250 likes | 401 Views
A data flow approach to interoperability. By Arve Meisingset. Purpose. Provide a generic architecture of one system Not (p) and ( ¬ p) within one system Provide a methodological basis for architecture specifications Mappings between data (format), not function blocks.
E N D
A data flow approach to interoperability By Arve Meisingset
Purpose • Provide a generic architecture of one system Not (p) and (¬ p) within one system • Provide a methodological basis for architecture specifications Mappings between data (format), not function blocks
Program:(X-3)*2+(Y+1) X 3 Y x ? - + 1 y u w 2 * y v ? x + Z 1a
Control:(X-3)*2+(Y+1) 3 x ? - + 1 y 2 * y ? x + 1b
Data flow:(X-3)*2+(Y+1) X 3 Y - + 1 u w 2 * v + Z 1c
Precedence graph:(X-3)*2+(Y+1) X 3 Y (X-3)*2+(Y+1) Z 2a
Precedence graph:as relations between data only X 3 Y (X-3)*2 +(Y+1) Generic processes Functions as subordinate (algorithmic) methods (X-3)*2 +(Y+1) Z 2c
Schema input schema Generic processor enforcing data instances according toSchema of data classes output 3a
Instances instances processors input schema internal data output Two-way transformations 3b
Population processors schema population Reading and writing in population 3c
External and Internal Schemata ES schemata IS ES populations EP IP EP External and Internal populations 4
EP Central Application Schema schemata ES IS AS ES EP AP IP populations 5
EP AP IP Three Schema Architecturei.e. a compiler architecture for data transformation System schema ES AS IS Er Ar Ir System processor System population EL AL IL 6
LS CS TS OS TS DS PS IP Data Transformation Architecture System schema External schema Application schema Internal schema Lr Cr Tr Or Tr Dr Pr External processor Application processor Internal processor LP CP TP OP TP DP PP 7
IP DS PS DS TS LS TS CS LS LS PS OS CS Data Flow between Layers Application schema enforcement of terminologies and consistency Presentation schema 8 All data may not be transformed to the concept form
TS OS CS TS CS LS TS DS OS TS Or Tr Tr Lr Cr Or Tr Tr Cr Dr Candidate interfaces transformation Li Ci Ti Oi Ti Di communication Transformation between different data within processes Communication of same data between processes Therefore, implementation processes (yellow) are dual to the schema architecture (pink) 9
Database Nesting Developer & system manager access to end user help Meta schema Schema End user Processor 10
Executable code Database Dictionary Database of schemata Code generation Meta schema ES AS IS Developer & system manager IS code generation End user Processor 11
Executable code Executable code Executable code Database Dictionary database Tool specification Bootstrapping Processor Tool developer Developer & system manager Processor End user Processor 12
lL lC lT lO lT lD lP pL pC pT pO pT pD pP The Schema Cube Developer & system manager view Tool developer’s view plP cL tL oL tL tD dL End user view 13
Example Presentation of dictionary data to developer lL lC lT lO lT lD lP cL The notations are placed in a two dimensinal grid tL oL tL tD dL X.11 pL pC pT pO pT pD pP BER
lL lC lT lO lT lD lP pL pC pT pO pT pD pP Reguirements for the seven schemata and mappings between them and different requirements for different media cL tL oL tL tD dL
Approaches to architecture • Control flow • Data flow • Precedence graphs • Data transformation • Nesting
Usage ? Focus of SG17 ?
Future work Development of RM • 1. other candidate reference models, • 2. criteria and perspectives of reference models, • 3. comparisons with other reference models, such as the ODP, • Use of RM • 4. requirements for each element of the reference model, • 5. each notation has to be evaluated for use in each element of the reference model, • 6. a map has to be developed on the combined use of notations.