Hello all
I was at the OGC meeting this week, which finished today. Below is a few
points where a see medium-term connections with SIS:
Simple Features (geometry)
-----------------------
A draft for the new ISO 19107 specification is now complete. Among other
talks, I have hear interesting explanations about why apparently
surprising geometries like "chlotoid" are defined. Chlotoids are used
(in my understanding) in rails (trains) as a transition from straight
line to arc of circle in order to avoid discontinuity with the
centrifuge pseudo-force. The recent train accident in Spain is a tragic
illustration of the importance of those factors.
Simple Features for CORBA is not used any more. Simple Features for
OLE/COM is still used by ESRI, but will probably not be maintained as an
OGC standard.
Temporal Discussion Working Group
-------------------------------
The group is exploring how we could define the temporal dimension in a
data request. OGC often uses URL for specifying CRS, and the same
approach is considered for time axis. They will need to support various
temporal datum and units. Example of what it may looks like:
http://.../JulianDate?axis-label="real-date";
http://.../JulianDate?axis-label="simulation-date";
A time series of N images would be describe by only one coverage and one
geometry. The GML elements like <gml:TemporalCRS> and <gml:CompoundCRS>
may have enough capability for this need. A query may looks like:
http://.../subset=E(someWestLongitude, someEastLongitude)&
subset=N(someSouthLatitude, someNorthLatitude)&
unix=(startTime, endTime)
where startTime and endTime would be in seconds since January 1st 1970
in the above example, or in day since the julian calendar epoch if
"unix" were replaced by "JulianCalendar", etc. In the reply, the element
in the <offsetVector> at the time dimension would be in the same unit.
OpenMI
-------
If I understood correctly, OpenMI is a framework for allowing
inter-operability between numerical models. It seems to have similar
goals than GeoAPI, but applied to numerical models instead than
geographic systems. OpenMI is developed in .NET for now, but may provide
Java interfaces later. OpenMI has a property for defining a CRS as a
string, so a collaboration with GeoAPI may be possible here if they
replace their String by an
org.opengis.referencing.crs.CoordinateReferenceSystem object. This will
not happen in OpenMI 2, but may be considered for OpenMI 3.
OpenMI is available on SourceForge and is submitted at OGC for
standardisation as document number 11-014r2. The vote ends on October 1st.
Big Data
-------
They were a talk about Big Data, but it seems to me that the needs could
be addressed by existing standards. We may need however to explain what
would be "good practice" when using OGC standards for big data.
Metadata
--------
Vote for the new version of ISO 19115 is under way at ISO and will ends
on January 21, 2014. If everything goes well, after that date the new
ISO 19115 standard will be final. GeoAPI and Apache SIS would need to be
updated accordingly.
Web Coverage Tiles Services
------------------------
Web Coverage Services (WCS) is facing a problem similar to what Web Map
Services (WMS) faced a few years ago: it is slow in some circumstances.
WMS solved the problem by defining a tiling services, so the WCS group
is considering to apply the same approach. This need has been brought by
the Meteo-Oceanography working group. They were discussion about whether
it should be a new OGC working group, or do the work in the existing WCS
working group. They choose to use the existing group.
Energy
------
I never saw this group before... They had a talk about Pipeline Open
Data Standard (doing Pipeline Integrity Management, identity threads,
high consequence areas, etc.)
Moving features (geometries in motion)
---------------------------------
A draft standard is under way. A use case were presented with a study of
pedestrians, car and trains motions in a Japanese city which allowed the
researchers to compare maps of motion modes during the week and during
the weekend. In this use case, such maps could be used for tuning the
public transportation system.
CRS WKT
--------
Well Known Text (WKT) version 2 is close to final. The document is in
the comments period. ESRI tested it extensively. Implementation in
Apache SIS is under way (after I finished the work on matrix).
CITE tests
--------
The CITE working group is thinking hard about how to make CITE tests
more appealing to developers. They were discussions about TestNG/Maven
integration, the delay between questions and answers on the CITE mailing
list, making easier to use (less configuration), etc. GeoAPI may be able
to help on some of those aspects. We plan to start be setting up a
common place where test files could be put and get the ball rolling for
possible closer integration.
Documentation
-------------
OGC is exploring ways to publish their specification as HTML pages. That
would make like easier for Apache SIS, since the javadoc could link
directly to some clauses. The URL would be
http://docs.opengeospatial.org (no content yet).
Martin
