Submissions to VOTech Kickoff meeting

18-19 Nov, 2004, Cambridge, UK

Here are all the submissions for the VOTechKickoff meeting plus links to papers (wiki pages) that more fully explain the topic under submission (when complete, I'll package all these papers into a pdf file).

Presentation Submissions

Paper-only Submissions

---

Abstracts DS1/2

Introduction to VOTech (Andy Lawrence)

I will summarise the aims of the VOTECH project, its place within the EuroVO vision, and will try to set some goals for our meeting.

Project Management Approach (Tony Linde)

I will outline how the VOTech project will be managed: the personnel and groups involved; how decisions will be taken and communicated; how code will be developed, stored and released; and the means of communication - web sites, wikis and mailing lists.

VOTech Science Requirements Approach (Nic Walton)

This presentation will set forth current ideas for capturing science requirements to inform activities across the programme. Note will be made of the mechanisms by which the project will be scientifically steered, through the formation of an external VOTech Science Group and an internal science contact group with representatives from the individual DS areas. Finally note will be made of the science interactions between the three main components of the embryonic EuroVO project.

Further Science Requirements from the AVO (Paolo Padovani)

The AVO project provides two further ways to collect science requirements for the VOTech project: input from the Science Working Group which went beyond the AVO mandate and the AVO Science Reference Mission. I will briefly discuss both of these.

AVO: Lessons Learned in Technical Coordination (Markus Dolensky)

This is an experience report on the technical coordination in the AVO project. It is about methods, technical infrastructure and specific aspects of such a networking activity.

---

Abstracts DS3-6

Overview of Shibboleth (Alan Robiette)

JISC has taken the decision to implement the Shibboleth technology, developed by the Internet2 organisation in the United States, as the basis of the next generation of access management services in UK education and research. This session will discuss what Shibboleth is, how it works, and how it is being enhanced to provide support for virtual organisations. The intention is to provide sufficient background to allow VOTech to assess its suitability as a possibility for the project's authentication/authorisation infrastructure.

Agent architectures: Smarter software for astronomers (Alasdair Allan)

In the real world, people rarely interact with other people, or data, in a hierarchical manner. If they do, they lose time and efficiency.

It's not necessary to have an global view of topology of the data, so long as you know about some of it, and know other people who know the rest. The emerging VO architecture can leverage the peer-to-peer nature of agent based solutions to make the software smarter, more efficient and more responsive to the user. The purpose of an agent is to know what a user wants, even if they don't know themselves. Establishing architectures and building agent based solutions for the VO will flatten the network topology, enabling links between related datasets to form more quickly, and make the resulting network more robust.

Access to virtual data (Anita Richards & Simon Garrington)

Astronomical data are often archived in forms which need further processing before they can be analysed or compared with other data, for example the application of quality-based flags to ISO spectra. In many cases different users need different products, such as stacked or single HST images, different resolutions for interferometry data. This can be achieved if data providers support standard data models and tools, with help from VOs (such as via Opticon andRadioNet).

Tree Codes and their Application to Fast and Efficient Algorithms (Bob Nichol)

I will review multi-resolution tree codes and their use in building complex astronomical data-mining applications.

How to Adapt existing Archives to VO: The ISO and XMM-Newton cases (Christophe Arviset)

Adapting existing archives to VO is an easy task if the archives have been properly engineered. There is no real need to change any already existing data structure, nor in your metadata databases, neither in your data repositories. Translation layers should do the work.

We present the ISO and XMM-Newton cases, where we adapted our existing archives to certain VO protocols.

A Column-Oriented Data Explorer (Clive Page)

Astronomical data mining will mostly involve tabular datasets; data exploration and cleaning operations need to be carried out interactively.

Relational DBMS are very inefficient for many of these operations, especially those involving sequential scans of large tables. Performance can be improved by one or two orders of magnitude using techniques such as column-oriented storage and parallel execution, which makes all the difference between batch and interactive working. The features of a proposed system will be outlined, and a basic prototype demonstrated.

The AstroGrid 'Solar Movie Maker' Application (Elizabeth Auden & Silvia Dalla)

AstroGrid's solar science case demonstrates the use of Astrogrid infrastructure to search for solar images, analyze data with Solarsoft, and generate MPEG movies. The IDL Solarsoft package provides routines to retrieve raw images at the required wavelength and resolution, calibrate them and correct for solar rotation taking place between images.

The movie maker tool operates on the data as follows:

• A unix shell script generates a non-interactive IDL session that applies instrument-specific Solarsoft algorithms to the solar images before amalgamating them into a single MPEG file.
• A local installation of the AstroGrid CEA software is configured to deploy the tool as a web service.
• Finally, the tool is registered with AstroGrid, allowing interaction with remote registry searches, user-defined workflows, job schedulers, and the MySpace virtual storage area.

An Agent-based Predictive Engine for Variable Star Observing (Eric Saunders)

We consider the problem of designing and implementing an autonomous intelligent agent that can make the most efficient use of the finite amount of telescope time available to an observer. Such an agent must be able to robustly quantify how well an observing schedule is proceeding and thus make changes, in real time, to optimise such a schedule. We present a system of metrics that will form the predictive algorithmic engine for the eSTAR intelligent agents. We discuss the broader applicability to the field of variable star observation.

Grid-enabled access to databases: current activities (Fabio Pasian)

Harmonisation of the VO infrastructure and user tools with the developments being carried out within the various national and European Grid projects is necessary.

The Italian astronomical community is active in the process of making scientific applications Grid-aware. One of such applications (production of simulated data from the Planck mission) has been proposed to EGEE for the use of its infrastructure. Within these projects, high priority is being given to the proper integration of archives and databases with the Grid.

This presentation briefly describes the solutions currently found or being pursued in this area and their applicability to the VO.

Visualisation and processing of multiband images (Francois Bonnarel & Mireille Louys)

We'll present our projects in the field of Multiband Image Analysis. We propose to explore different methods for data set reduction, segmentation and visualisation. We emphasize the strong need for astronomical validation.

Interoperability Standards (Francoise Genova, Francois Bonnarel, Pierre Fernique, Andre Schaaff)

I will recall the main topics for development of interoperability standards dicussed in the IVOA framework, and give a few examples of topics on which further development and experimentation are needed, to deal with the VO heterogeneity and complexity.

A Neural approach to clustering and visualization in multiparametric spaces (Giuseppe Longo)

The talk will present some recent tools implemented by the AstroNeural collaboration (probabilistic Principal surfaces, ensemble classifiers, competitive evolution on data) which seem to offer great advantages with respect to more traditional visualization and data mining tools. The Astroneural tools have so far found an application in several fields of activity: from geophysics, to bioinformatics to astroparticle physics.

VObs Service-to-Service Security (Guy Rixon)

Requirements are summarized for authorization, authentication and accounting in the use of VO services. Some effects of these requirements on AAA frameworks are noted.

Hypercell - Visualization of Multivariate Data (Ken Brodlie)

The talk will describe an approach to the visualization of multivariate data, by exploring in an organised fashion a sequence of low-dimensional subspaces. The concept has been implemented as a set of modules in IRIS Explorer. Although the tool is quite general in application, we have applied it in collaboration with Bob Mann from Edinburgh University to visualize a data table of observations of attributes taken from the SuperCOSMOS Science Archive. The talk will also mention work at Leeds in building visualization web services.

Real Applications for the Virtual Observatory (Malcolm Currie)

This paper outlines the capabilities of Starlink software and describes how Starlink has been preparing for the Virtual Observatory, ready to provide real applications for the VO.

Starlink table/VOTable handling software (Mark Taylor)

The Starlink Tables Infrastructure Library (STIL) is an open-source, pure Java library for multi-format input, output and processing of tabular data. It is designed to be efficient and scalable, and as far as we know provides the only full implementation of the VOTable format. We therefore suggest its use for VO-Tech components which need to read or write VOTables. Starlink's TOPCAT is an example of a GUI program based on STIL.

In addition, we explain the format-neutral, data model-centric thinking behind STIL's architecture (also used elsewhere in the Starlink Java codeset), and discuss it as a productive way to write data-handling software in a heterogeneous environment such as the VO.

Intelligent Access to Foreign Data (Norman Gray)

The obvious, and current, solution to the problem of exchanging data structures is to convert data into and out of a consensus model, but this is now labouring under the burdens of complication and often unhappy compromises. I wish to explore a complementary approach, which gives applications the ability to directly utilise foreign data models (or ontologies) using Semantic Web and AI technologies.

This uses relatively mainstream ideas from the Information Retrieval community, and functions at a high enough level that the results are potentially applicable, not just to the VO community, but to the HEP community also, only now waking up to the problems of semantic Interop.

Science-driven VOTech Tools (Paolo Padovani)

I will review the tools which had been suggested by the AVO Science Working Group but which went beyond the AVO mandate. I will also discuss some tools that VOTech WILL have to build to reach its goals.

Extending the CEA (Paul Harrison )

...

Sky oriented resource discovery (Patricio Ortiz)

Registries will host thousands of entries of resources containing data for billions of celestial objects. Identifying the resources with objects located in the vicinity of a list of targets is of primary importance as it is to avoid querying resources with negative resoults. A fine granularity method to represent a resource's sky coverage is presented. This method reduces the number of null-return queries from about 400% to less than 10% the true number of resources producing matches. This gain means a reduction in bandwidth and resource usage. Algorithms and transport methods are available to make this another registry resource.

celFilter: a sky oriented resource discovery (Patricio Ortiz)

Registries will host thousands of entries of resources containing data for billions of celestial objects. Identifying the resources with objects located in the vicinity of a list of targets is of primary importance as it is to avoid querying resources with negative resoults.

A fine granularity method to represent a resource's sky coverage is presented. This method reduces the number of null-return queries from about 400% to less than 10% the true number of resources producing matches. This gain means a reduction in bandwidth and resource usage.

Algorithms and transport methods are available to make this another registry resource.

VOSpec: A tool for handling Virtual Observatory compliant Spectra (Pedro Osuna)

The VOSpec tool can superimpose spectra coming from VO-compatible resources that conform to the VO Simple Spectral Access Protocol (SSAP) with three added extra fields in the VOTable through the use of an algorithm based on dimensional analysis.

Towards a resource discovery engine? (Riccardo Smareglia)

...

Ontologies (Sebastien Derriere)

Ontologies and Semantic Web are very active research fields. Applications of the techniques and standards developed in those fields should enable new capabilities for resource discovery, and data/metadata sharing.

I will discuss how the VO could leverage on those technologies and already existing VO standards (Registry, UCD, VOQL) to develop new components, allowing intelligent resource discovery for astronomers and software agents.

When are Agents good data miners? (Tim Naylor)

By exploiting the emerging VO architecture, agent technologies can offer interesting data mining capabilities. We describe the collaborative agent model eSTAR has used to interact with grid enabled telescopes, and how this model can be applied to more general data mining problems. We discuss use cases where this model has a clear advantage.

The Possibilities of Virtual Storage (Tony Linde)

AstroGrid's MySpace component provides a user with virtual storage facilities for file objects, most typically workflows developed using the AstroGrid portal and VOTables returned from queries on remote datasets. Although the files may be stored in any number of locations around the world, the user has a view of them as a single nested set of directories (folders and containers are other terms used). The AstroGrid-2 project will extend the functionality offered by MySpace and will, in conjunction with the NVO MyDB project, offer to store database tables: this will create the VOSpace vision under discussion in IVOA quarters.

In this talk I will offer areas in which VOTech might conduct research that will extend the functionality of VOSpace. Some ideas are:

• to allow the user to upload scripts or code and have them run remotely or as part of a workflow
• to extend VOSpace to the user's own computer disk space
• to interface to Grid datastores

Publishing Ideas and Arguments (Tony Linde)

looking at the ScholOnto project for ways that we might represent annotations or comments relating to resources

A Grid-enabled Theoretical Data Archive in the VO framework (Ugo Becciani)

Many theoretical researchers use CINECA high performance computing resources in the framework of the agreement between CINECA and the National Institute for Astrophysics (INAF). In the agreement the MPP Cineca system is used and more than 100 Projects have been approved in the last two years. The produced data must be made available to the community. Other institutions (e.g. Catania Astrophysical Observatory) own a local MPP system and produce data from simulation that can be also made available to the community. We will propose a pilot project for a federation of theoretical data archives that will be developed following the standard of the VO and will use the grid to access the distributed data.

AstroMD: a VO enabled tool for Scientific Visualization and Data Analysis (Ugo Becciani)

AstroMD is a package for supporting the visualization and analysis of astrophysical three-dimensional data. It was developed in the Cosmo.Lab project (OACT, CINECA and other institutions), funded by the EU FP5. This tool gives a 3D graphic representation of data exploiting the most advanced visualization technology, and has several built-in tools which allow the user an efficient manipulation and analysis of data. We will integrate the AstroMD tool with VO services: connection to a VO web service, retrieval and dealing with data in the VOTable, FITS or ASCII formats. Data can be visualized using various techniques: points rendering, volume rendering, isosurfaces.

---

Dead Links

VOTechKickoffCall VOTechKickoffSubmissionsUpload