You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@sis.apache.org by Martin Desruisseaux <ma...@geomatys.fr> on 2013/02/22 13:38:56 UTC
RangeSet, and work for volunteer (if any)
Hello all
The recently committed RangeSet class was the last part of Range-related
classes. The main use case is to represent the range of available data
in a time series, and the holes where data are missing. Imagine for
example a time series for thousands of remote sensing images, each image
being representative on one week (e.g. NOAA's AVHRR data). A developer
can use a loop for inserting the temporal Range of each images in a
RangeSet. Overlapping ranges are merged, which should result in a single
large range covering for instance 10 years of data. But if there is some
holes (some weeks without data), then the RangeSet will contains two or
more Range instances, with the time of missing data between the ranges.
This make easy to spot holes by examining the RangeSet content, or to
display it in a widget.
However for licensing reasons, the implementation of
RangeSet.remove(E,E) method has not been ported. If there is any
volunteer for implementing that method, I have filled a JIRA task for
this purpose:
https://issues.apache.org/jira/browse/SIS-79
Martin
Re: RangeSet, and work for volunteer (if any)
Posted by Martin Desruisseaux <ma...@geomatys.fr>.
Le 22/02/13 22:30, Joe White a écrit :
> Chris, you read my mind. I'm supposed to be back in Atlanta on Sunday
> night, so I can look into it on Monday.
Thank you very much Joe!
I also wonder... would you be eventually interested to contribute to
GeoAPI [1] too? This is not exactly SIS, but closely related since SIS
aims to be a GeoAPI implementation. Scope would be eventually larger,
since - if GeoAPI has some success - it would hopefully be implemented
by more software than just SIS. It is a "kind of work" slightly
different than SIS, in that SIS is more about extensive coding, while
GeoAPI is more about looking carefully to international standards.
GeoAPI is an Open Geospatial Consortium (OGC) working group, but
volunteers can contribute without being OGC member since the SVN
repository is hosted by SourceForge. However most GeoAPI work has to be
given OGC copyright, but the license is BSD-like.
One GeoAPI work that I'm thinking about is the "Geospatial integrity of
geoscience software (GIGS)" [2] test suite. GIGS tests were created by
the authors of the EPSG database (a very important piece of information
in a SIS). Those tests are provided as Excel documents. In GeoAPI, we
started to implement them as JUnit tests [3]. We ported the series 2000
and series 3000 tests, but more series need to be ported. Libraries like
SIS inherit automatically those test suites, so GeoAPI tests could be
used for making comparative analysis of correctness of different
geographic software (I think it could be a nice FOSS4G presentation - we
will be able to compare SIS with Proj.4).
However one inconvenient is that Apache SIS is not yet able to run those
GIGS tests. If you were working on GIGS tests and wanted to try them,
you would have to test with the Geotk library for now (a LGPL library
which is the source of the code ported and relicensed to Apache)... It
would be a temporary situation, since my goal is to finish porting the
Geotk referencing module (and hopefully more than just that module)
before the Nottingham FOSS4G meeting. However I think it is perfectly
normal if you prefer to focus on SIS, so GIGS tests are really just an
idea that I tell just in case.
Regards,
Martin
[1] http://www.geoapi.org/
[2] http://www.epsg.org/gigs.html
[3]
http://www.geoapi.org/geoapi-conformance/apidocs/org/opengis/test/referencing/gigs/package-summary.html
Re: RangeSet, and work for volunteer (if any)
Posted by Joe White <wh...@gmail.com>.
Chris, you read my mind. I'm supposed to be back in Atlanta on Sunday
night, so I can look into it on Monday.
Joe
On Feb 22, 2013 4:18 PM, "Mattmann, Chris A (388J)" <
chris.a.mattmann@jpl.nasa.gov> wrote:
> Thanks Martin.
>
> I'll be at ApacheCon NA 2013 next week but perhaps during that time can
> take a look.
>
> Otherwise, Joe White was around before looking for a way to contribute,
> Joe this may be a cool way :)
>
> Cheers,
> Chris
>
> On 2/22/13 4:38 AM, "Martin Desruisseaux"
> <ma...@geomatys.fr> wrote:
>
> >Hello all
> >
> >The recently committed RangeSet class was the last part of Range-related
> >classes. The main use case is to represent the range of available data
> >in a time series, and the holes where data are missing. Imagine for
> >example a time series for thousands of remote sensing images, each image
> >being representative on one week (e.g. NOAA's AVHRR data). A developer
> >can use a loop for inserting the temporal Range of each images in a
> >RangeSet. Overlapping ranges are merged, which should result in a single
> >large range covering for instance 10 years of data. But if there is some
> >holes (some weeks without data), then the RangeSet will contains two or
> >more Range instances, with the time of missing data between the ranges.
> >This make easy to spot holes by examining the RangeSet content, or to
> >display it in a widget.
> >
> >However for licensing reasons, the implementation of
> >RangeSet.remove(E,E) method has not been ported. If there is any
> >volunteer for implementing that method, I have filled a JIRA task for
> >this purpose:
> >
> >https://issues.apache.org/jira/browse/SIS-79
> >
> > Martin
> >
>
>
Re: RangeSet, and work for volunteer (if any)
Posted by "Mattmann, Chris A (388J)" <ch...@jpl.nasa.gov>.
Thanks Martin.
I'll be at ApacheCon NA 2013 next week but perhaps during that time can
take a look.
Otherwise, Joe White was around before looking for a way to contribute,
Joe this may be a cool way :)
Cheers,
Chris
On 2/22/13 4:38 AM, "Martin Desruisseaux"
<ma...@geomatys.fr> wrote:
>Hello all
>
>The recently committed RangeSet class was the last part of Range-related
>classes. The main use case is to represent the range of available data
>in a time series, and the holes where data are missing. Imagine for
>example a time series for thousands of remote sensing images, each image
>being representative on one week (e.g. NOAA's AVHRR data). A developer
>can use a loop for inserting the temporal Range of each images in a
>RangeSet. Overlapping ranges are merged, which should result in a single
>large range covering for instance 10 years of data. But if there is some
>holes (some weeks without data), then the RangeSet will contains two or
>more Range instances, with the time of missing data between the ranges.
>This make easy to spot holes by examining the RangeSet content, or to
>display it in a widget.
>
>However for licensing reasons, the implementation of
>RangeSet.remove(E,E) method has not been ported. If there is any
>volunteer for implementing that method, I have filled a JIRA task for
>this purpose:
>
>https://issues.apache.org/jira/browse/SIS-79
>
> Martin
>