3.7 Visualisation

3.7.1 Vision

At present, there is no unified visualisation software system for solar system physics. Although many plotting tools exist, these tend to be specific to particular missions or research groups. Since many of these packages use the same languages (e.g IDL) and the similar basic routines, there is great scope for duplication of effort. It is proposed that advanced visualisation software should be at the heart of the UKSSVO. Users would interact with the Virtual Observatory via a visual interface at every stage from data selection to production of publication quality graphics. Good visualisation capabilities will be central, both to data selection and to manipulation and combination of data sets using UKSSVO science tools. High-quality graphical representations of measured data and scientific results will also enhance the accessibility of the project to the non-scientist, and will therefore also be important in outreach terms.

The development of visualisation capabilities for the UKSSVO will be strongly informed by our vision of the Solar System as a single time-dependent three-dimensional system. The basic radial co-ordinate systems are common to solar, planetary and STP work, the only difference being the choice of origin. Using co-ordinate transformation tools within the VO, different data sets will be easily transformed between such co-ordinate systems and integrated into a single solar system reference frame. The system will be designed to allow the user to move easily between the visual and query interfaces, in order to refine search results and manipulate data products.

In order to develop the visualisation interface for the UKSSVO, existing graphical capabilities (such as QSAS and the UCL CTIP visualisation environment, CVE) need to be brought together and adapted to interface with the data handling software, search engines etc supported by the Virtual Observatory. Appropriate connections with metadata will also have to be established in order to define what types of image manipulation are possible for each of the various data sets.

Current development work for Astrogrid already includes work on visualisation standards and techniques. This development of visualisation systems for the UKSSVO would be carried out in close collaboration with the developers of Astrogrid, drawing on their experience and expertise, with the aim of developing a user-friendly and intuitive system that actively facilitates advanced data analysis.

3.7.2 Extension of current tools

Relatively simple facilities will be delivered by adapting existing plotting tools and packages to work with the SSVO data objects. In the first instance, the Treeview application from Starlink will be extended to allow simple plotting of solar system data formats, in parallel with the development of the HDX capabilities for handling these formats. Simple plotting will be provided by a plotting engine based on QSAS components or CDAWlib. In addition, we will work with existing packages such as QSAS and SolarSoft to extend their capabilities so that they can work with the SSVO (see WP610 and 620).

3.7.3 Grid visualisation using GAPtk

It is proposed to begin the development the advanced visualisation system outlined above using the Grid Applications Portals toolkit (GAPtk) developed by the Grid Visualisation Group in the RAL e-Science Centre (see WP360). The facilities of this toolkit will be combined with the elements of the existing domain-specific visualisation tools in QSAS and CVE.

The primary goal of the e-Science Centre visualisation activities is to build a set of generic and scalable Grid and Web based advanced visualisation services for data analysis and to provide a high level application programming interface (API) to embed these visualisation services within any problem solving environment. The Grid visualisation toolkit, GAPtk, is designed to meet this goal and has already been applied to the NERC funded GODIVA project and the Excitations Visualisation Environment (EVE) for the real-time visualisation and coupled experimental simulation of condensed matter physics studies.

The GAPtk visualisation services are extensive and extend from a single dimensional visualisation to a multi-dimensional complex visualisation scenarios choreographed by combining individual visualisation services. Examples of these services include the generation of Grid enabled isosurface generation (for real-time rendering and scalability), slices and animations created on the fly with play back using a distributed architecture.

The GAPtk toolkit comprises: generic visualisation services; a sample set of common data analysis and simulation application utilities; and the API that enables scientists to compose or assemble more complex analysis scenarios from the services. At the heart of the toolkit is a central application visualisation server, which mediates between a variety of desktop applications (custom software), problem solving environments (e.g. MATLAB, IDL) and portals (browser based interfaces) and a variety of third party data access services and the Grid fabric layer. It adds valuable intelligence to the communication layer and hides the complexities of distributed computing. The server also contains some generic application and visualisation services, for instance, server- side rendering of publication quality images or generating an animation sequence from the geometry computed. The server also contains software modules that coordinate the input and output from various external services, adding context sensitive semantic information to the outputs before sending these to the clients.

The GAPtk server has three interfaces. The server's client interface uses a Web services communication protocol, except for very large datasets. The server's data interface to third-party data query and search Web services again uses Web services communication mechanisms. A third server interface communicates with the Grid fabric layer to obtain secure data and compute resources using appropriate Grid and Web protocols such as GridFTP and SOAP.

Generic application and visualisation services are implemented as atomic components within the server which an application developer can then use to compose a complex visualisation scene. This provides an efficient way to implement flexible higher order data visualisations as the user is able to compose and superpose the variables of his problem space within a visualisation scenario. For example, it supports the variability and scaling required in each data dimension as appropriate for a given application context to make the complex visualisation scenario semantically correct, intuitive and easy to explore. This allows an application programmer to create geometry with superposed data structures that will render these with appropriate scaling and behaviour at the client end. Work is also continuing on Grid-enabling basic visualisation algorithms to support near real-time data exploration.

Currently there are three separate high level scripting interfaces under active development to allow applications access to the services and interfaces of the GAPtk server. Using these, it is possible to construct customised task-based user interfaces on the client-side desktop, for a wide range of applications using a variety of problem solving environments (PSE). This feature helps users to build on existing knowledge and investment in application specific problem solving environments and also supports users who wish to develop their own more advanced Grid-aware environments, for instance for distributed, collaborative computational or experimental steering.

There are two other server-side APIs allow new Web services, applications and Grid fabric layers to be linked in. Where an application is already either Grid-enabled or parallelised, a simple services based wrapper enables it to be linked into and made available via the GAPtk framework.

Staff at the CCLRC e-Science Centre at RAL have considerable experience of these techniques, and the proposal calls for their involvement at a level of approximately 2 staff years per year in order to make these facilities available within the UKSSVO (see Appendix B).