Hello all
There is some notes about the OGC sessions that I saw yesterday and today.
GeoAPI
----------------------------------------
We presented the proposed content for GeoAPI 3.1 [1]. The slides
summarized the following points:
* Guidelines applied in the process of deriving the Java interfaces
from UML.
* Three examples about why we should derive Java interfaces from UML,
not from XSD.
* Kind of compatible changes.
* Kind of incompatible changes.
* Aftermath: about 200 compatible changes and 30 incompatible ones.
* Mapping ISO internationalisation objects (PT_FreeText) to Java
(Locale, Charset, InternationalString).
They were no motion. The next work will be to update the GeoAPI
specification accordingly.
OGC has a GitHub repository [2] where some standards, test suites or
educational materials are developed. Maybe we should move the GeoAPI
repository (currently on SourceForge) there.
Metadata (ISO 19115):
----------------------------------------
Work on XML schemas for the new ISO 19115:2014 standard is ongoing in an
ISO-TC211 repository on GitHub [3]. Maybe Apache SIS could start looking
at it for the transition to the new XML schema, now that we have
transitioned the Java API. However I think it would depend on volunteer
resources: if someone is interested in following ISO-TC211/XML work,
update SIS implementation accordingly and provide feedbacks to
ISO-TC211/XML, than would be great! Otherwise we may just wait for the
final XSD before to begin this work.
The Earth Science Information Partners (ESIP) web site is worth to note.
For example in the "Documentation connections" page [4], "Documenting
Resource Identification", there is a mapping between concepts (e.g.
"resource title", "keyword", "bounding box", etc.) and various metadata
dialects (ISO, DIF, ECHO, ECS, FGDC, etc.). In Apache SIS, we currently
use ISO 19115 as the "lingua franca" of spatial metadata. So if someone
implemented a reader for the other formats, we could use those mapping
for storing the information in SIS metadata objects.
ESIP also summarize the concepts that are recommended by the different
dialects. So we can see that the "title" is recommended by 7 different
metadata dialects, the "bounding box" by 5 different dialects, the
language by 2 different dialects, etc. This may be a way to prioritize
the information.
The is a strong interest for a tools that quantify the metadata
completion. There is already an online tool on the NOAA web site (I
think). This is an other thing that SIS could do if there is volunteer.
Geometries (ISO 19107)
----------------------------------------
The group is at only two weeks of submitting the new ISO 19107 draft to
ISO as a community draft. Past that point, it would be much more
difficult to edit. The new standard greatly simplify the previous
version. Segmentations and boundary objects are replaced by ordinary
geometry objects. B-Splines and NURBES are considered mathematically
equivalent, so the new standard uses only polynomial spline interface.
The were also changes to harmonize with SQL/MM Spatial.
Vertical CRS
----------------------------------------
The meteo-oceanography group needs to be able to express the
transformations between various vertical CRS definitions, for example
from a z value measured as a pressure to a z value in metres. The usual
way to do this kind of work for horizontal CRS is to store the
transformation steps for various (sourceCRS, targetCRS) tuples in a
database, in particular the EPSG database [5]. However it is not yet
clear if this approach could be applicable to the kind of vertical
coordinate transformations done in the meteorological community because
of the large amount of definitions (e.g. at least one for each airport)
and there time-varying nature. One alternative approach that this group
wishes to explore is whether we should try to encore more explicit
information in the vertical datum [6] for allowing libraries to infer
the transformation steps without an EPSG-like database.
Workflow:
----------------------------------------
Similar to Web Processing Service (WPS), but discussion about:
* Defining choices between alternative processes depending on which
data are available.
* Feedback or interaction mechanism (e.g. know if a job is stalled, etc.).
* If a step in WPS requires a login/password, how to pass the request
in the chain of process up to the user.
This is still in exploratory phase.
Martin
[1] http://www.geoapi.org/snapshot/javadoc/content.html
[2] https://github.com/opengeospatial
[3] https://github.com/ISO-TC211/XML
[4] http://wiki.esipfed.org/index.php/Category:Documentation_Connections
[5] http://epsg-registry.org/
[6]
http://www.geoapi.org/snapshot/javadoc/org/opengis/referencing/datum/VerticalDatum.html
