Thank you, Martin!
> On Oct 1, 2018, at 4:59 PM, Martin Desruisseaux
> <[email protected]> wrote:
>
> Three weeks ago was hold an OGC meeting in Stuttgart (Germany).
> Consolidated slides giving only the main points of each presentation is
> available at [1]. Below are a few elements that I noted. As usual this
> is only my understanding and may contain mistakes/approximations:
>
>
> New working groups being formed:
>
> * Pipeline working group
> * Statistical working group
> * Blockchain and distributed ledger technologies working group
> * Aviation (actually reviving a dormant working group)
> * GeoAI (Artificial Intelligence) working group
>
> The Pipeline Working Group is proposed by the the oil and gas industry.
> In this environment of highly volatile products under high pressure in
> underground infrastructures difficult to access, their requirements on
> international standards prefer rigidity over flexibility. This reminded
> me Tim Boudreau's blog "Maven's inflexibility is its best feature" (from
> the point of view of an IDE).
>
> The Statistical Domain Working Group will work on integration of
> statistics and geospatial, use of metadata to link them,
> disclosure/privacy issues, etc. On artificial intelligence (AI), an ISO
> Technical Committee has been created in 2017: ISO/IEC JTC1 SC42, with
> their first meeting in April 18 in Beijing (China). The OGC GeoAI
> working group plans to coordinate their work with this ISO TC. IEEE is
> launching an initiative for ethical considerations in AI.
>
>
> New programmatic elements:
>
> The Coordinate Reference Systems (CRS) working group made a report on
> their new features: dynamic datum, three-dimensional projected CRS,
> geoid-based vertical CRS, triaxal ellipsoids for planetary applications,
> etc. Those new features are not yet implemented in Apache SIS, while I
> still hope to do that somewhere in 2019. The Moving Features Working
> Group published a corrigenda introducing new elements -
> MovingFeatureType and DynamicAttributeType - which fill a hole in
> current editions of ISO standards on which GeoAPI interfaces are based.
> This element has been used for completing the org.opengis.feature Java
> package in GeoAPI. Not yet implemented in Apache SIS neither, but on the
> radar.
>
> The GeoAPI working group took a look on the new C++ API of the PROJ
> library, created as part of the "gdalbarn" crow-founding. The C++ API is
> inspired by GeoAPI and follows closely ISO 19111:2018 model. It seems
> close to what could be a "GeoAPI in C++" API. Other elements briefly
> shown in the meeting was the tentative Python API (in addition of
> current Java API) and the specification draft [2] (still work in progress).
>
> For drawing symbols on a map, we did not had formal definition of
> programmatic elements that we could use in GeoAPI. The closest thing
> that we had was XML schema of symbology encoding. OGC is now preparing a
> Symbology Conceptual Model. The core part is available in draft stage at
> [3] and seems better suited to GeoAPI.
>
>
> New format, service or registry:
>
> The GeoPackage format may be extended to include "Common Data Base"
> (CDB) vector data - a synthetic environment database for storing
> computer simulation that represents activities at a high level of
> realism. CDB were previously using Shapefiles. The move to GeoPackage
> would allow to reduce the number of files and disk usage.
>
> The Web Feature Service (WFS) 3.0 draft is available at [4]. The first
> release (tentatively next year) will not cover all WFS 2.0
> functionalities. WFS 2 and WFS 3 are expected to exist in parallel for
> some time. Since WFS is also an ISO standard, WFS 3 will use a new ISO
> number for avoiding to discard WFS 2.
>
> There is a need to allow other authorities (e.g Astronomical Union,
> Oceanographers) to register new Coordinate Reference Systems (CRS), e.g.
> for depth axes measured by pressure. We have not yet clarified what
> would be the process for submitting new CRS to OGC. This is a recurring
> demand.
>
> The concept of DataCube became popular recently, especially in the
> context of Earth Observation. A DataCube can be understood as a
> specialization of the Coverage structure defined by ISO 19123 in the
> special case where data are gridded. The number of dimensions can be
> anything, and DataCubes shall support efficient trimming and slicing.
> Actually this is not really new. As far as Apache SIS is concerned for
> example, there is nothing to change; it is already in-line with the
> requirements enumerated at OGC. Those requirements are:
>
> * Shall support gridded data of between two to four spatial-temporal
> dimensions.
> * Uses the Geographic Coverage model as defined in OGC Coverage
> Implementation Schema (CIS).
> * Shall prepare and structure raw observation data to support analysis
> with minimal additional preparation when analysis is requested.
> * Shall provide metadata for the user that defines the provenance of
> all source data, data preparation methods and structure.
> * Shall support analytics on all domain axes alike, irrespective of
> the axis spatial or temporal semantics.
> * Shall allow efficient trimming and slicing along any number of axes
> from a datacube in a single request.
> * Shall allow tuning of performance to anticipated user query patterns.
> * Shall provide server-side analytic processing on the data including
> composite extraction, processing, filtering, and fusion tasks in
> ad-hoc fashion.
> * Shall support access to metadata.
>
>
> Other new stuffs:
>
> In addition of standards, OGC was used to publish "Best practice"
> documents which are not standards but describe some recommended uses of
> standards. This category of document has been refined in two categories:
> "Best practice" (as before) and "Community practice". The main
> difference is that "Community practice" describe the use of de-factor
> standards that originated outside OGC.
>
> The OGC Naming Authority is defining identifiers in the
> http://www.opengis.net namespace. But this namespace is only for
> formerly approved OGC resources. For resources under development, a new
> namespace has been defined: http://dev.opengis.net.
>
> A data quality vocabulary is being developed at W3C [5]. It is not a
> standard yet.
>
>
> Tests
>
> Current beta version of CITE tests cover the following:
>
> * CSW 2.0.2 (current version: 1.18)
> * GeoPackage 1.0 (current version: 1.1)
> * GeoPackage 1.2 (current version: 0.7)
> * SensorThings API 1.0 (current version: 1.2)
> * WFS 2.0 (current version: 1.30)
> * WMS 1.1 (current version: 1.18)
> * SWE Common 2.0 (current version: 0.3)
> * WMTS 1.0 NSG Profile (current version: 0.3)
> * GeoPackage 1.2 NSG Profile (current version: 0.5)
> * WMS 1.3 NSG Profile (current version: 0.2)
>
>
> [1] https://portal.opengeospatial.org/files/?artifact_id=81129
> [2] http://www.geoapi.org/snapshot/standard_document.html
> [3] https://portal.opengeospatial.org/files/?artifact_id=80686&version=1
> [4]
> https://cdn.rawgit.com/opengeospatial/WFS_FES/3.0.0-draft.1/docs/17-069.html
> [5] https://www.w3.org/TR/vocab-dqv/
>
> Martin
>
>
