[Geoserver-devel] Multidim WCS @ Geoserver meeting 2

GeoServer GeoServer Developer
1

Just want to confirm:

The people interested in Multidimensional WCS are coming to Geoserver
Meeting #2 tomorrow, right? I have a couple of stupid questions:

Has anyone investigated the options for requesting data via time? I'd
really like to see a comparison of how one would accomplish this using
GADS-WCS, Galeon IE, GML 3, "Regular WCS"?

Can a "Regular WCS" host a coverage with a CompoundCRS?

Is a BBOX defined to be 2D, or does it track with the CRS of the coverage
(e.g., can it become 3D or 4D given a 3D CRS or a 4D CompoundCRS?)

Thanks for your kind patience with my ignorance.

Bryce

2

Hi,

I cant make the times for this meeting unless I get bitten by a vampire later today.

However, I have had some thoughts about handling time...

Case in question is an SDI for the marine sector. Also been looking at Meterological forecasts. Lots of issues relating to time ]-;

Time is complex - at the very least we have four different co-existing semantics for time:
- time of observation
- time the observation pertains to (eg a model run forecasting the future)
- optionally, time the forecast is valid for (before a more up to date one will supercede it)
- time of publication

On top of that, we have "granularity"
* daily mean
* weekly
* monthly
* yearly
* three-day rolling avergae

etc

data needs to have the same granularity to be comparable, in general.

Then there is currency
* start date
* end date
* how current is current?
* is this the best available (most current)

A couple of principles we tried to follow:
1) multiple services should be encouraged to have common view of time
2) WMS, WFS and WCS should share common semantics for time

As a result, we created formal profiles of WMS for time enablement. WCS is akin to WMS. (These documents will be available online once the portal server environment is available for us to install the portal - I'll send the URL - contact me for the word docs).

These profiles establish vocabularies that allow us to
a) discover services in the catalogue
b) all the user to send meaningful requests to specific services

I would suggest that from an implementation point of view, WFS and WCS views of the same data should be consistently configured somehow, even if a special case named dimension is used for one interface (TIME=) and an ordinary property is used for WFS

(Hopefully we have established by this point that the exact semantics of TIME might be a little complex - and mapping TIME to a single axis of an n-D coverage is only the most trivial case. At the very least we should consider what it takes to be consistent within a WMS that operates across a set of features with an equivalent time property as well as a raster coverage.)

The correct solution will be the simplest one that does not break the different views :slight_smile:

Rob

Bryce L Nordgren wrote:

Just want to confirm:

The people interested in Multidimensional WCS are coming to Geoserver
Meeting #2 tomorrow, right? I have a couple of stupid questions:

Has anyone investigated the options for requesting data via time? I'd
really like to see a comparison of how one would accomplish this using
GADS-WCS, Galeon IE, GML 3, "Regular WCS"?

Can a "Regular WCS" host a coverage with a CompoundCRS?

Is a BBOX defined to be 2D, or does it track with the CRS of the coverage
(e.g., can it become 3D or 4D given a 3D CRS or a 4D CompoundCRS?)

Thanks for your kind patience with my ignorance.

Bryce

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Geoserver-devel mailing list
Geoserver-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel
  

3

Hello,

Has anyone investigated the options for requesting data via time? I'd
really like to see a comparison of how one would accomplish this using
GADS-WCS, Galeon IE, GML 3, "Regular WCS"?

GADS-WCS is expected to provide similar format for requesting data as is
currently provided by GADS.

GADS assumes that one file represents 1 timestep which is a limitation.

GADS supports the following:

1. dataRequest()

User can choose the data format
Can easily download subsets of data
Uses start-stride-count semantics (familiar in community)

e.g.
dataRequest("FOAM_NINTH", "temperature", "CDF",
  "t", 0, 1, 20,
  "z", 0, 1, -1,
  "y", 100, 4, 400,
  "x", 300, 4, 600)

2. dataRequestNatural()
Same as dataRequest() but in natural units (degrees, metres .)

e.g.

dataRequestNatural("FOAM_NINTH", "temp", "CDF",
"t", "2004-06-01 00:00:00", "2004-06-22 00:00:00",
"z", "0", "10",
"y", "42", "64",
"x", "-26", "9")

One area which is not clear under current WCS spec is how to identify which
variable to request from a dataset.

Cheers,

Adit

-----Original Message-----
From: geoserver-devel-bounces@lists.sourceforge.net
[mailto:geoserver-devel-bounces@lists.sourceforge.net] On Behalf Of Rob
Atkinson
Sent: 08 August 2006 01:21
To: Bryce L Nordgren
Cc: geoserver-devel@lists.sourceforge.net
Subject: Re: [Geoserver-devel] Multidim WCS @ Geoserver meeting 2

Hi,

I cant make the times for this meeting unless I get bitten by a vampire
later today.

However, I have had some thoughts about handling time...

Case in question is an SDI for the marine sector. Also been looking at
Meterological forecasts. Lots of issues relating to time ]-;

Time is complex - at the very least we have four different co-existing
semantics for time:
- time of observation
- time the observation pertains to (eg a model run forecasting the future)
- optionally, time the forecast is valid for (before a more up to date
one will supercede it)
- time of publication

On top of that, we have "granularity"
* daily mean
* weekly
* monthly
* yearly
* three-day rolling avergae

etc

data needs to have the same granularity to be comparable, in general.

Then there is currency
* start date
* end date
* how current is current?
* is this the best available (most current)

A couple of principles we tried to follow:
1) multiple services should be encouraged to have common view of time
2) WMS, WFS and WCS should share common semantics for time

As a result, we created formal profiles of WMS for time enablement. WCS
is akin to WMS. (These documents will be available online once the
portal server environment is available for us to install the portal -
I'll send the URL - contact me for the word docs).

These profiles establish vocabularies that allow us to
a) discover services in the catalogue
b) all the user to send meaningful requests to specific services

I would suggest that from an implementation point of view, WFS and WCS
views of the same data should be consistently configured somehow, even
if a special case named dimension is used for one interface (TIME=) and
an ordinary property is used for WFS

(Hopefully we have established by this point that the exact semantics of
TIME might be a little complex - and mapping TIME to a single axis of an
n-D coverage is only the most trivial case. At the very least we should
consider what it takes to be consistent within a WMS that operates
across a set of features with an equivalent time property as well as a
raster coverage.)

The correct solution will be the simplest one that does not break the
different views :slight_smile:

Rob

Bryce L Nordgren wrote:

Just want to confirm:

The people interested in Multidimensional WCS are coming to Geoserver
Meeting #2 tomorrow, right? I have a couple of stupid questions:

Has anyone investigated the options for requesting data via time? I'd
really like to see a comparison of how one would accomplish this using
GADS-WCS, Galeon IE, GML 3, "Regular WCS"?

Can a "Regular WCS" host a coverage with a CompoundCRS?

Is a BBOX defined to be 2D, or does it track with the CRS of the coverage
(e.g., can it become 3D or 4D given a 3D CRS or a 4D CompoundCRS?)

Thanks for your kind patience with my ignorance.

Bryce

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job

easier

Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Geoserver-devel mailing list
Geoserver-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel
  
-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job
easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Geoserver-devel mailing list
Geoserver-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel

4

Ciao guys,
I will be there this evening.
Sorry but I have no time right now, I will discuss these topics these evening.

Simone.

On 8/8/06, Adit Santokhee <ads@anonymised.com> wrote:

Hello,

> Has anyone investigated the options for requesting data via time? I'd
> really like to see a comparison of how one would accomplish this using
> GADS-WCS, Galeon IE, GML 3, "Regular WCS"?

GADS-WCS is expected to provide similar format for requesting data as is
currently provided by GADS.

GADS assumes that one file represents 1 timestep which is a limitation.

GADS supports the following:

1. dataRequest()

User can choose the data format
Can easily download subsets of data
Uses start-stride-count semantics (familiar in community)

e.g.
dataRequest("FOAM_NINTH", "temperature", "CDF",
       "t", 0, 1, 20,
       "z", 0, 1, -1,
       "y", 100, 4, 400,
       "x", 300, 4, 600)

2. dataRequestNatural()
Same as dataRequest() but in natural units (degrees, metres .)

e.g.

dataRequestNatural("FOAM_NINTH", "temp", "CDF",
"t", "2004-06-01 00:00:00", "2004-06-22 00:00:00",
"z", "0", "10",
"y", "42", "64",
"x", "-26", "9")

One area which is not clear under current WCS spec is how to identify which
variable to request from a dataset.

Cheers,

Adit

-----Original Message-----
From: geoserver-devel-bounces@lists.sourceforge.net
[mailto:geoserver-devel-bounces@lists.sourceforge.net] On Behalf Of Rob
Atkinson
Sent: 08 August 2006 01:21
To: Bryce L Nordgren
Cc: geoserver-devel@lists.sourceforge.net
Subject: Re: [Geoserver-devel] Multidim WCS @ Geoserver meeting 2

Hi,

I cant make the times for this meeting unless I get bitten by a vampire
later today.

However, I have had some thoughts about handling time...

Case in question is an SDI for the marine sector. Also been looking at
Meterological forecasts. Lots of issues relating to time ]-;

Time is complex - at the very least we have four different co-existing
semantics for time:
- time of observation
- time the observation pertains to (eg a model run forecasting the future)
- optionally, time the forecast is valid for (before a more up to date
one will supercede it)
- time of publication

On top of that, we have "granularity"
* daily mean
* weekly
* monthly
* yearly
* three-day rolling avergae

etc

data needs to have the same granularity to be comparable, in general.

Then there is currency
* start date
* end date
* how current is current?
* is this the best available (most current)

A couple of principles we tried to follow:
1) multiple services should be encouraged to have common view of time
2) WMS, WFS and WCS should share common semantics for time

As a result, we created formal profiles of WMS for time enablement. WCS
is akin to WMS. (These documents will be available online once the
portal server environment is available for us to install the portal -
I'll send the URL - contact me for the word docs).

These profiles establish vocabularies that allow us to
a) discover services in the catalogue
b) all the user to send meaningful requests to specific services

I would suggest that from an implementation point of view, WFS and WCS
views of the same data should be consistently configured somehow, even
if a special case named dimension is used for one interface (TIME=) and
an ordinary property is used for WFS

(Hopefully we have established by this point that the exact semantics of
TIME might be a little complex - and mapping TIME to a single axis of an
n-D coverage is only the most trivial case. At the very least we should
consider what it takes to be consistent within a WMS that operates
across a set of features with an equivalent time property as well as a
raster coverage.)

The correct solution will be the simplest one that does not break the
different views :slight_smile:

Rob

Bryce L Nordgren wrote:
> Just want to confirm:
>
> The people interested in Multidimensional WCS are coming to Geoserver
> Meeting #2 tomorrow, right? I have a couple of stupid questions:
>
> Has anyone investigated the options for requesting data via time? I'd
> really like to see a comparison of how one would accomplish this using
> GADS-WCS, Galeon IE, GML 3, "Regular WCS"?
>
> Can a "Regular WCS" host a coverage with a CompoundCRS?
>
> Is a BBOX defined to be 2D, or does it track with the CRS of the coverage
> (e.g., can it become 3D or 4D given a 3D CRS or a 4D CompoundCRS?)
>
> Thanks for your kind patience with my ignorance.
>
> Bryce
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job
easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> Geoserver-devel mailing list
> Geoserver-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/geoserver-devel
>

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job
easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Geoserver-devel mailing list
Geoserver-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Geoserver-devel mailing list
Geoserver-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel

--
-------------------------------------------------------
Eng. Simone Giannecchini
President /CEO GeoSolutions

http://www.geo-solutions.it

-------------------------------------------------------

5

Rob Atkinson <rob@anonymised.com> wrote on 08/07/2006 06:20:47 PM:

Hi,

I cant make the times for this meeting unless I get bitten by a vampire
later today.

Wow. We only have Grizzlies. Good luck. :slight_smile:

However, I have had some thoughts about handling time...

All thoughts welcome. However, in mean time I discovered WMS 1.3.0,
Normative Annexes C & D. Looks like that might be the default unless
compelling arguments can be made to depart...

As a result, we created formal profiles of WMS for time enablement. WCS
is akin to WMS. (These documents will be available online once the
portal server environment is available for us to install the portal -
I'll send the URL - contact me for the word docs).

These profiles establish vocabularies that allow us to
a) discover services in the catalogue
b) all the user to send meaningful requests to specific services

Cna you stick them on my wiki page as attachments, or are they sensitive?
Maybe add a section under "Related Efforts" to describe them?

http://docs.codehaus.org/display/GEOS/Multidimensional+WCS

(I typed that from memory, I hope its right.)

I really want to look at this because it sounds like you put thought into
it...

Bryce

6

Summary of WCS part of mtg for those who couldn't make it:

1] Bryce found Annex C & D of the WMS 1.3.0 spec, specifically treating
multidimensional data.
2] We agreed to adopt this technique (even tho it's a WMS spec, not WCS)
because it may reflect the direction taken by the next WCS spec.
3] We're investigating how to handle various common tasks we want to ensure
we can handle. In general, we want to live within the spec as long as we
can, then invent our own solution only when we must.
4] Adit is prototyping a demo Geoserver/GADS to be ready by next week
(17th.) Simone/Alessio are helping.

People with action items from this meeting should be filling in their
section of the Multidimensional WCS wiki page over the next week.

http://docs.codehaus.org/display/GEOS/Multidimensional+WCS

If you think this strategy is dumb, speak up now.

Bryce

7

Bryce L Nordgren wrote:

1] Bryce found Annex C & D of the WMS 1.3.0 spec, specifically treating
multidimensional data.
2] We agreed to adopt this technique (even tho it's a WMS spec, not WCS)
because it may reflect the direction taken by the next WCS spec.
3] We're investigating how to handle various common tasks we want to ensure
we can handle. In general, we want to live within the spec as long as we
can, then invent our own solution only when we must.

If you think this strategy is dumb, speak up now.

No - not dumb. However, the thing to do is to recognise that the interface is a service binding onto the internal model.

We may not be able to get there in one step, but a coverage is also a feature, so we ought to be able to invoke a set of standard (and extensible) functions on the coverage via the WFS interface.

WMS + WCS = a small set of predefined functions
WFS + WPS = generalisations (which begs the question why we need both :slight_smile: )

I'd appreciate feedback from those closer to the current code base of how close we are to being able to hold a coverage data store and access it with a WFS filter.
If the answer is simply - we have stubs for the operations but havent yet implemented them, thats still a comfort. If we eblieve data is fundamentally different because of its storage artefact, thats going to cause more pain as we will increasingly deal with "discrete coverages" packed into netCDF etc.

Rob A.

Bryce

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Geoserver-devel mailing list
Geoserver-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel