220 likes | 339 Views
A Prototype Spatial Object Transfer Format (SOTF). Peter Woodsford ( peterw@lsl.co.uk) Laser-Scan Ltd., Cambridge, UK. www.laser-scan.com 6th EC-GI & GIS Workshop, Lyon, France, 28-30 June 2000. SOTF - Introduction. Many agencies now handle GeoInfo as Objects
E N D
A Prototype Spatial Object Transfer Format (SOTF) Peter Woodsford (peterw@lsl.co.uk) Laser-Scan Ltd., Cambridge, UK. www.laser-scan.com 6th EC-GI & GIS Workshop, Lyon, France, 28-30 June 2000.
SOTF - Introduction • Many agencies now handle GeoInfo as Objects • Spatial Object Transfer Standard • SOTF is a transfer format for geospatial data optimised for: • transfer across ‘non-live’ connections • archive • active object store to warehouse connections • OGC’s Geography Markup Language (GML) - both XML encoding of geospatial data
SOTF - Introduction (cont) • Origins in NIMA research contract • Survey of current transfer standards • Requirements specification • neutral, industry standard technology • remedying key shortcomings • incremental update • support for value added • flexible as regards topology • 2D and 3D (and potentially, temporal)
What is SOTF? • A data store imports/exports SOTF datasets, SOTF describes these processes and the demands on the data store • Currently an SOTF dataset is an XML encoding for geospatial data • similar to GML • very similar to [the first version of] SFXML
Use of XML • Current prototype uses XML v1.0 • parallels initial OGC offerings • Area of rapid development • particularly for schema definition, a key part of SOTF • Indexing not key to transfer format, but technology is emerging • Currently verbose, but emerging compression techniques (zip does a lot)
Why the word ‘object’? • SOTF has an object-oriented schema with: • features and feature types • properties and data types • SOTF supports multiple geometric properties per feature • SOTF supports both spatial and aspatial feature types
Why the word ‘object’? • SOTF supports multiple inheritance of feature types • SOTF supports light-weight, binary feature relationships • SOTF is designed to handle complex, structured geospatial data; • it does not support methods and behaviour.
SOTF data store requirements • Designed to work with both [object-]relational and object-oriented data stores • An SOTF dataset always includes an explicit schema • Currently SOTF does this with a fixed DTD • GML profile 1 would not be suitable • To support export of an SOTF dataset a data store • should provide feature identifiers that persist between exports • may provide ability to retain a previous state
SOTF and topology support • SOTF has no in-built topology support • However topological feature types (e.g. faces, edges and nodes) can be described using base SOTF concepts • To work between multiple data stores it is necessary to agree on a common ‘topological sub-schema’ • A sub-schema describes an optional, but standardised, part of the schema
Data producers and data consumers • These two communities have differing requirements: • large vs. small geographic extent • fixed schema vs. ad-hoc inclusion of extra data • time tabled release vs. spontaneous demand • SOTF provides a number of mechanisms to address these differences
Incremental update • Level of granularity is the feature • Tag features as ‘new’, ‘modified’ or ‘deleted’ • Requires persistent feature identifiers • Requires a data store to be able to ‘difference’ states
Data supplier Data consumer SOTF dataset Area-of-Interest
Area-of-Interest • SOTF works at the level of features • granularity of incremental update • granularity of references • SOTF does not require features to be split along artificial tile boundaries • To support ‘area-of-interest’ SOTF requires features to support the concepts of extent and dependency in the data store.
Data supplier (with pre-defined ‘tile’ structure) Pre-generated, stock-piled, set ofSOTF datasets Data consumercombines SOTF datasets Combining SOTF datasets
Value Adding • SOTF provides rules to determine if two schema are compatible. • Since SOTF datasets always include a schema this allows: • schema evolution at the data producer to be communicated to the data consumer • ‘compatible’ additions to the schema to be carried out by the data consumer in support of value-adding • update does not invalidate value-added data
Summary - key techniques • Incremental (or change-only) update • depends on persistent feature (object) ID’s • Support for export ‘area-of-interest’ • avoids ‘tiling’ • Can be combined at receiver • permits ‘stockpiling’ by issuer • Supports value-adding • possible through explicit schema description
SOTF current status • SOTF has been developed to a working prototype by Laser-Scan under contract from NIMA • uses subset of Digital Nautical Chart data • demonstrates the key techniques • NIMA are keen to see further development of something by somebody that addresses these requirements provided: • it is compatible with emerging standards such as GML • there is interest from a wider community
SOTF future status • Concepts under consideration for the evolution of GML within OGC – plays into Web access • Possible definition of content for Transaction Encoding Specification - Interoperability • Possible unifying role in emerging set of XML-based transfer formats - Transfer • Up to the community!