190 likes | 290 Views
Interoperability A simple case for standards. Kim Finney JCADM – Rome 2007. Data Exchange Scenario. Instruments/Measurements. Observations. computer to computer communication. Local data store.
E N D
Interoperability A simple case for standards Kim Finney JCADM – Rome 2007
Data Exchange Scenario Instruments/Measurements Observations computer to computer communication Local data store
Data Exchange ScenarioWhich attributes can be considered to be capturing the same information ? CTD Measurements Seal Observations Species Abundance Age Weight Sex Location Mammal type Occurrence Size Age Male/Female Region Project name Cruise Instrument type Principal Investigator name Quality control flags Depth/Temperature values Depth salinity values Latitude Longitude Ship Voyage Equipment Owner Depths Salinity/Temp/Oxygen Start position Stop position
Data Exchange ScenarioNote that the “data type” of the values can also be different? Seal Observations Mammal type = Weddell seal Occurrence = 67 Size = 56 Age = 4 Male/Female =5:1 Region = 151 10 E 70 12 S Species = Leptonychotes weddellii Abundance = 66 Age = five yrs Weight = 65 Sex = 20F30M Location = CS
Data Exchange ScenarioHow do we know what units of measure, codes, CRS have been used ? Seal Observations Species = Leptonychotes weddellii Abundance = 66 (INDIVIDUALS) Age = five yrs old (AV. YEARS) Weight = 65 (AV. Kg) Sex = 20F30M (TOTAL OF EACH) Location = CS (PLACE NAME CODE) Mammal type = Weddell seal pups Occurrence = 67 (INDIVIDUALS) Size = 56 (cm AV. LENGTH) Age = 4 (AV. MONTHS) Male/Female = 5:1 (RATIO) Region = 151 10 E 70 12 S (LAT/LONG CRS)
Data Exchange ScenarioEncodings and formatting patterns can also be very different ? Seal Observations Weddell seal 67 64 56 56 4 4 5:1 4:1 151 10 E 70 12 S 151 10 E 70 12 S L. Wedelli, 66, 5years, 65, 20F30M, CS L. Wedelli, 23, 2years, 64, 22F30M, CS L. Wedelli, 44, 5years, 54, 23F30M, CS L. Wedelli, 23, 3years, 56, 24F30M, CS L. Wedelli, 109, 2years, 68.4, 19F30M, CS
Semantics • Collaborators need to be able to agree on: • The definitions of the real-world objects to be exchanged Weddell Seal pups ? Seals ? Marine mammals ? Leptonychotes weddellii ? All of the above ?
Including attributes • How to model the characteristics of these real-world objects
Encoding • How to encode representations of these real-world objects to exchange them in cyberspace so that machines – not just humans can interpret their meaning. <taxon_specimen> <genus>Leptonychotes</genus> <species> weddellii</species> <common_name>weddell seal</common_name> <observation_date>12Jan06</observation_date> <id>123</id> <sex>F</sex> <weight uom="http://UnitsDictionary.xml#Kg>76 </weight> <age uom="http://UnitsDictionary.xml#yrs>8</age> </taxon_specimen>
Reference Systems • Standard ways of representing real world object geometry or at least declare reference systems. e.g. Swath data e.g. Radar data
Declaration of units of measure • Standard ways of representing location
Controlled Vocabularies • Dictionaries of terms • Term, textual definition • Ranges, permissible codes, units of measure (optional) • Ontologies • As above but includes relationships between terms and rules (axioms).
Ontologies • Many definitions • A shared and common understanding of a domain that can be communicated between people and across application systems (Gruber, 1993). class individual Classes or concepts => is instance of Sub-class is part of class Sub-classes => Relationships individual has value of property Properties or slots => class restricts property
Semantic Interoperability Semantics Encodings Symbols Protocols Feature Catalogue use use Symbology Catalogue
Common PatternsFeature Catalogue • Feature Catalogue • Stores definitions of features (objects) & information about their attributes & relationships between features (semantics). CTD Profile {Feature} Position Bottom Depth Depth Parameter Project Platform Feature Attributes Project {Feature} ownedBy hostedBy Project Name Project Description CTD Profile Platform {Feature} generatedBy Platform Type Platform Name Procedure {Feature} Procedure Instrument Type Processing Method
Common PatternsSymbology Catalogue • Symbology Catalogue • Stores portrayal rules and symbols for displaying features (objects). • Associates features (e.g. CTD) with symbol(s) which can have scale and other styling dependencies.
Common PatternsEncodings (syntax + semantics) GML - An Open Geospatial Consortium (OGC) Standard. - Tuned for representing geometry and topologies. - Widely adopted to exchange and manipulate data with a spatial component. - 600 page specification. - Syntax makes no allowance for incorporating other languages, no axioms supported. - All derivatives of this language are called Application Profiles.
Common PatternsEncodings + Protocol • OGC (GML-based) web service protocols • Commonly used to encode and send data between systems WFS feature data WMS maps WCS coverage data <Whales> <Whale> <gml:name>Minke</gml:name> <gml:whaletype> <gml:String srsName="EPSG:4326"> <gml:coordinates> 0,5.0 20.6,10.7 80.5,60.9 </gml:coordinates> </gml:String> </gml:whaletype> <classification>marinemammal</classification> <number>CAAB Code 34567</number> </whale> </whales> Supports a few well known service functions including: getCapabilities() describeFeature() getFeature()
CTD GML WFS Encoding • <IMOS:Dataset> • {include and import statements for re-using other GML-based ontology elements} • <IMOS:ObservationCollection> {encompasses all deployments and observation members} • {properties describing bounding box, time, etc for all deployments} • <IMOS:Observation> {first observation member} • {project, platform, location, time, procedure, composite phenomena properties} • <IMOS:FeatureofInterest> {CTD feature} • {CTD specific properties encoded in a specialised CSML Profile Feature} • </IMOS:FeatureofInterest> • <IMOS:Result> {CTD data values} • <IMOS:domainSet> • {CTD domain properties and values encoded in a specialised CSML Profile Coverage} • </IMOS:domainSet> • <IMOS:rangeSet> • {CTD range properties and values encoded in a specialised CSML Profile Coverage} • </IMOS:rangeSet> • </IMOS:Result> • </IMOS:Observation> • <IMOS:Observation> {second observation member – pattern repeats} • …………… • …………… • </IMOS:Observation> • </IMOS:ObservationCollection> • </IMOS:Dataset>