Template for comments / Country: Germany / Date: 2016-04-28 / Document: TG for WCS-based download services v1.0rc1 / Project: MIWP-7b /MIG-T
Chapter/ Section
(e.g. 3.1) / Paragraph/Figure/Table/
(e.g. Table 1) / Type of comment1 / Comments / Proposed change / Resolution
Title / ed / Either integrate the WCS guidance into the existing TG for download services (WFS + ATOM) or rename the existing ones.
Foreword / 1.1 / 1st paragraph: Directive 2007/2/EC of the European Parliament and of the Council [Directive 2007/2/EC], … /
Table 2: SPATIAL DATA THEMES in Annex II of [INSPIRE Directive] / ed / Be consistent with other similar documents. / Directive 2007/2/EC of the European Parliament and of the Council [INS DIR], … /
SPATIAL DATA THEMES in Annex II of [INS DIR]
(all) / (all Examples) / ed / Drop "&" in the last row. / e.g. Example 5: WCS&request=GetCapabilities
1 / Paragraph 4, last sentence / ge / „INSPIRE conformant WCS“ - such a service does not exists / Please use s.th. like „INSPIRE conformant Download Service based on WCS“
1 / Table 2 and Table 3 / ed / Please mark those themes for which no other modelling option than coverages are possible.
1.1 / 4th paragraph: This technical guidance shows how the operations required by the INS NS ...
last paragraph: The Spatial Object Types defined in INS ISSDS that... / ed / Put all references in square brackets. / This technical guidance shows how the operations required by the[INS NS] ...
The Spatial Object Types defined in[INS ISSDS]that ...
1.2 / 3rd paragraph: … is included in the Generic Conceptual Model in 9.9.4 / ed / Usethe correct name and add the reference. / is included in the INSPIRE Generic Conceptual Model in 9.9.[INS GCM]
1.2 / p.14, 1st paragraph: Coverage functions are used... / ed / Maybe the mentioning of "functions" is confusing for the general audience. / Check wording.
1.2 / p. 14, 2nd paragraph: ... properties of spatial object types where the type of the property value is a realisation of one of the types... / ed / Not sure people will understand this / Check wording.
1.2 / p. 14, 2nd paragraph: ... ISO 19123 / te / The 19123 coverage model is abstract, and, therefore, allows many non-interoperable concretizations. / Better refer to GMLCOV/CIS which is based on 19123 (too) and concise enough to be interoperable.
1.2 / p. 14, 3rd paragraph, 1st bullet point: ... a grid for which there is an affine transformation between the grifcoordinates and the coordinates of a coordinate reference system (see figure 2, left). / te / This is high-level and abstract. / Add a stanza like "constant spacing along the axes".
1.2 / p. 14, 3rd paragraph, 2nd bullet point: a grid associated with a transformation that can be used to convert grid coordinate values to values of coordinates referenced to a coordinate reference system (see Figure 2, right). / ed / Maybe mention the grid is irregularly spaced?
1.2 / p. 14, 4th paragraph: In addition, some themes make reference to the types TimeValuePair and Timeseries defined in Taylor, Peter (ed.),OGC®WaterML 2.0: Part 1 – Time series, v2.0.0,Open Geospatial Consortium, 2012. These provide a representation of the time instant/value pairs, i.e. time series (see Figure 3). / te / This is confusing, as coverages as well may contain time axes and (unfortunately) WaterML is not harmonized in any way with GMLCOV/CIS.
1.2 / p. 14, last paragraph: Where possible, ... / This seems like a conclusion. / Shift it after the GMLCOV/CIS discussion below.
1.2 / p. 14, last paragraph: ... these coverage types... / te / What does “these” refer to? WaterML != OGC coverages.
1.2 / p. 15, last paragraph: ... extension to GML 3.2.1... / ed / Suggest, for clarification, “extension to the conceptual model of GML 3.2.1 which can be mapped to GML or any other suitable format”.
1.2 / p. 16, paragraph below figure 4: ... GML coverage... / te / Can be any suitable format. / Drop “GML”.
2 / INSPIRE, Implementing Directive 2007/2/EC of the European Parliament and of the Council as regards interoperability of spatial datasets and services, INSPIRE Directive / ed / Work uniform with other similar documents (see e.g. Draft “Technical Guidance for INSPIRE Spatial Data Services and services allowing spatial data services to be invoked”). / INS DIRINSPIRE, Implementing Directive 2007/2/EC of the European Parliament and of the Council as regards interoperability of spatial datasets and services
2 / (all) / ed / Add missing references. / Add e.g. INS GCM, OGC 09-146r1, OGC WCS, ISO 19101 etc.
2 / List of references / ed / Please use another header for the list of OGC specs and INSPIRE CRS than „normative“ references. Normative are only those, which implements legal bindings (Directive, Regulations, Implementing Rules, Decisions). The other things like Technical Guidance and Guidelines concerns only technical things.
3.1 / (1) … conceptual schema for data required by one or more applications [ISO 19101] etc. / ed / Write all references in bold. / conceptual schema for data required by one or more applications[ISO 19101]
3.1 / (12) NOTE 2: The INSPIRE generic concept model document …
(23) NOTE: … and in the INSPIRE Generic Conceptual model document is used synonymously with spatial object. [Note modifiedfrom INSPIRE GenericConceptualmodel] / ed / Use the correct name. / The INSPIRE Generic Conceptual Model document …
… and in the INSPIRE Generic Conceptual Model document is used synonymously with spatial object. [Note modified from INSPIRE Generic Conceptual Model]
3.1 / (5) coordinate reference systems
Systems for uniquely referencing spatialinformation in space as a set of coordinates (x, y, z) and/or latitude and longitude and height, based on a geodetic horizontal and vertical datum[INSPIRE Directive] / ed / Systems for uniquely referencing spatialand/or temporalinformation in space as a set of coordinates (x, y, z,t) and/or latitude and longitude and height, based on a geodetic horizontal and vertical datum[INSPIRE Directive]
3.1 / (6)coverage
spatial object that acts as a function to return values from its range for any direct position within its spatial, temporal or spatiotemporal domain, in accordance with ISO 19123:2007 [INS ISDSS]
EXAMPLE Orthoimage, digitalelevation model (as grid or TIN), point grids etc. / ed / (1)coverage
spatial object that acts as a function to return values from its range for any direct position within its spatial, temporal or spatiotemporal domain, in accordance with ISO 19123:2007 [INS ISDSS]
EXAMPLE Orthoimage,Image timeseries,digitalelevation model (as grid or TIN), point grids etc.
3.1 / (12)geographical grid systems
NOTE 2 ... Thus, a 'geographical grid' is equivalent to an ISO 19123 coverage. ... / ed / Gridded coverage?
3.1 / (12)geographical grid systems
NOTE 2 ... The unqualified term 'grid' may refer either to a grid geometry or a geographical grid (coverage) depending on the context. / ed / For the reader’s understanding, what is the difference? / Add clarification for the difference between grid geometry and geographical grid (coverage).
3.1 / (13)GMLcoverage / ed / Should not contain “GML”, this has led to sustained confusion. / Drop "GML".
3.1 / (13)GMLcoverage
... subclass (specialization) of a Coverage ... / te / Not sure I can follow this, and hard to understand anyway. The term suggests a simple interpretation, which is consistent with OGC: a coverage encoded in GML. As such, I am not sure it is needed as a specific term, unless we define GeoTIFFcoverages, NetCDFcoverages, and the like as well. / Drop term (13).
3.1 / (17) range (of a coverage)
... feature attribute... / ed / Suggest to drop these words – it is not clear what they mean, and they are not necessary either. / Drop "feature attribute".
3.1 / (20) spatial data, NOTE / te / Very much agreed with the NOTE. Does it include temporal?
3.1 / (21) spatialdataservices
operations... / te / Service offers a set of operations, but is not an operation. Example: a WFS instance is not an operation. / Check wording.
3.4 / 2nd paragraph: For example,[INS NS]uses the abbreviated title for the document as shown below:
INSPIRE Network Services Regulation, INS NS, COMMISSION REGULATION (EU) No 1088/2010 of 23 November 2010 amending Regulation (EC) No 976/2009 as regards download services and transformation services / ed / Be consistent with other similar documents (e.g. Draft “Technical Guidance for INSPIRE Spatial Data Services and services allowing spatial data services to be invoked”) / For example,[INS NS]uses the abbreviated title for the document as shown below:
COMMISSION REGULATION (EU) No 1088/2010 of 23 November 2010 amending Regulation (EC) No 976/2009 as regards download services and transformation services
3.5 / p. 23, 2nd paragraph, 1st bullet point: Updates to the Coverage Data Model (GML 3.2.1 Application Schema - Coverages version 1.0.1, OGC 09-146r2, becomes Coverage Implementation Schema 1.1). / ed / Replace "becomes" with "is advanced to" or similar expression. / Updates to the Coverage Data Model (GML 3.2.1 Application Schema - Coverages version 1.0.1, OGC 09-146r2,is advanced toCoverage Implementation Schema 1.1)
3.5 / p. 23, 2nd paragraph, last bullet point:SelectCapabilities extension to WCS core, being adopted / ge / Not sure about this, there are numerous reservations on this spec. Maybe we should not anticipate anything in this (normative) document.
4 / p. 24, 1st paragraph / ed / Replace "exist" with "exists" and add a sentence to clarify the need of a WCS guidance. / Other Technical Guidanceexistsfor describing implementations using other service interfaces such as for Atom Syndication Format, WFS, and SOS.However, only WCS allows extraction and processing of coverage subsets.ThisWCSguidance is based on the abstract model established in the INSPIRE Network Services Regulation [INS NS].
4 / Paragraph 1 / ge / The referenced TG for SOS is not published yet. / Mark this TG as unpublished.
4 / p. 25, Table 4, row 1 and 2 / ed / Not sure about meaning of this row 2: another service? Detailingofrow 1 = headline? / Check and clarify.
4 / p. 25, Table 4, last row: "NOT ABLE TO SUPPORT" / te / Based on the explanation below, maybe it is better understandable to leave out this “NOT ABLE” and just state “By definition, if a service ….” / Drop "NOT ABLE TO SUPPORT".
4 / Table 4 / te / Text „Open request (query) parameters” in column 6 / Move to column 5
4 / p. 25, paragraph below table 4: ...Conversely, when a service has access to a live database(for example one in which coverage data is being continuously updated), ... / te / Maybe the differentiating criterion is not “file vs database” – a file can just as well be updated regularly. / Leave out file/database exemplification.
4 / Paragraph below Table 4, third sentence / ed / Typo in „when you provision“. / Replace “provision” with „provide“.
4 / p. 26, figure 5: Each file has it's own spatial data set identifier; Each coverage has it's own spatial data set identifier / ed / Replace "it's" with "its". / Each file hasitsown spatial data set identifier; Each coverage hasitsown spatial data set identifier
4 / p. 26, paragraph below figure 5: ...a coverage can be mapped to the term spatial object. / ed / A very abstract view – maybe innocent readers understand this wrong. / Check and clarify.
4 / Third paragraph, below Figure 5 – second point / ge / The term „conditional“ is somewhat misleading. This operations are mandatory if the provider will give access through a direct access service! / Check and clarify.
4.1.1.1 / Table 5, first row, Response, point 5: „which must include at least one required CRS“ / ge / Is this really mandated by the directive? In case that data is provided as managed by the provider, without the possibility to fulfil the INSPIRE data model, then the data can be distributed unaltered. No transformation of CRS should be required. / Check and clarify.
4.1.1.1 / Table 5, „Get Download Service Metadata Response“:„will be a WCS Capabilities document, which includes“ ...“the spatial datasets metadata“ / ge / Spatial dataset metadata cannot be included in the capabilities document. It is always referenced via xlink:href elements in MetadataURL tags! / Check and clarify.
4.1.1.1 / Table 5, „Get Spatial Data Set“ / ge / Spatial Data Set Identifier is defined to be a request parameter. How can this be done with a GetCoverage request? In this request the coverage identifier is normally used. The Spatial Data Set Identifier came from the dataset metadata and may be given in the WCS capabilities, but is not a request parameter for the GetCoverage request.
4.1.1.2 / Title: „Conditional download options“ / ge / Don't use the term conditional. Maybe „optional“ is better, cause there is no legal demand for providing direct access download services! / Change title.
4.1.1.2 / Table 9 / te / The DescribeCoverage request of a WCS delivers already a description of the „spatial object“, which is in the case of a coverage the coverage itself. / The Recommended WCS-based implementation should be the same as in Table 7.
4.1.1.2 / Table 10 / te / The requirement to implement the WCS processing extension to realize a direct access download service forms a big obstacle for many service providers. All of the requirements for a direct access service from IR NS can be provided by the GetCoverage operation: e.g. domain subsetting, bounding box, slicing etc. What more query functionality would be needed? And coordinate transformation is not a requirement for INSPIRE download services. If there is really more functionality needed than WCS 2.0 core delivers, it would be better to split the requirements in a part, where WCS 2.0 core is sufficient, and another part, where the WCS processing extension is needed. / Check carefully, if the WCS processing extension must be mandatory to implement a direct access service. If not, define the GetSpatioObject Operation by the use of the GetCoverage function.
4.3.1 / p. 32, example 2 / ge / Should this document maybe support best practice and adhere to the conventions of upper case for the parameter names? I know, some implementations are lax, but IMHO being a standard it should stick with the regulations.
4.2 / Table 12, point 2 – Direct Access Download Options / ge / Definition is incomplete. / C, shall be M if the download service provides direct access to spatial data sets
4.4 / Table 13 / ge / The Table may be omitted and as an alternative a list of CRS should be referenced
5.1 / p. 35:OGC 09-146r2,OGC® GML Application Schema for Coverages, version 1.0 / ge / For all standards referenced, suggest to use just major version number. Reasons:
  • By definition, all 1.x are compatible, therefore anything that works with 1. should also work with 1.1. Confining this would introduce artificial constraints
  • Later changes, say from 1.0 to 1.1 will require major work, such as edits + new version
/ Use just major version number for all standards referenced.
5.1 / p. 35:OGC 09-146r2,OGC® GML Application Schema for Coverages, version 1.0
Conformanceclassesused:
  • gml-coverage
/ te / This seems to express that only GML encodings are wanted. / Add further classes, such as other-format (allowing binary formats like GeoTIFF) and multipart (GML header + binary format).
5.1 / p. 35, TG Requirement 1: Implementations shall conform to OGC 09-110r4 Conformance Class ‘core WCS’ / ge / Recommend to not use concrete document numbers, better reference the version desired (reasons, see above). / Reference the version desired.
5.1 / p. 35, paragraph below TG Requirement 1 / ed / Rephrase paragraph. / The WCS core specification specifies the core operations requiredto be implementedby any WCS,remaining agnostic of the request encoding; protocol bindings areWCSextensions.This way, functionalityandencodingareseparated.
5.1 / p. 35, 2nd paragraph below TG Requirement 1:
Two protocol bindings are currently supported by WCS 2.0: / te / Three bindings, actually: GET, POST, SOAP. REST is under preparation.
BTW, you list only one. / Replace "Two" with "Three" and add missing bindings in the list.
5.1 / p. 35, 3rd paragraph below TG Requirement 1: Note that theXML/SOAP Protocol Binding Extensionspecification in addition to conforming to WCS core also conforms to the following classes: / te / Do the others not conform to OWS Common 2.0? Is there any source for this statement? / Check and clarify.
5.1 / p. 36, 1st paragraph:
Conformance classes used:
  • HTTP POST
  • SOAP encoding
/ te / Trying to interpret this I think it wants to say: “INSPIRE WCS Download Services use the following protocol bindings”. If so, shouldn’t this be a formal requirement?(if it is about WCS SOAP, that has only 1 class) / Check and clarify.
5.1 / p. 36, 2nd and 4th paragraph / ed / Drop "this". / Support for thethisprotocol binding is indicated in a WCS GetCapabilities response as:
5.1 / p. 36, 2nd paragraph: OGC 09-147r3, WCS 2.0 Interface Standard – KVP Protocol Binding Extension, version 1.0 / ed / This formatting is very confusing. Only late I understood that the paragraphs below go _under_ this bullet. Indentation might help. / Check and change format.
5.1 / p. 36, paragraph above TG Requirement 2: You must support... / ed / Replace "You" with "An implementation". / An implementationmust support...
5.1 / p. 36, paragraph above TG Requirement 2: ... you should note that it is the XML/SOAP binding that isproposedfor the ISO coverage service implementation. / te / Caveat: only informal discussion currently, no concrete ISO opinion stated. / Check and rephrase, if applicable.
5.1 / p. 36, TG Requirement 2: Implementations shall support at least one of the two supported protocol bindings (KVP or POST/XML) / ed / Replace "supported" with "WCS" / Implementations shall support at least one of the twoWCSprotocol bindings (KVP or POST/XML)
5.1 / TG Recommendation 4 / ge / The service should support both KVP and XML/SOAP. Maybe we should recommend POST.
5.1 / p. 36, paragraph below TG Recommendation 4: GMLCOV is already included as part of WCS core, i.e. if you make a GetCoverage request and don’t specify any output format the range set will be delivered in XML. / te / No, a coverage is delivered in its Native Format, whatever this might be.
CAVEAT: GMLCOV is not GML! / Check and rephrase.
5.1 / Paragraph below TG Recommendation 4 / ge / Maybe there is a contradiction in OGC 09-110r4? At p.24: "The default format for the coverage response is the coverage’s Native Format." At p.26 it is defined as GMLCOV : "The response to a successful GetCoverage request is a coverage as per [OGC 09-146r2]." Remark: OGC 09-146r2 = GML 3.2.1 Application Schema - Coverages ("GMLCOV")
5.1 / p. 37, Table 14 / ge / Based on the points made in the table, why not establish a recommendation for say JSON for 1D; TIFF for 2D; NetCDF for 3D, 4D? Or maybe tie format recommendations to the Annex themes?
As it stands, clients might have a hard time confronted with the myriads of format choices.
5.1.1 / p. 39, TG Requirement 4: When using either the XML/POST or GET/KVP protocol bindings, the name ... / te / Language is not contingent on protocol, so why not remain generic so that future extensions will not need to touch this. / Drop "When using either the XML/POST or GET/KVP protocol bindings,".
5.1.1 / Second paragraph / ed / Strike last sentence „Note that strictly …“
5.1.1.1 / Second paragraph / ge / Maybe extend list for keywords / Titles, Abstracts, Keywords
5.1.1.2 / p. 40, last paragraph: ... it is up to the implementing Member State to decide... / ge / Is this special for this choice? Why not implementing institution, for example? / Check and rephrase, if applicable.
5.1.1.2 / p. 44, TG Requirement 6: A network service [Download Service]... / ed / Drop "network service". / A Download Service ...
5.1.1.2 / Example 11 / ge / The use of the namespace „://“ is not common to the other TG documents / Maybe this should be altered in all other documents or be omitted.
5.1.1.3 / p. 48, paragraph below Example 12: A service supports French and English and the service default language is French / ed / Strange sentence, maybe incomplete? / Check and rephrase.