column-ordered flat-file dataset can be faster that a relational database (because astronomical catalogues tend to have wide tables). This task seeks to extend that work to see whether it is possible to devise a parser which takes an ADQL query, and works out whether it would be better to execute the query on the database or the flat-file version. An alternative approach, which may also be addressed, is to build covering indexes in the database for those columns (or combinations thereof) for which column-ordered flat-file copies would otherwise be made, and see whether the database query engine can work out that it can answer the query using those, rather than the parent table.

T03: Scripted access to TOPCAT graphics

This task was held over from Stage06, as MarkTaylor was not able to devote any time to it during that Stage. The task has two stages: (i) assess what can be done with a reasonable amount of effort to provide scripted access to some of the graphical functionality in TOPCAT, in the manner that STILTS does for the core data manipulation functionality; and (ii) undertake deployment and testing of such a server-side system at a data centre, probably ROE.

T04: GridSphere portlet access to STILTS

Data centres may wish to offer access to local services via their websites, to complement the access provided through desktop applications, such as Aladin and the AstroGrid VODesktop. Portals may offer a way of doing that, and this task is to assess whether this may be readily done using a technology called RAPID, which has been developed by the National e-Science Centre in Edinburgh. RAPID automatically builds GridSphere portlets from XML descriptions of tools (i.e input parameters, flags, etc), and it may be possible to generate these XML documents for CEA/UWS apps using relatively straightforward XSLT scripting. To assess this possibility, initial work will be undertaken to produce RAPID portlets for the CEA version of STILTS.

T05: Dissemination of recent standards

This heading includes a number of related activities all designed to increase the uptake of recently proposed standards, such as: release of a new Aladin server, together with CAMEA and a related Java library for Characterization; release of a reference implementation of a SAADA database with Characterization mapping; and discussion with ESAC and ESO regarding mappings between databases and the Characterization model.

T06: Developing Characterization 2

This task centres on paving the way towards a second stage Characterization model, based on the recognised limitations of and omissions from what has been proposed to date.

T07: Further development of SAMP/PLASTIC

This task includes both the development of the draft standard - for discussion at the Trieste Interop, with the goal of reaching Proposed Recommendation level - and of reference implementations of that standard.

T08: Further development of SIA2 protocol

The task here is to develop a working draft for discussion at Trieste and to write a test implementation within the Aladin server.

T09: Footprint standard

This will be a collaboration with Gretchen Greene to define a set of Use Cases for discussion at Trieste, based on experience with the ESO/CDS and JHU footprint service initiatives.

T10: Chracterization data model applications

Two application areas are envisaged: completion of prototyping of VOEvent/archive image matching, in collaboration with Elizabeth Auden; and using characterization of datasets in CDS workflows.

T11: Develop modified version of MAF2.2

A modified version of v2.2 of the Multimod Application Framework will be developed in Catania, and will form the basis for future versions of VisIVO which have the same reliability under Linux/Mac OS Z as under Windows.

T12: Release of VisIVO 1.5

This will provide enhanced support for 2D grids, dataset animation, Mac OS X support and further documentation.

T13: Development of VisIVOServer

The VisIVOServer will provide a command-line tool to run on the server side and which provides users with static 2D images of 3D datasets. One specific application of this approach will be to the SDSS database, in collaboration with Portsmouth.

T14: Improvement of VONeural GRID-launcher

Continuing work on the VONeural GRID-launcher will address how Grid resources can be employed for the execution of jobs generated by VONeural, in a way which is maximally interoperable with existing VO middleware.

T15: Deployment of further VONeural services

The on-going roll-out of VONeural services will continue with VONeural-PPS, a parallel version of VONeural_MLP and a pipeline prototype.