Hello,
I have a client with data that spans the dateline (Northern Russian
coast to Northern Alaskan coast). Is there a way to do this in
geonetwork? Just putting in the lat/long gives us all the longitudes
outside the region of interest.
Thanks
Karen
On 5/9/12 3:39 PM, Karen Remick wrote:
Hello,
I have a client with data that spans the dateline (Northern Russian
coast to Northern Alaskan coast). Is there a way to do this in
geonetwork? Just putting in the lat/long gives us all the longitudes
outside the region of interest.
I believe that we had fixed this in the indexing in GeoNetwork some time ago as we dealt with Pacific and global data. The typical way to do this was to break the metadata footprint into two records each of which does not span the dateline, and allow for queries to be decomposed to do the same.
Kai - did you commit this index and search change to GeoNetwork?
Doug.
Thanks
Karen------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at http://sourceforge.net/projects/geonetwork
--
Douglas D. Nebert
Senior Advisor for Geospatial Technology, System-of-Systems Architect
FGDC Secretariat Tel/Fax:+1 503 454-6248 Cell:+1 703 459-5860
Shouldn't it just work by having the most Westerly longitude to be
greater than the most easterly longitude? If the logic is right then the
comparison should be 'Most Westerly Longitude < search longitude < Most
easterly Longitude' etc.
John
On Wed, 2012-05-09 at 17:56 -0400, Douglas Nebert wrote:
On 5/9/12 3:39 PM, Karen Remick wrote:
> Hello,
> I have a client with data that spans the dateline (Northern Russian
> coast to Northern Alaskan coast). Is there a way to do this in
> geonetwork? Just putting in the lat/long gives us all the longitudes
> outside the region of interest.
I believe that we had fixed this in the indexing in GeoNetwork some time
ago as we dealt with Pacific and global data. The typical way to do this
was to break the metadata footprint into two records each of which does
not span the dateline, and allow for queries to be decomposed to do the
same.Kai - did you commit this index and search change to GeoNetwork?
Doug.
> Thanks
> Karen
>
>
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
> _______________________________________________
> GeoNetwork-users mailing list
> GeoNetwork-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/geonetwork-users
> GeoNetwork OpenSource is maintained at http://sourceforge.net/projects/geonetwork
>
Hi John,
That's how I thought it worked too, e.g if we had a dataset
that spanned the dateline we have e.g +176E as the 'westBoundLongitude'
and -176W as the 'eastBoundLongitude', in xml:
<gmd:EX_GeographicBoundingBox>
<gmd:westBoundLongitude>
<gco:Decimal>176</gco:Decimal>
</gmd:westBoundLongitude>
<gmd:eastBoundLongitude>
<gco:Decimal>-176</gco:Decimal>
</gmd:eastBoundLongitude>
<gmd:southBoundLatitude>
<gco:Decimal>-20</gco:Decimal>
</gmd:southBoundLatitude>
<gmd:northBoundLatitude>
<gco:Decimal>-14</gco:Decimal>
</gmd:northBoundLatitude>
</gmd:EX_GeographicBoundingBox>
Karen,
Is this the way you had entered your west/east longitudes i.e
is the west one +ve and the east one -ve?
Andrew
----- Original Message ----- From: "john.hockaday" <john.hockaday@anonymised.com>
To: "Douglas Nebert" <ddnebert@anonymised.com>
Cc: <geonetwork-users@lists.sourceforge.net>
Sent: Friday, May 11, 2012 10:50 AM
Subject: Re: [GeoNetwork-users] spanning the dateline
Shouldn't it just work by having the most Westerly longitude to be
greater than the most easterly longitude? If the logic is right then the
comparison should be 'Most Westerly Longitude < search longitude < Most
easterly Longitude' etc.John
On Wed, 2012-05-09 at 17:56 -0400, Douglas Nebert wrote:
On 5/9/12 3:39 PM, Karen Remick wrote:
> Hello,
> I have a client with data that spans the dateline (Northern Russian
> coast to Northern Alaskan coast). Is there a way to do this in
> geonetwork? Just putting in the lat/long gives us all the longitudes
> outside the region of interest.
I believe that we had fixed this in the indexing in GeoNetwork some time
ago as we dealt with Pacific and global data. The typical way to do this
was to break the metadata footprint into two records each of which does
not span the dateline, and allow for queries to be decomposed to do the
same.Kai - did you commit this index and search change to GeoNetwork?
Doug.
> Thanks
> Karen
>
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
> _______________________________________________
> GeoNetwork-users mailing list
> GeoNetwork-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/geonetwork-users
> GeoNetwork OpenSource is maintained at > http://sourceforge.net/projects/geonetwork
>------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at http://sourceforge.net/projects/geonetwork
Hi Doug,
Looking at how GN 2.6.3 behaves when you enter a bounding box
on the advanced search form. If you enter a a box crossing dateline with west-east 100E/-100W,
north/south -14S/-20S Geonetwork.log file indicates a polygon query like this:
<geometry>POLYGON((
-100.0000 -20.0000,
100.0000 -20.0000,
100.0000 -14.0000,
-100.0000 -14.0000,
-100.0000 -20.0000))
</geometry>
Trouble is this box is completely 'outside' the box you wanted crossing the dateline.
Instead this box runs across the Greenwich meridian (0). You can
also see same box drawn on the query map. Guess that explains
Karen's result of getting results outside the region of interest. I tried
flipping the west-east co-ords and same problem.
Problem is the sofware doesn't know if you wanted your box to cross the
dateline(180) or greenwich (0). As you say you could break the query into 2 parts to get
the results you want but that's tedious. Perhaps as enhancement the user
could specify that the box crossed the Dateline or Greenwich and then
the box gets composed how you want.
Cheers,
Andrew
----- Original Message ----- From: "Douglas Nebert" <ddnebert@anonymised.com>
To: "andrew walsh" <awalsh@anonymised.com>; "John Hockaday" <John.Hockaday@anonymised.com>
Sent: Friday, May 11, 2012 2:56 PM
Subject: Re: [GeoNetwork-users] spanning the dateline
On 5/10/12 10:11 PM, andrew walsh wrote:
Hi John,
That's how I thought it worked too, e.g if we had a dataset
that spanned the dateline we have e.g +176E as the 'westBoundLongitude'
and -176W as the 'eastBoundLongitude', in xml:Yes, that is the way one should encode either the query or the metadata, but the cartesian logic is broken, and can be achieved only through two queries. I thought we fixed this about a year ago.
<gmd:EX_GeographicBoundingBox>
<gmd:westBoundLongitude>
<gco:Decimal>176</gco:Decimal>
</gmd:westBoundLongitude>
<gmd:eastBoundLongitude>
<gco:Decimal>-176</gco:Decimal>
</gmd:eastBoundLongitude>
<gmd:southBoundLatitude>
<gco:Decimal>-20</gco:Decimal>
</gmd:southBoundLatitude>
<gmd:northBoundLatitude>
<gco:Decimal>-14</gco:Decimal>
</gmd:northBoundLatitude>
</gmd:EX_GeographicBoundingBox>Karen,
Is this the way you had entered your west/east longitudes i.e
is the west one +ve and the east one -ve?Andrew
----- Original Message -----
From: "john.hockaday"<john.hockaday@anonymised.com>
To: "Douglas Nebert"<ddnebert@anonymised.com>
Cc:<geonetwork-users@lists.sourceforge.net>
Sent: Friday, May 11, 2012 10:50 AM
Subject: Re: [GeoNetwork-users] spanning the datelineShouldn't it just work by having the most Westerly longitude to be
greater than the most easterly longitude? If the logic is right then the
comparison should be 'Most Westerly Longitude< search longitude< Most
easterly Longitude' etc.John
On Wed, 2012-05-09 at 17:56 -0400, Douglas Nebert wrote:
On 5/9/12 3:39 PM, Karen Remick wrote:
Hello,
I have a client with data that spans the dateline (Northern Russian
coast to Northern Alaskan coast). Is there a way to do this in
geonetwork? Just putting in the lat/long gives us all the longitudes
outside the region of interest.I believe that we had fixed this in the indexing in GeoNetwork some time
ago as we dealt with Pacific and global data. The typical way to do this
was to break the metadata footprint into two records each of which does
not span the dateline, and allow for queries to be decomposed to do the
same.Kai - did you commit this index and search change to GeoNetwork?
Doug.
Thanks
Karen------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at
http://sourceforge.net/projects/geonetwork------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at
http://sourceforge.net/projects/geonetwork------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at http://sourceforge.net/projects/geonetwork--
Douglas D. Nebert
Senior Advisor for Geospatial Technology, System-of-Systems Architect
FGDC Secretariat Tel/Fax:+1 503 454-6248 Cell:+1 703 459-5860
Thank you Andrew, That is exactly what is happening.
Perhaps the next version could have a check box that would put the box over
the dateline instead of Greenwich, or take values <-180 or >180 (ie 170 to
190 or -170 to -190) to mean lapping over the dateline. In the mean time
I'll tell the clients that they should make 2 files, both with complete data
and a note telling people searching the metadata that the data covers the
whole region.
Karen
--
View this message in context: http://osgeo-org.1560.n6.nabble.com/spanning-the-dateline-tp4963898p4968823.html
Sent from the GeoNetwork users mailing list archive at Nabble.com.
The other way is to use 0-360 longitudes instead of +-180, but you may need to tidy up the basemap as well, to support longitude values from -180 to +360.
Note that EPSG claimed that both +-180 & 0-360 longitude spaces are valid under EPSG:4326, it is up to the implementation to sort this out.
In the FOSS world to help address this, GMT is pretty seamless across 180, Postgis has a ST_Shiftlongitude() function and proj (cs2cs) allows the user to specify the central meridian of the reprojected coordinates, defaulting to 0. Open layers has a flag which sort of works... I'm not sure what specific approach has been used in Gepnetwork, but multiple polygons does work.
Brent Wood (from NZ where this is indeed a common issue!)
--- On Fri, 5/11/12, andrew walsh <awalsh@anonymised.com> wrote:
From: andrew walsh <awalsh@anonymised.com>
Subject: Re: [GeoNetwork-users] spanning the dateline
To: "Douglas Nebert" <ddnebert@anonymised.com>
Cc: geonetwork-users@lists.sourceforge.net
Date: Friday, May 11, 2012, 7:17 PM
Hi Doug,
Looking at how GN 2.6.3 behaves when you enter a bounding box
on the advanced search form. If you enter a a box crossing dateline with
west-east 100E/-100W,
north/south -14S/-20S Geonetwork.log file indicates a polygon query like this:
<geometry>POLYGON((
-100.0000 -20.0000,
100.0000 -20.0000,
100.0000 -14.0000,
-100.0000 -14.0000,
-100.0000 -20.0000))
</geometry>
Trouble is this box is completely 'outside' the box you wanted crossing the
dateline.
Instead this box runs across the Greenwich meridian (0). You can
also see same box drawn on the query map. Guess that explains
Karen's result of getting results outside the region of interest. I tried
flipping the west-east co-ords and same problem.
Problem is the sofware doesn't know if you wanted your box to cross the
dateline(180) or greenwich (0). As you say you could break the query into 2
parts to get
the results you want but that's tedious. Perhaps as enhancement the user
could specify that the box crossed the Dateline or Greenwich and then
the box gets composed how you want.
Cheers,
Andrew
----- Original Message -----
From: "Douglas Nebert" <ddnebert@anonymised.com>
To: "andrew walsh" <awalsh@anonymised.com>; "John Hockaday"
<John.Hockaday@anonymised.com>
Sent: Friday, May 11, 2012 2:56 PM
Subject: Re: [GeoNetwork-users] spanning the dateline
On 5/10/12 10:11 PM, andrew walsh wrote:
Hi John,
That's how I thought it worked too, e.g if we had a dataset
that spanned the dateline we have e.g +176E as the 'westBoundLongitude'
and -176W as the 'eastBoundLongitude', in xml:Yes, that is the way one should encode either the query or the metadata, but
the cartesian logic is broken, and can be achieved only through two queries. I
thought we fixed this about a year ago.<gmd:EX_GeographicBoundingBox>
<gmd:westBoundLongitude>
<gco:Decimal>176</gco:Decimal>
</gmd:westBoundLongitude>
<gmd:eastBoundLongitude>
<gco:Decimal>-176</gco:Decimal>
</gmd:eastBoundLongitude>
<gmd:southBoundLatitude>
<gco:Decimal>-20</gco:Decimal>
</gmd:southBoundLatitude>
<gmd:northBoundLatitude>
<gco:Decimal>-14</gco:Decimal>
</gmd:northBoundLatitude>
</gmd:EX_GeographicBoundingBox>Karen,
Is this the way you had entered your west/east longitudes i.e
is the west one +ve and the east one -ve?Andrew
----- Original Message -----
From: "john.hockaday"<john.hockaday@anonymised.com>
To: "Douglas Nebert"<ddnebert@anonymised.com>
Cc:<geonetwork-users@lists.sourceforge.net>
Sent: Friday, May 11, 2012 10:50 AM
Subject: Re: [GeoNetwork-users] spanning the datelineShouldn't it just work by having the most Westerly longitude to be
greater than the most easterly longitude? If the logic is right then the
comparison should be 'Most Westerly Longitude< search longitude< Most
easterly Longitude' etc.John
On Wed, 2012-05-09 at 17:56 -0400, Douglas Nebert wrote:
On 5/9/12 3:39 PM, Karen Remick wrote:
Hello,
I have a client with data that spans the dateline (Northern Russian
coast to Northern Alaskan coast). Is there a way to do this in
geonetwork? Just putting in the lat/long gives us all the longitudes
outside the region of interest.I believe that we had fixed this in the indexing in GeoNetwork some time
ago as we dealt with Pacific and global data. The typical way to do this
was to break the metadata footprint into two records each of which does
not span the dateline, and allow for queries to be decomposed to do the
same.Kai - did you commit this index and search change to GeoNetwork?
Doug.
Thanks
Karen------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at
http://sourceforge.net/projects/geonetwork------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at
http://sourceforge.net/projects/geonetwork------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at
http://sourceforge.net/projects/geonetwork--
Douglas D. Nebert
Senior Advisor for Geospatial Technology, System-of-Systems Architect
FGDC Secretariat Tel/Fax:+1 503 454-6248 Cell:+1 703 459-5860
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at http://sourceforge.net/projects/geonetwork
Hi Karen, Brent,
Thanks.
Thinking about the maths when we enter a bounding box
it GN asks for a west bound co-ordinate (say +175) and an
east bound coordinate (say -175). So west > east when we
cross the dateline, different to case when box crosses Greenwich (0)
e.g -100 -> +100 then west < east. So when
west > east we know dateline has been crossed in which case GN sends
off the appropriate query that gets metadata-dataset extents crossing dateline
properly.
Not very convenient if your clients are needing to split up their
datasets before/after dateline just so they can discovered. It shouldn't
be that the software limitation is dictating the data management.
Andrew
----- Original Message ----- From: "KRemick" <kremick@anonymised.com>
To: <geonetwork-users@lists.sourceforge.net>
Sent: Saturday, May 12, 2012 3:22 AM
Subject: Re: [GeoNetwork-users] spanning the dateline
Thank you Andrew, That is exactly what is happening.
Perhaps the next version could have a check box that would put the box over
the dateline instead of Greenwich, or take values <-180 or >180 (ie 170 to
190 or -170 to -190) to mean lapping over the dateline. In the mean time
I'll tell the clients that they should make 2 files, both with complete data
and a note telling people searching the metadata that the data covers the
whole region.Karen
--
View this message in context: http://osgeo-org.1560.n6.nabble.com/spanning-the-dateline-tp4963898p4968823.html
Sent from the GeoNetwork users mailing list archive at Nabble.com.------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at http://sourceforge.net/projects/geonetwork
Hi Andrew,
I agree, but that is the effect of using a cartesian geometry to represent a sphere. Always some kludge to work around it. This is not just a Geonetwork issue, but a general GIS one.
See this article 
http://it.slashdot.org/story/07/02/25/2038217/software-bug-halts-f-22-flight
Supposedly any polygon has an inside & an outside. Draw a circle representing a polygon on a globe. What happens when the polygon you draw is the equator? Can we rely on the right hand rule?
The other issue (perhaps more on-topic) is that software other than Geonetwork interrogates the coordinates (bounding box) via CSW, or the coordinates are harvested by GN from a WFS or WMS layer. It also needs to work with OpenLayers for mapping as well as searching. So any solution must be:
compliant with the spec for the profile
interoperable with other tools
Which makes it far from straightforward, & not just a GN problem.
Cheers,
Brent Wood
--- On Mon, 5/14/12, andrew walsh <awalsh@anonymised.com> wrote:
From: andrew walsh <awalsh@anonymised.com>
Subject: Re: [GeoNetwork-users] spanning the dateline
To: "KRemick" <kremick@anonymised.com>
Cc: geonetwork-users@lists.sourceforge.net
Date: Monday, May 14, 2012, 12:06 PM
Hi Karen, Brent,
Thanks.
Thinking about the maths when we enter a bounding box
it GN asks for a west bound co-ordinate (say +175) and an
east bound coordinate (say -175). So west > east when we
cross the dateline, different to case when box crosses Greenwich (0)
e.g -100 -> +100 then west < east. So when
west > east we know dateline has been crossed in which case GN sends
off the appropriate query that gets metadata-dataset extents crossing dateline
properly.
Not very convenient if your clients are needing to split up their
datasets before/after dateline just so they can discovered. It shouldn't
be that the software limitation is dictating the data management.
Andrew
----- Original Message -----
From: "KRemick" <kremick@anonymised.com>
To: <geonetwork-users@lists.sourceforge.net>
Sent: Saturday, May 12, 2012 3:22 AM
Subject: Re: [GeoNetwork-users] spanning the dateline
Thank you Andrew, That is exactly what is happening.
Perhaps the next version could have a check box that would put the box over
the dateline instead of Greenwich, or take values <-180 or >180 (ie 170 to
190 or -170 to -190) to mean lapping over the dateline. In the mean time
I'll tell the clients that they should make 2 files, both with complete data
and a note telling people searching the metadata that the data covers the
whole region.Karen
--
View this message in context:
http://osgeo-org.1560.n6.nabble.com/spanning-the-dateline-tp4963898p4968823.html
Sent from the GeoNetwork users mailing list archive at Nabble.com.------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at
http://sourceforge.net/projects/geonetwork
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at http://sourceforge.net/projects/geonetwork
Hi Brent,
Please see comments in line below:
On Fri, 2012-05-11 at 15:34 -0700, pcreso@anonymised.com wrote:
The other way is to use 0-360 longitudes instead of +-180, but you may need to tidy up the basemap as well, to support longitude values from -180 to +360.
The trouble is that according to ISO 19115 the bounding box range is
-180 to +180. If we use 0-360 then that shouldn't be valid entries in
the ISO 19115 metadata.
Note that EPSG claimed that both +-180 & 0-360 longitude spaces are valid under EPSG:4326, it is up to the implementation to sort this out.
This may be so but it is not a valid range for ISO 19115 bounding box.
In the FOSS world to help address this, GMT is pretty seamless across 180, Postgis has a ST_Shiftlongitude() function and proj (cs2cs) allows the user to specify the central meridian of the reprojected coordinates, defaulting to 0. Open layers has a flag which sort of works... I'm not sure what specific approach has been used in Gepnetwork, but multiple polygons does work.
I think that there is a major bug in the way GoeNetwork handles the
search. If the longitudes are -150 and +150 and if the search crosses
the International Date Line then the most Westerly Longitude should be
+150 and the most Easterly Longitude should be -150.
If the search crosses the 0 longitude then the most Westerly Longitude
should be -150 and the most Easterly Longitude should be +150.
We have been using this algorithm for decades and there is no flaw in it
at all. Nearly all the Australian metadata uses this concept and it is
compliant to ISO 19115. GN should follow this algorithm.
My two cents worth.
Thanks.
John Hockaday
Brent Wood (from NZ where this is indeed a common issue!)
--- On Fri, 5/11/12, andrew walsh <awalsh@anonymised.com> wrote:
From: andrew walsh <awalsh@anonymised.com>
Subject: Re: [GeoNetwork-users] spanning the dateline
To: "Douglas Nebert" <ddnebert@anonymised.com>
Cc: geonetwork-users@lists.sourceforge.net
Date: Friday, May 11, 2012, 7:17 PMHi Doug,
Looking at how GN 2.6.3 behaves when you enter a bounding box
on the advanced search form. If you enter a a box crossing dateline with
west-east 100E/-100W,
north/south -14S/-20S Geonetwork.log file indicates a polygon query like this:<geometry>POLYGON((
-100.0000 -20.0000,
100.0000 -20.0000,
100.0000 -14.0000,
-100.0000 -14.0000,
-100.0000 -20.0000))
</geometry>Trouble is this box is completely 'outside' the box you wanted crossing the
dateline.
Instead this box runs across the Greenwich meridian (0). You can
also see same box drawn on the query map. Guess that explains
Karen's result of getting results outside the region of interest. I tried
flipping the west-east co-ords and same problem.Problem is the sofware doesn't know if you wanted your box to cross the
dateline(180) or greenwich (0). As you say you could break the query into 2
parts to get
the results you want but that's tedious. Perhaps as enhancement the user
could specify that the box crossed the Dateline or Greenwich and then
the box gets composed how you want.Cheers,
Andrew
----- Original Message -----
From: "Douglas Nebert" <ddnebert@anonymised.com>
To: "andrew walsh" <awalsh@anonymised.com>; "John Hockaday"
<John.Hockaday@anonymised.com>
Sent: Friday, May 11, 2012 2:56 PM
Subject: Re: [GeoNetwork-users] spanning the dateline> On 5/10/12 10:11 PM, andrew walsh wrote:
>> Hi John,
>>
>> That's how I thought it worked too, e.g if we had a dataset
>> that spanned the dateline we have e.g +176E as the 'westBoundLongitude'
>> and -176W as the 'eastBoundLongitude', in xml:
> Yes, that is the way one should encode either the query or the metadata, but
> the cartesian logic is broken, and can be achieved only through two queries. I
> thought we fixed this about a year ago.
>
>> <gmd:EX_GeographicBoundingBox>
>> <gmd:westBoundLongitude>
>> <gco:Decimal>176</gco:Decimal>
>> </gmd:westBoundLongitude>
>> <gmd:eastBoundLongitude>
>> <gco:Decimal>-176</gco:Decimal>
>> </gmd:eastBoundLongitude>
>> <gmd:southBoundLatitude>
>> <gco:Decimal>-20</gco:Decimal>
>> </gmd:southBoundLatitude>
>> <gmd:northBoundLatitude>
>> <gco:Decimal>-14</gco:Decimal>
>> </gmd:northBoundLatitude>
>> </gmd:EX_GeographicBoundingBox>
>>
>> Karen,
>>
>> Is this the way you had entered your west/east longitudes i.e
>> is the west one +ve and the east one -ve?
>>
>> Andrew
>>
>>
>> ----- Original Message -----
>> From: "john.hockaday"<john.hockaday@anonymised.com>
>> To: "Douglas Nebert"<ddnebert@anonymised.com>
>> Cc:<geonetwork-users@lists.sourceforge.net>
>> Sent: Friday, May 11, 2012 10:50 AM
>> Subject: Re: [GeoNetwork-users] spanning the dateline
>>
>>
>>> Shouldn't it just work by having the most Westerly longitude to be
>>> greater than the most easterly longitude? If the logic is right then the
>>> comparison should be 'Most Westerly Longitude< search longitude< Most
>>> easterly Longitude' etc.
>>>
>>> John
>>>
>>> On Wed, 2012-05-09 at 17:56 -0400, Douglas Nebert wrote:
>>>> On 5/9/12 3:39 PM, Karen Remick wrote:
>>>>> Hello,
>>>>> I have a client with data that spans the dateline (Northern Russian
>>>>> coast to Northern Alaskan coast). Is there a way to do this in
>>>>> geonetwork? Just putting in the lat/long gives us all the longitudes
>>>>> outside the region of interest.
>>>> I believe that we had fixed this in the indexing in GeoNetwork some time
>>>> ago as we dealt with Pacific and global data. The typical way to do this
>>>> was to break the metadata footprint into two records each of which does
>>>> not span the dateline, and allow for queries to be decomposed to do the
>>>> same.
>>>>
>>>> Kai - did you commit this index and search change to GeoNetwork?
>>>>
>>>> Doug.
>>>>> Thanks
>>>>> Karen
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Live Security Virtual Conference
>>>>> Exclusive live event will cover all the ways today's security and
>>>>> threat landscape has changed and how IT managers can respond. Discussions
>>>>> will include endpoint security, mobile security and the latest in malware
>>>>> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
>>>>> _______________________________________________
>>>>> GeoNetwork-users mailing list
>>>>> GeoNetwork-users@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/geonetwork-users
>>>>> GeoNetwork OpenSource is maintained at
>>>>> http://sourceforge.net/projects/geonetwork
>>>>>
>>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Live Security Virtual Conference
>>> Exclusive live event will cover all the ways today's security and
>>> threat landscape has changed and how IT managers can respond. Discussions
>>> will include endpoint security, mobile security and the latest in malware
>>> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
>>> _______________________________________________
>>> GeoNetwork-users mailing list
>>> GeoNetwork-users@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/geonetwork-users
>>> GeoNetwork OpenSource is maintained at
>>> http://sourceforge.net/projects/geonetwork
>>>
>>
>> ------------------------------------------------------------------------------
>> Live Security Virtual Conference
>> Exclusive live event will cover all the ways today's security and
>> threat landscape has changed and how IT managers can respond. Discussions
>> will include endpoint security, mobile security and the latest in malware
>> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
>> _______________________________________________
>> GeoNetwork-users mailing list
>> GeoNetwork-users@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/geonetwork-users
>> GeoNetwork OpenSource is maintained at
>> http://sourceforge.net/projects/geonetwork
>>
>
>
> --
> Douglas D. Nebert
> Senior Advisor for Geospatial Technology, System-of-Systems Architect
> FGDC Secretariat Tel/Fax:+1 503 454-6248 Cell:+1 703 459-5860
>
>------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at http://sourceforge.net/projects/geonetwork
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
GeoNetwork-users mailing list
GeoNetwork-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geonetwork-users
GeoNetwork OpenSource is maintained at http://sourceforge.net/projects/geonetwork