[MetaCRS] Testing framework available to any GeoAPI implementation

[Martin Davis]

Norm,

That would be GREAT if you can post a test file!  I'd love to run that thru
Proj4J and see what breaks...  8^)

I know your reasons for liking CSV, and mostly I agree with them.  For
various reasons I often end up using a text editor to work with the files,
and then they are quite difficult to read and get correct.

XML is the obvious other choice for format, and it seems like it might be
easier to read in a text editor.  It also provides some "tech cred".
 Anyway, there's no reason that we can't specify both formats - they should
be easily interchangable.

Martin

On Tue, Aug 16, 2011 at 8:59 AM, Norm Olsen <norm.olsen at autodesk.com> wrote:

>  I have adopted the MetaCRS format for internal testing use, and hope to
> post a test file with some 195,000 test points in the next month or so.  I
> really like the CSV format.  It’s easy to parse (CS-MAP includes a CSV file
> object written in C++) and you can use spreadsheet applications to examine,
> manipulate, and analyze the data fairly easily.****
>
> ** **
>
> Norm****
>
> ** **
>
> *From:* metacrs-bounces at lists.osgeo.org [mailto:
> metacrs-bounces at lists.osgeo.org] *On Behalf Of *Martin Davis
> *Sent:* Tuesday, August 16, 2011 9:33 AM
> *Cc:* metacrs at lists.osgeo.org
> *Subject:* Re: [MetaCRS] Testing framework available to any GeoAPI
> implementation****
>
> ** **
>
> Even better, why not provide these tests in the MetaCRS Test format?  That
> way they are declarative, platform and API-independent, and hence more
> easily usable by all the various CRS projects.****
>
> ** **
>
> (I realize that this format hasnt got much traction beyond - ahem - Proj4J,
> and I have my reservations about CSV as a format, but nevertheless it has
> been presented and formalized and been proven in use).****
>
> ** **
>
> Martin D2****
>
> On Mon, Aug 15, 2011 at 6:36 AM, Martin Desruisseaux <
> martin.desruisseaux at geomatys.fr> wrote:****
>
> On a related topic with my previous email, there is some news on the
> "testing" front (a while ago we talked on this mailing list about a shared
> database for test points).
>
> The GeoAPI project provides a "geoapi-conformance" module which executes
> many JUnit tests on arbitrary implementations of the GeoAPI interfaces. I
> tried to make the test as implementation-independent as possible. For
> example the current implementation uses only the points provided in the
> "Example" sections of EPSG guidance notice, or the example points provided
> by some authoritative sources like the French National Mapping Agency (IGN)
> - they are not points calculated by a particular implementation.
>
> The tests are extensively documented, as you can see from here:****
>
>    -
>    http://www.geoapi.org/geoapi-conformance/apidocs/org/opengis/test/referencing/MathTransformTest.html#testMercator1SP%28%29
>    ****
>
>
> The tests are also customizable. If your implementation does not support
> the calculation of projection derivatives, and if you need to relax the
> tolerance threshold value, just override the test method as below:****
>
> @Override****
>
> public void testMercator1SP() throws FactoryException, TransformException {****
>
>     isDerivativeSupported = false;  // MathTransform.derivative(...) not supported by our implementation.****
>
>     tolerance = 0.1;                // Relax the tolerance threshold value from 1 cm to 10 cm.****
>
>     super.testMercator1SP();****
>
> }****
>
>
> By creating GeoAPI wrappers for Proj.4 (see my previous post), Proj.4 win a
> new tests suite for free :-) (or maybe to be more realist, I should said
> that the tests suite win a validation from a well etablished library).
>
>     Regards,
>
>         Martin****
>
>
> _______________________________________________
> MetaCRS mailing list
> MetaCRS at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/metacrs****
>
> ** **
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/metacrs/attachments/20110816/c7471620/attachment-0001.html