Some slightly sketchy notes of Paul Walk’s talk
Paul says: the real challenges are around:
Business case
IPR
etc.
Technical issues not trivial, but insignificant compared to other challenges
We aren’t building a system here – but thinking about an environment … although probably will need to build systems on top of this at some point
‘The purple triangle of interoperability’!:
- Shared Principles
- Technical Standards
- Community/Domain Conventions
Standards are not the whole story
- Use (open) technical standards
- Require standards only where necessary
- Avoid pushing standards to create adoption
- Establish and understand high-level principles and ‘explain the working out’ – support deeper understanding
Paul suggests some ‘safe bets’ in terms of approaches/principles:
- Use Resource Oriented Architecture
- identify persistently – global and public identities to your high-order entities (e.g. metadata records, actual resources)
- URLs (or http URIs) is a sensible default for us (although not the only game in town)
- use HTTP and REST
Aggregation is a corner-stone of RDTF vision – so make your resources a target for aggregation:
- use persistent identifiers for everything
- adopt appropriate licensing
- ‘Share alike’ maybe easier than ‘attribution’
Paul still a little sceptical of ‘Linked Data’ – it’s been the future for a long time. Tools for Linked Data still not good enough – can be real challenge for developers. However, we should be a
Quote Tom Coates: “Build for normal users, developers and machines” – and if possible, build the same interface for all three [hint, a simple dump of RDF data isn’t going to achieve this!]
Expect and enable users to filter – give them ‘feeds’ (e.g. RSS/Atom) – concentrate on making your resources available
Paul sees slight risk we embrace ‘open at the expense of ‘usability’ – being open is an important first step – but need to invest in making things useful and usable
Developer friendly formats:
- XML has a lot going for it, but also quite a few issues
- well understood
- lots of tools available
- validation is a pain
- very verbose
- not everything is a tree
- JSON has gained rapid adoption
- less verbose – simple
- enables client side manipulation
Character encodings – huge number of XML records from UK IRs are invalid do to character encoding issues
Technical Foundations:
- Work going on now – Will be a website ETA June 2011
- JISC Observatory will gather evidence of ‘good use’ of technical standards etc
- Need to understand federated aggregation better
Questions for data providers:Do you want to provide a data service, or just provide data?