In general terms, the eCCP1 project is investigating e-science for quantum chemistry. More specifically, the project aims to study the requirements for enabling effective use of Grid resources by the quantum chemistry community, considering:
- Grid middleware, such as the Globus Toolkit and EGEE
- Grid resources, including Condor Pools, JISC clusters and HPCx
- Client tools (developed within the project and based on middleware)
- Current approaches to standards in data and command representation, such as the Chemical Markup Language
The emergence of Grid technologies has provided a generalised framework for the interoperability of computational codes. However, such interoperability is hindered by a lack of data format standards. The eCCP1 project tracks and develops, where necessary, emerging standards in computational chemistry data and command representation, with the aim of producing a common format for quantum chemistry data. XML has been chosen as the meta-language, principally due to its wide spread adoption and the availability of mature tools. This makes an XML format easily interpretable by disparate groups of collaborators. The requirements for the data-model are currently being gathered from the international quantum chemistry community. This is via project meetings such as the NeSC meetings ‘Toward a Common Data and Command Representation for Quantum Chemistry’ and ‘CCP/e-Science Workshop’, by the eCCP TWiki site (http://grids.ac.uk/eccp
) and email list (email@example.com
). Formatted data will be supported by meta-data, specified in a meta-data schema, providing information including the pedigree and purpose of data. Consideration will also be given to the development of an ontology for the domain, specifying relationships between formatted data-types. Such formatted data with supporting meta-data and ontology has advantages outside those of code interoperability. These include making data available to wider community and promoting appropriate data reuse. It also reduces the problems associated with the error prone process of converting between many different file formats.
The project aims to develop tools that can be deployed and later customised by CCP1 code developers. This will include the development of GUI tools (such as the CCP1 GUI) that can be used with a range of different CCP1 codes and implement the Grid functionality. This development involves investigating existing and developing new XML tools that can be used to manipulate the formatted data. This would include consideration of languages (for APIs), interfaces (e.g. SAX/DOM/Mixed) and functionality. Such tools would be responsible for interacting with data and meta-data repositories, which could include structural and basis set libraries.
The current initial aim is to produce a reference implementation of the above, with tools available at NeSCForge (http://forge.nesc.ac.uk/projects/eccp
) and data, meta-data and semantic models available at the eCCP TWiki site (http://grids.ac.uk/eccp
- 05 May 2004
Topic revision: r1 - 05 May 2004 - 13:04:00 - PhilipCouch