[Requests] TML (RFPC 33) Response

Simon.Cox@csiro.au Simon.Cox at csiro.au
Thu Jun 8 04:47:19 EDT 2006


Part A is to be completed once per evaluator per comment submission.  Please iterate over Part B as needed.

 PART A
 ------
 1. Evaluator: Simon Cox, Simon.Cox at csiro.au

 2. Submission: OGC Request 33: OpenGIS® Transducer Markup Language (TML): Request for Public Comments  


 PART B
 ------
1. Specification Section Number - 2 (Conformance)
2. Criticality - Editorial
3. Comments - It is stated that this specification "implement concepts defined in the ISO 19100 series of International Standards". From examination of the content it is not clear how this is true. There are almost no references to ISO 19100 standards. 

1. Specification Section Number - 3 - Normative references 
2. Criticality - Editorial/Major/Baseline harmonization
3. Comments - Relevant standards ISO/IEC 11404, ISO 19103, ISO 19104, ISO 19118, ISO 19123, ISO DIS 19136 are missing from the reference list

1. Specification Section Number - 4 - Terms and defs
2. Criticality - Major/Baseline harmonization
3. Comments - Some of the terms in this clause appear in the standard set of definitions maintained by ISO 19104, or are the primary subject of another standard in the ISO 19100 series. The local dictionary should not contain definitions that differ from these

1. Specification Section Number - 4.29, 6.9
2. Criticality - Editorial
3. Comments - Discussion of standardization of phenomena/property dictionary needs references to relevant OGC specs that deal with this topic - O&M &/or SWE architecture

1. Specification Section Number - 5.3, 6
2. Criticality - Major
3. Comments - UML is the preferred notation for data models, as indicated by the notation summary in clause 5.3. However, there is no UML in the relevant clauses of the specification (clause 6). This makes it difficult to fully comprehend the model, or to relate the XML implementation to the model. See also comment on clause 5.5. 

1. Specification Section Number - 5.5
2. Criticality - Editorial/Major/Baseline harmonization
3. Comments - The statement that the XML Schema for TML "shall reflect the accompanied UML and schema perfectly" is confusing: (a) there is no UML in the spec; (b) there is no description of how such a reflection would work across different languages - this should reference to ISO 19118 &/or ISO DIS 19136 (aka GML 3.2) Annex E (Encoding rule)

1. Specification Section Number - 6
2. Criticality - Editorial
3. Comments - The introductory 3 paragraphs should be in a sub-clause - the OGC/ISO ocumentation standards do not allow text and subsections to be immediate siblings. 

1. Specification Section Number - 6.3
2. Criticality - Editorial
3. Comments - It should be made explicit that "live-streaming data" _samples_ the phenomenon at discrete time instants.

1. Specification Section Number - 6.12
2. Criticality - Editorial/Major/Baseline harmonization
3. Comments - A "data stream" is effectively a special case of ISO 19123 CV_DiscreteCoverage, where the "geometry" property of the element/CV_GeometryValuePair is constrained to have type TM_Instant (see ISO 19108). The various encodings discussed in the rest of the document are _implementations_ of a "DiscreteTimeInstantCoverage" datamodel. Linking the discussions to this abstract data-model is an important connection to the standards baseline. 

1. Specification Section Number - 6.16
2. Criticality - Major/Baseline harmonization
3. Comments - Use of the terms "Set" and "Unit" are inconsistent with ISO/IEC 11404 Clauses 6.8 and 8.4 (both entitled "Aggregate datatypes"). dataSet appears to correspond most closely with "Record". dataUnit appears to correspond most closely with "Item" (from the select-item operation). 

1. Specification Section Number - 6.18
2. Criticality - Editorial/Major/Baseline harmonization
3. Comments - Sequencing rules should be compared with CV_SequenceRule described in ISO 19123 clause 8.15 and Annex D. 

1. Specification Section Number - 6.22
2. Criticality - Editorial
3. Comments - Orphaned sub-clause at beginning of clause should have a heading. Numbering of sub-clauses is also a little wierd. 

1. Specification Section Number - 6.22, 6.23
2. Criticality - Major/harmonization
3. Comments - There appears to be a lot of scope overlap with SensorML in these clauses. A common model (in UML?) and terminology should be used. 

1. Specification Section Number - 7
2. Criticality - Major
3. Comments - The XML encoding is not derived from GML. This is noted as a future goal. 

However, this could be made easier now at little cost by following the GML encoding patterns now. In particular through use of 
(i) element-content in preference to XML attributes for data, 
(ii) normalization of common objects with use of xlinks for cross-referencing, supported by URI's rather than text for identifiers, and 
(iii) interleaving elements representing Objects and properties. 
The latter would also be made easier by use of UML following the ISO 19103 profile for system modelling. Note that this maps quite directly into XML Spy diagrams. 
The coding standards in the XML Schemas, particularly concerning global type definitions and element declarations are also inconsistent with conventional practice in OGC data standards. 

1. Specification Section Number - All
2. Criticality - Major/harmonization
3. Comments - Overall, TML is currently a "standalone" specification: it does not explicitly re-use any components provided by baseline OGC or ISO specifications, as supported by the XML Namespace and XML Schema <import> mechanisms. This will mean that processors and interfaces compatible with existing OGC data standards must be completely re-written to deal with TML, with essentially no component re-use.



More information about the Requests mailing list