issue warning to user when performing datum shift (OGR, GDAL)

[Dylan Beaudette-2]

Hi everyone,


In light of some recent thought, experimentation, and discussion [1] I
am wondering if it would be at all possible or even desirable to warn
the user when a datum shift operation is about to be performed. Within
the US, and assumming that the NADCON grid was installed with Proj4,
transforming between NAD27 and NAD83 should work just as expected.
However, in other parts of the world, this may not be the case.
Therefore, it might be a good idea to follow the model used in GRASS,
where the user is specifically prompted for datum transform parameters
whenever defining a new locatio, in these situations.

Any thoughts / reasons not to do this?

Cheers,

1. http://casoilresource.lawr.ucdavis.edu/drupal/node/259


Dylan
_______________________________________________
Gdal-dev mailing list
Gdal-dev@...
http://lists.maptools.org/mailman/listinfo/gdal-dev

[Frank Warmerdam]

Dylan Beaudette wrote:

> Hi everyone,
>
>
> In light of some recent thought, experimentation, and discussion [1] I
> am wondering if it would be at all possible or even desirable to warn
> the user when a datum shift operation is about to be performed. Within
> the US, and assumming that the NADCON grid was installed with Proj4,
> transforming between NAD27 and NAD83 should work just as expected.
> However, in other parts of the world, this may not be the case.
> Therefore, it might be a good idea to follow the model used in GRASS,
> where the user is specifically prompted for datum transform parameters
> whenever defining a new locatio, in these situations.
>
> Any thoughts / reasons not to do this?

Dylan,

Well, the most obvious reason not to do it is that the GDAL utilities
are not interactive.  So there is no opportunity to ask.

I think it would be very nice if there was a mechanism to ask stuff
like ogr2ogr to describe what it will do in a transformation but even doing
that would be pretty challenging.

Best regards,
--
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam@...
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | President OSGF, http://osgeo.org

_______________________________________________
Gdal-dev mailing list
Gdal-dev@...
http://lists.maptools.org/mailman/listinfo/gdal-dev

[Dylan Beaudette-2]

On Monday 31 July 2006 13:58, Frank Warmerdam wrote:

> Dylan Beaudette wrote:
> > Hi everyone,
> >
> >
> > In light of some recent thought, experimentation, and discussion [1] I
> > am wondering if it would be at all possible or even desirable to warn
> > the user when a datum shift operation is about to be performed. Within
> > the US, and assumming that the NADCON grid was installed with Proj4,
> > transforming between NAD27 and NAD83 should work just as expected.
> > However, in other parts of the world, this may not be the case.
> > Therefore, it might be a good idea to follow the model used in GRASS,
> > where the user is specifically prompted for datum transform parameters
> > whenever defining a new locatio, in these situations.
> >
> > Any thoughts / reasons not to do this?
>
> Dylan,
>
> Well, the most obvious reason not to do it is that the GDAL utilities
> are not interactive.  So there is no opportunity to ask.
>
> I think it would be very nice if there was a mechanism to ask stuff
> like ogr2ogr to describe what it will do in a transformation but even doing
> that would be pretty challenging.
>
> Best regards,

Hi Frank,

Thanks for the quick reply. I suppose that this is a no-brainer issue for
those that fully understand their data. Perhaps a warning message to STDERR
when datum shifting ?

Cheers,

Dylan



--
Dylan Beaudette
Soils and Biogeochemistry Graduate Group
University of California at Davis
530.754.7341
_______________________________________________
Gdal-dev mailing list
Gdal-dev@...
http://lists.maptools.org/mailman/listinfo/gdal-dev

[Ed McNierney-4]

Dylan -

I understand the concern, but I agree with Frank on the implementation
issues.  May I suggest that another approach would be to avoid the
"default" behavior and require an explicit statement by the user of the
datum shift method to be used?  I realize that backward compatibility
requires accepting current transformations "as is", but a statement
like:

ogr2ogr -s_srs 'prj=latlong datum=NAD83' -t_srs '+proj=utm +zone=11
+datum=NAD27 +method=NADCON' output.shp input.shp

Might help avoid any mystery about what "default" behavior is.  A
compile-time "strict" option might be required to turn this feature on.

I don't think that a command like the one above should warn the user
that a datum shift is being performed, any more than the user should be
warned that the coordinates of the points will be projected.  The
command line asks for a datum shift, so the user shouldn't be surprised
if one is performed.

        - Ed


Ed McNierney
President and Chief Mapmaker
TopoZone.com / Maps a la carte, Inc.
73 Princeton Street, Suite 305
North Chelmsford, MA  01863
Phone: [phone]
Fax: [phone]
ed@...

-----Original Message-----
From: gdal-dev-bounces@...
[mailto:gdal-dev-bounces@...] On Behalf Of Dylan
Beaudette
Sent: Monday, July 31, 2006 4:41 PM
To: gdal-dev
Subject: [Gdal-dev] issue warning to user when performing datum shift
(OGR,GDAL)

Hi everyone,


In light of some recent thought, experimentation, and discussion [1] I
am wondering if it would be at all possible or even desirable to warn
the user when a datum shift operation is about to be performed. Within
the US, and assumming that the NADCON grid was installed with Proj4,
transforming between NAD27 and NAD83 should work just as expected.
However, in other parts of the world, this may not be the case.
Therefore, it might be a good idea to follow the model used in GRASS,
where the user is specifically prompted for datum transform parameters
whenever defining a new locatio, in these situations.

Any thoughts / reasons not to do this?

Cheers,

1. http://casoilresource.lawr.ucdavis.edu/drupal/node/259


Dylan
_______________________________________________
Gdal-dev mailing list
Gdal-dev@...
http://lists.maptools.org/mailman/listinfo/gdal-dev

_______________________________________________
Gdal-dev mailing list
Gdal-dev@...
http://lists.maptools.org/mailman/listinfo/gdal-dev

[Dylan Beaudette-2]

Ed -

I agree completely - The user is responsible for how he/she chooses to
use/mis-use the software.

Perhaps I as asking for something (that most of us don't want for the stated
reasons) but which is worth mentioning when someone is getting started. Maybe
a couple good mistakes are a better reminder than a casual notice from  the
software.

Cheers,

Dylan

On Monday 31 July 2006 17:10, Ed McNierney wrote:

> Dylan -
>
> I understand the concern, but I agree with Frank on the implementation
> issues.  May I suggest that another approach would be to avoid the
> "default" behavior and require an explicit statement by the user of the
> datum shift method to be used?  I realize that backward compatibility
> requires accepting current transformations "as is", but a statement
> like:
>
> ogr2ogr -s_srs 'prj=latlong datum=NAD83' -t_srs '+proj=utm +zone=11
> +datum=NAD27 +method=NADCON' output.shp input.shp
>
> Might help avoid any mystery about what "default" behavior is.  A
> compile-time "strict" option might be required to turn this feature on.
>
> I don't think that a command like the one above should warn the user
> that a datum shift is being performed, any more than the user should be
> warned that the coordinates of the points will be projected.  The
> command line asks for a datum shift, so the user shouldn't be surprised
> if one is performed.
>
> - Ed
>
>
> Ed McNierney
> President and Chief Mapmaker
> TopoZone.com / Maps a la carte, Inc.
> 73 Princeton Street, Suite 305
> North Chelmsford, MA  01863
> Phone: [phone]
> Fax: [phone]
> ed@...
>
> -----Original Message-----
> From: gdal-dev-bounces@...
> [mailto:gdal-dev-bounces@...] On Behalf Of Dylan
> Beaudette
> Sent: Monday, July 31, 2006 4:41 PM
> To: gdal-dev
> Subject: [Gdal-dev] issue warning to user when performing datum shift
> (OGR,GDAL)
>
> Hi everyone,
>
>
> In light of some recent thought, experimentation, and discussion [1] I
> am wondering if it would be at all possible or even desirable to warn
> the user when a datum shift operation is about to be performed. Within
> the US, and assumming that the NADCON grid was installed with Proj4,
> transforming between NAD27 and NAD83 should work just as expected.
> However, in other parts of the world, this may not be the case.
> Therefore, it might be a good idea to follow the model used in GRASS,
> where the user is specifically prompted for datum transform parameters
> whenever defining a new locatio, in these situations.
>
> Any thoughts / reasons not to do this?
>
> Cheers,
>
> 1. http://casoilresource.lawr.ucdavis.edu/drupal/node/259
>
>
> Dylan
> _______________________________________________
> Gdal-dev mailing list
> Gdal-dev@...
> http://lists.maptools.org/mailman/listinfo/gdal-dev

--
Dylan Beaudette
Soils and Biogeochemistry Graduate Group
University of California at Davis
530.754.7341
_______________________________________________
Gdal-dev mailing list
Gdal-dev@...
http://lists.maptools.org/mailman/listinfo/gdal-dev

[Matt Wilkie-2]

> I agree completely - The user is responsible for how he/she chooses to
> use/mis-use the software.

So if I understand this correctly, the summary of repsonses is: this is
a logical bug rather than functional one, and the fix is not
implementation but documentation.

The first comment from Paul K on the CA Soils site would be a good
"patch", reproduced here and heavily edited:

    There is no such thing as an accurate "default" set of datum
transformation parameters for a datum. OGR uses (NADCON) by default,
which is the most accurate available for North America, but in the
general case (worldwide) this is very hard to determine and there
normally is no such thing as a default. The transformation that should
be used depends on the exact area covered, accuracy required etc. In
other words, user beware and do your homework.
    -- http://casoilresource.lawr.ucdavis.edu/drupal/node/259

I'm not sure where in the docs this warning should go though. The OGR
Projections Tutorial (http://ogr.maptools.org/osr_tutorial.html) is
likely to be avoided by end users as it's about how to code projection.
Perhaps the cs2cs man page? http://proj.maptools.org/man_cs2cs.html

cheers,

--
matt wilkie
--------------------------------------------
Geographic Information,
Information Management and Technology,
Yukon Department of Environment
10 Burns Road * Whitehorse, Yukon * Y1A 4Y9
867-667-8133 Tel * 867-393-7003 Fax
http://environmentyukon.gov.yk.ca/geomatics/
--------------------------------------------

_______________________________________________
Gdal-dev mailing list
Gdal-dev@...
http://lists.maptools.org/mailman/listinfo/gdal-dev

[Dylan Beaudette-2]

On Tuesday 01 August 2006 11:02, Matt Wilkie wrote:
> > I agree completely - The user is responsible for how he/she chooses to
> > use/mis-use the software.
>
> So if I understand this correctly, the summary of repsonses is: this is
> a logical bug rather than functional one, and the fix is not
> implementation but documentation.
>
I would agree.

> The first comment from Paul K on the CA Soils site would be a good
> "patch", reproduced here and heavily edited:
>
>     There is no such thing as an accurate "default" set of datum
> transformation parameters for a datum. OGR uses (NADCON) by default,
> which is the most accurate available for North America, but in the
> general case (worldwide) this is very hard to determine and there
> normally is no such thing as a default. The transformation that should
> be used depends on the exact area covered, accuracy required etc. In
> other words, user beware and do your homework.
>     -- http://casoilresource.lawr.ucdavis.edu/drupal/node/259

Wording is concise and to the point.

> I'm not sure where in the docs this warning should go though. The OGR
> Projections Tutorial (http://ogr.maptools.org/osr_tutorial.html) is
> likely to be avoided by end users as it's about how to code projection.
> Perhaps the cs2cs man page? http://proj.maptools.org/man_cs2cs.html
>
> cheers,

I feel that this "reminder" should find its way into the docs somewhere, but
other than possibly a very general README or the like, it would be a little
out of place.

Just to reiterate, I am not suggesting that we dumb down the powerfull, and
often non-interactive tools that we use and love. However, I think that a
carefully placed reminder might save a new user from a bit of grief- and at
the same time promote the idea that people must fully _understand_ their data
and what they are _doing_ to it.

Cheers,


--
Dylan Beaudette
Soils and Biogeochemistry Graduate Group
University of California at Davis
530.754.7341
_______________________________________________
Gdal-dev mailing list
Gdal-dev@...
http://lists.maptools.org/mailman/listinfo/gdal-dev