Water Quality Data Publishing Architecture
Not just components…. System behaviour –User needs –All users with a stake in success Governance –= User + Supplier Confidence Renewal and extension Use Cases Semantics (Information Architecture)
RM-ODP Tried and true methodology Viewpoints –Enterprise –Information –Computational (components and interfaces) –Engineering –Technology
CANRI Relatively easy – proven, don’t need to do much ad-hoc to build complete systems Also supported and framed by CANRI computational architecture Governance, Information, Engineering –Less mature, domain specific, needs work –regardless of computational platform
Robustness Build for the future –Musn’t break as new products or applications added –Anticipate extension, integration, functions –survive change of organisations/roles Distributable –All components may be moved, upgraded, rebuilt
Enterprise Viewpoint Emerging data standards Modular standards –Easier –Better –More stable –Some technology demands Lifecycle planning
Data standards need lifecycle planning –Initial implementation –Key harmonisation processes –Roadmap Review Gap analysis (what would we like to re-use?) Adopt/Adapt Engage with interested parties Versions Gateways/Deprecation Communication
Information Viewpoint Features Have properties With domains (allowable values) Each property –Name –Namespace –Definition –Lifecycle –Roadmap
Computational Viewpoint WFS is fine Can be made to support large compressed data if required (no streaming of gigabytes expected) WFS + –Decouple Query semantics from response –Change request to OGC –Powerful, safer, more predictable
Technical Viewpoint Geoserver –Used in SEEGrid to prove scientific data can be served via WFS using external data standards –Limited object models, but still better than GIS-only –Spatial and aspatial related data –Based on v1.2 – on a branch –Issues accepted in core roadmap –Needs refactoring of geotools libraries
Geoserver Strategy Harmonise with v1.3 Compare with Use Cases, Data model –Enterprise Viewpoint –Information Architecture –Engineering Considerations –Computational (fixed (WFS + query), phew!)