INSPIRE Thematic Clusters

Layer naming

2377 Views

We have been looking at the issue of Layer Names (described in section 11.1 of the data specification) and have identified two issues:

1. For GE.GeologicUnit there two recommended styles, for lithology and age. However, in a service, layer names have to be unique so should we use GE.GeologicUnit.Lithology and GE.GeologicUnit.AgeOfRocks as the layer names so that each layer can have a single default style? We are not sure if changing the layer name in INSPIRE is permissible or if the layer names are mandatory. If they are mandatory then either we would need to have distinct services for lithology and age or we would have to provide two styles for the GE.GeologicUnit layer and leave it to the client to state which style it wishes to use.

2. A similar issue exists for different types of GE.GeologicUnit, for example derived from maps of different scale or type (eg bedrock or superficial deposit etc). Should we have a layer naming convention that allows us to put out all these different types of GE.GeologicUnit as part of a single service (eg GE.GeologicUnit.[Geographical extent] of the data in the layer, then [Data owner organization] (not service provider), then [Language code] (if non-default as per service naming conventions), then [Scale], then [Theme]) or do we have a separate service for each 'type' of GeologicUnit.

John

Replies

    • Public

    By James PASSMORE

    So for example BGS currently has a single WMS (view service) with the intention of serving harmonized geology data to clients such as the OneGeology Europe portal:

    http://ogc2.bgs.ac.uk/cgi-bin/BGS_OGE_Bedrock_and_Surface_Geology/ows?service=WMS&request=GetCapabilities&

    This single service has six layers with names as shown in the table below.

    Table of layer names in a current OneGeology-Europe View service
    Current Layer Name INSPIRE Layer Name * Suggested rename
    OGE_1M_surface_GeologicStructure
    GE.GeologicFault OR GE.GeologicFold GE.GeologicFault.BGS.EN.1M.Surface
    OGE_1M_bedrock_GeologicStructure
    GE.GeologicFault OR GE.GeologicFold GE.GeologicFault.BGS.EN.1M.Bedrock
    OGE_1M_surface_GeologicUnit
    GE.GeologicUnit GE.GeologicUnit.BGS.EN.1M.Surface.Lithology
    OGE_1M_surface_GeologicUnit_age
    GE.GeologicUnit GE.GeologicUnit.BGS.EN.1M.Surface.Age
    OGE_1M_bedrock_GeologicUnit
    GE.GeologicUnit GE.GeologicUnit.BGS.EN.1M.Bedrock.Lithology
    OGE_1M_bedrock_GeologicUnit_age
    GE.GeologicUnit

    GE.GeologicUnit.BGS.EN.1M.Bedrock.Age

    * Section 11.1 Layers to be provided by INSPIRE view services provides a table that gives layer names like GE.GeologicUnit, GE.GeologicFault, GE.GeologicFold.

    It is a requirement of any WMS service that all layer names are unique, so are we saying that we must split up this service into four different services to be compliant with the specification, or was the intention that an INSPIRE compliant geology view service would serve up layers of type GE.GeologicUnit and allow different layer names.

    I see that the data specification tells us (11 Portrayal):
    A view service may offer several layers of the same type, one for each dataset that it offers data on a specific topic., so there must be some leeway on the layer names...

    ref: http://inspire.ec.europa.eu/documents/Data_Specifications/INSPIRE_DataSpecification_GE_v3.0.pdf

    • Public

    By Clemens PORTELE

    It was left to each TWG to decide about the structure and content of the layers for an application schemaThe Generic Conceptual Model only requires that every spatial object which is geo-referenced is part of at least one layer (req 64). Obviously the GE TWG decided to have a layer "GE.GeologicUnit" and use different styles for representations like lithology/age/etc, rather than layers. As this layer is now also in the Regulation it would require changing the legal text to make a change as far as I can see.

    Regarding layer names and titles, my reading of the Regulation is that both layer name and layer title (in the supported languages) as well as the content (column "spatial object type") is fixed and cannot be changed.

    However, the View Service could still provide additional layers like "GE.GeologicUnit.Lithology" or "GE.GeologicUnit.BGS.EN.1M.Surface.Lithology" as long as there is the layer "GE.GeologicUnit" that meets the requirements of the legal text (and technical guidance).

    If the data spec states that "a view service may offer several layers of the same type", then this seems to be a mistake, or be based on some definition of "layer type" that is not clear to me (I am not sure the term is defined somewhere), or it does not apply to WMS or WMTS.

    • Public

    By James PASSMORE

    I've been told that an earlier draft of the specification had 'layer type' and not 'layer name' in table 11.1, so perhaps the change to name here could itself  be seen as the drafting error rather than the inclusion of the "a view service may offer several layers of the same type" statement.

     

     

     

     

    • Public

    By Clemens PORTELE

    Maybe. But now layer name and title are both in the table in the data spec - and more importantly also in the Regulation.

    • Public

    By James PASSMORE

    It is unfortunate it's also in the regulation, as the data spec has no legal binding and we could have perhaps thrashed out what was intended/useful as opposed regulated.

     

     

     

     

     

     

    • Public

    By Tomas LINDBERG

    What is the practical implications of this discussion? I am not very into technical details of the internal structure of WMS/WFS map servers, and are a little confused sometimes what a "WMS-service" is. I'll try to describe how I have been thinking of this, please set me straight if I am misunderstanding, and give comments on how you are planning to structure your services.

    We are using Geoserver for our WMS's, and so far we have been giving all our layers unique names on that se in a similar way that has been proposed (scale, type, sampling frame etc). We have "packeged" our layers into "view services" using static capability documents, separate services for separate "geologic maps" or collections (one view service for 1M Bedrock, another for 250K Bedrock etc). We think this a a logical way of "linking" datasets and services and metadata for us and users, but maybe we are overly GIS/Map orientated.

    So we are both using descriptive unique names on the layers and separate view services.

    What do you mean with a "view service" on a Geoserver? Doesn't layernames need to be unique on a server? I know there are workspaces in Geoserver that I guess could be used to keep layers with the same name in different workspaces. When you say you need to create separate view services if layernames are to be the same, would that mean create a workspace for every service in Geoserver?

    As for the issue with different styles I would not as a first choice make separate layers for lithology and age just to be able to display both as a default style. There could be many styles to a layer, even if these to has a special "status".

     

     

    • Public

    By James PASSMORE

    Going on Clemens remarks and thinking a little laterally, as I understand it, the following service complies with the letter of the regulation, if not the spirit.  It includes the required layer names and titles, but doesn't change the original OneGeology-Europe layer names or titles.

    http://ogc2.bgs.ac.uk/cgi-bin/BGS_OGE_Bedrock_and_Surface_Geology_in1/ows?service=WMS&request=GetCapabilities&

     

    Use of Group layering to add required INSPIRE layer names to a View service

    But perhaps we could go a little further and develop a naming schema for the grouped layers like that suggested above as this does actually seem useful.

    For example like:

    http://ogc2.bgs.ac.uk/cgi-bin/BGS_OGE_Bedrock_and_Surface_Geology_in2/ows?service=WMS&request=GetCapabilities&

    Here I have renamed the original layer names, but kept the layer titles.  I have also applied the recommended names for the styles to those layers.  The requirement that each layer has a default style is met even for the GE.GeologicUnit group layer as below:

    Layer GE.GeologicUnit style "default" requested explicitly:

    http://ogc2.bgs.ac.uk/cgi-bin/BGS_OGE_Bedrock_and_Surface_Geology_in2/ows?SERVICE=WMS&VERSION=1.3.0&REQUEST=GetMap&BBOX=236796.23161151638487354,680760.14754756249021739,386155.25756803149124607,815391.38101961952634156&CRS=EPSG:27700&WIDTH=933&HEIGHT=841&LAYERS=GE.GeologicUnit&STYLES=default&FORMAT=image/png&DPI=96&MAP_RESOLUTION=96&FORMAT_OPTIONS=dpi:96&TRANSPARENT=TRUE&

    Layer GE.GeologicUnit default style used implicitly:

    http://ogc2.bgs.ac.uk/cgi-bin/BGS_OGE_Bedrock_and_Surface_Geology_in2/ows?SERVICE=WMS&VERSION=1.3.0&REQUEST=GetMap&BBOX=236796.23161151638487354,680760.14754756249021739,386155.25756803149124607,815391.38101961952634156&CRS=EPSG:27700&WIDTH=933&HEIGHT=841&LAYERS=GE.GeologicUnit&STYLES=&FORMAT=image/png&DPI=96&MAP_RESOLUTION=96&FORMAT_OPTIONS=dpi:96&TRANSPARENT=TRUE&

    Similarly for sub layers e.g

    Layer GE.GeologicFault.BGS.EN.1M.Surface default style used implicitly:

    http://ogc2.bgs.ac.uk/cgi-bin/BGS_OGE_Bedrock_and_Surface_Geology_in2/ows?SERVICE=WMS&VERSION=1.3.0&REQUEST=GetMap&BBOX=236796.23161151638487354,680760.14754756249021739,386155.25756803149124607,815391.38101961952634156&CRS=EPSG:27700&WIDTH=933&HEIGHT=841&LAYERS=GE.GeologicFault.BGS.EN.1M.Surface&STYLES=&FORMAT=image/png&DPI=96&MAP_RESOLUTION=96&FORMAT_OPTIONS=dpi:96&TRANSPARENT=TRUE&

    Layer GE.GeologicFault.BGS.EN.1M.Surface style "GE.GeologicFault" requested explicitly:

    http://ogc2.bgs.ac.uk/cgi-bin/BGS_OGE_Bedrock_and_Surface_Geology_in2/ows?SERVICE=WMS&VERSION=1.3.0&REQUEST=GetMap&BBOX=236796.23161151638487354,680760.14754756249021739,386155.25756803149124607,815391.38101961952634156&CRS=EPSG:27700&WIDTH=933&HEIGHT=841&LAYERS=GE.GeologicFault.BGS.EN.1M.Surface&STYLES=GE.GeologicFault&FORMAT=image/png&DPI=96&MAP_RESOLUTION=96&FORMAT_OPTIONS=dpi:96&TRANSPARENT=TRUE&

     

     

    • Public

    By James PASSMORE

    Tomas,

    So for some definitions, as I understand it, a Web Map Service (WMS) is a type of View service, as it provides a a view onto some data (an image).  It is different from a Download service such as a Web Feature Service (WFS), or Web Coverage Service (WCS) that provide access to the data itself.

    As far as portrayal is concerned, certainly it is possible to have two or more different styles associated with a single dataset say lithology, chronostratigraphy, and  lithostratigraphy, so you get differently coloured images (on WMS GetMap requests) and associated legends (on WMS+SLD GetLegendGraphic requests) depending on which style you choose, however I agree that you might want to to go further with your separation and create actual layers.  This latter approach for example allows you to create custom WMS GetFeatureInfo responses.

    You are correct that in a WMS the individual layer names MUST be unique, so  you cannot have more than one layer called GE.GeologicUnit.  Unfortunately the data specification somehow seems to have forgotten this aspect.  Ideally the data specification should I think be amended to take this into consideration, however it is currently as it is, so the only way to have a single service with multiple layers representing different geology at different scales etc is to use group layering...

     

     

     

    • Public

    By John LAXTON

    Just to point out that the sentence 'A view service may offer several layers of the same type, one for each dataset that it offers data on a specific topic'  is in the default text of section 11 of the data specification so will occur in all data specifications (not just Geology). As James says maybe the first column of the table in section 11.1 was meant to be 'Layer Type', but again 'Layer Name' was in the default specification template so will occur in all specifications.

    The issue of different portrayals for age and lithology could have been addressed if we had made these distinct layers. However all data providers will have a range of different data for each layer, such as of varying scale. This will be the case for all layers in all data specifications I would imagine and the sentence stating a service may have several layers of the same type had this in mind I assume. As this is likely to be an issue covering all INSPIRE services maybe it is something the maintenance group should be looking at.

    • Public

    By James PASSMORE

     I've made some slight changes to my second example service which is based on the WMS configuration of our current OneGeology-Europe harmonized geology service.

     http://ogc2.bgs.ac.uk/cgi-bin/BGS_OGE_Bedrock_and_Surface_Geology_in2/ows?service=WMS&request=GetCapabilities&

    Clemens, or anyone, can you let me know if you think this service complies with the layer naming REQUIREMENTS of the Implementing Rules.

    Thanks

    James

     

This discussion is closed.

This discussion is closed and is not accepting new comments.