Hello all
The Open Geospatial Consortium (OGC) had a meeting last week in Delft,
Netherlands. Over 300 people attended in person, and 100+ online. A
subset (315 slides) of the presentations is available at [1], and a
shorter subset (38 slides) at [2]. The next sections in this email are a
mix of personal notes and copies of a template written by OGC staff in [3].
[1]https://portal.ogc.org/files/?artifact_id=107819
[2]https://portal.ogc.org/files/?artifact_id=107799
[3]https://portal.ogc.org/files/?artifact_id=107820
OGC Web Services to OGC API transition
The full set of capabilities offered by the OGC Web Services Standards
(e.g., Web Map Service (WMS), Web Feature Service (WFS), etc.) have now
been reflected in published OGC API Standards or in work in or
approaching final approval vote. For example, OGC API – Maps – Part 1
can be a replacement for WMS (see [1] for more details). Feature API –
part 3 – filtering and CQL2 are going to vote. Over the coming months,
OGC will establish a process and resources to aid in transition to the
more modern Standards, while still ensuring that the user community
recognizes that the legacy web services are still functional and valuable.
Changes proposal for Referencing by Coordinates
One of the mean services of Apache Spatial Information System (SIS) is
its implementation of ISO 19111 / OGC Topic 2 (Referencing by
Coordinates). Apache SIS has been used during OGC TestBed last summer
for prototyping the use of ISO 19111 / OGC Topic 2 standards in space
and for planets other than Earth. The experience gained has been used
for proposing changes to the existing standards: addition of a new
CelestialBody class for identifying the planet, addition of a new
MinkowskiCS class for use with Einstein's special relativity,
generalization of some existing properties, etc. Since the change
proposals concern a joint standard between OGC and ISO, and since ISO is
already revising ISO 19111 with their own changes right now, a
discussion was about how to move forward in sync with ISO. Current
consensus is to start this work at OGC first (as we did for previous
revisions).
Another discussion was about how to move forward with a standard JSON
encoding of Coordinate Reference Systems (CRS). The PROJ project
proposed their own encoding, named PROJJSON. This encoding is a
translation of Well-Known Text (WKT) encoding, with a one-by-one
matching. But WKT is not a model, it is an encoding derived from the
model described in ISO 19111, and is not a one-by-one mapping of that
model. Currently, the only encoding covering fully the ISO 19111 model
is the Geographic Markup Language (GML), but that encoding is based on
an older version of ISO 19111. The working group plans to use PROJJSON
as a starting point, but if we want JSON to replace GML, we may need to
modify or extend PROJJSON. The group will try to avoid gratuitous
incompatible changes, but I think that some changes should be expected.
Other sync between OGC and ISO standards
Temporal reference systems were used to be defined in a separated
standard, ISO 19108. However this standard, published in 2002, does not
fit very well in current OGC/ISO standards. Part of its material has
been absorbed in latest revision of above-cited ISO 19111. The remaining
parts were discussed in a separated OGC working group and resulted in a
new abstract model, which is going to electronic vote soon. My
understanding is that ISO 19108 will no longer be used and will be
replaced by the new OGC abstract topic, with no direct equivalence on
ISO side.
ISO 19123-1 and ISO 19123-3 are two abstract specifications, for
coverage and for processing respectively. They were adopted in April
2023 and became OGC Topic 6. The ISO/OGC relationship is similar to what
has been done with ISO 19111 / OGC Topic 2: the OGC abstract topic has
the same content as the ISO standard, with only editorial differences
(mostly formatting). ISO 19123-3 defines coverage constructors (how to
build a coverage from scratch or derive from an existing coverage),
coverage condensers (summary information from a coverage) and an
expression language. ISO 19123-4 is currently in development and will be
about tiling. See [1] for more overview.
OGC topic 1 – spatial schema (ISO 19107) has a convoluted history. It
was an OGC abstract topic, then dropped, but will now come bask as topic
0. Number 0 is used because number 1 is already used for something
different now.
Moving features / GeoAPI / OpenEO
The working group has defined an abstract model, which is now moving to
the vote state at OGC. This model is strongly inspired by the Moving
Feature JSON encoding, but also introduces some new classes. See [1] for
some UML diagrams. I hope to translate those new classes to Java
interfaces in the OGC GeoAPI project when time will allow.
We had no GeoAPI session during this meeting, because of lack of
preparation time. But work is in progress now for upgrading GeoAPI
`org.opengis.referencing` packages from the ISO 19111:2007 model to the
ISO 19111:2019 model. Apache SIS is upgraded in parallel for testing the
GeoAPI changes. The hope is to have all changes completed for discussion
in the next OGC meeting in June (Montréal).
OpenEO is a project with goals similar to GeoAPI, but targeting
different languages and API levels. OpenEO provides common API for
Python, R, JavaScript and Julia. OpenEO is submitted to OGC as a
community standard. The submitted parts include API and processes,
described on https://api.openeo.org/ and https://processes.openeo.org/
respectively. The submission is going to the public comment phase. See
[1] for more overview.
Climate Resilience Domain Working Group
Geospatial Reporting Indicators have been discussed in the context of
Land Degradation. Means to exchange indicator information reporting the
degree of land degradation (or influencing factors) is not standardized
in the community. OGC members are proposing a new Standard Working Group
to develop such standardized reporting indicators, possibly as extended
functionality of work in Analysis Ready Data.
Futures directions
OGC Standards try to met the FAIR principle: Findable, Accessible,
Interoperable, Reusable. OGC has been traditionally focused on a path
from Findable to Reusable. It resulted in a lot of metadata, because
different users may be interested in different aspects of a feature, and
those aspects needed to be described in metadata in order to be
findable. But the raise of A.I. changes the focus. All reusable data can
be consumed by A.I., which them make those data findable. So the path
can also be in the opposite direction, from Reusable to Findable.
Martin
