Hey Martin,

Did I miss the 0.3 VOTE? It's possible since I've been so behind
in email but I didn't see a VOTE thread, RESULT thread, etc.

I'll reply to the point sin detail below in a bit.

Cheers,
Chris

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Chris Mattmann, Ph.D.
Senior Computer Scientist
NASA Jet Propulsion Laboratory Pasadena, CA 91109 USA
Office: 171-266B, Mailstop: 171-246
Email: [email protected]
WWW:  http://sunset.usc.edu/~mattmann/
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Adjunct Assistant Professor, Computer Science Department
University of Southern California, Los Angeles, CA 90089 USA
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++






-----Original Message-----
From: Martin Desruisseaux <[email protected]>
Organization: Geomatys
Reply-To: "[email protected]" <[email protected]>
Date: Tuesday, July 30, 2013 11:00 AM
To: Apache SIS <[email protected]>
Subject: Replaced LatLon and LatLonRect

>Hello all
>
>Starting SIS 0.4 work, I removed LatLon and LatLonRect as per [1] and
>[2]. Those classes were deprecated in SIS 0.3 (not yet released). Those
>classes are replaced by DirectPosition2D and Envelope2D, which more
>closely follow ISO/OGC objects.
>
>The LatLonRect.getJavaRectangles() method has been ported to
>Envelope2D.toRectangles() [3]. Some differences in the Envelope2D method
>compared to the LatLonRect one are:
>
>  * Returns an empty array if the envelope is empty.
>  * Do not shift the coordinates by +90°N and +180°E. If such shift is
>    desired, it will be the work of coordinate transformation services.
>  * The longitude axis and the 180° limit is fetched by inspection of
>    the associated Coordinate Reference System (CRS) object, rather than
>    being hard-coded. Longitude is not necessarily the first or second
>    axis, since it depends on the CRS definition. We may have no
>    longitude at all (e.g. projected coordinates). Limit may be 200
>    grads instead than 180°. The wraparound axis may be something else
>    than longitude (e.g. time axis in climatology).
>  * A multi-dimensional variant (> 2D) of this method is provided in
>    GeneralEnvelope.toSimpleEnvelopes().
>
>
>The LatLon.getNormLon() method has been refactored as
>GeneralDirectPosition.normalize(). The main difference is that the new
>method normalizes the ordinate values in-place instead than returning
>normalized longitude, because we don't know in advance which axis will
>be normalized (we may have no longitude at all, etc.).
>
>While the above has been tested locally on my machine, it may take a few
>weeks before it works as described on trunk because it requires an
>implementation of GeographicCRS, which has not yet been ported to SIS.
>I'm starting this port this week.
>
>One remaining class to refactor is LatLonPointRadius. This class
>contains the following methods:
>
>  * getCircularRegionApproximation
>  * getRectangularRegionApproximation
>
>
>I guess that those methods should be moved elsewhere, in some
>computational classes. However I don't know yet where to put them. One
>possible candidate would be org.apache.sis.distance.DistanceUtils, but
>this is not really distance computation... However it may be a temporary
>position since I would like to propose modification to DistanceUtils
>later too (something along the line of the "geodetic calculator"
>provided in the geotk).
>
>Does anyone would like to share comments, though or concerns?
>
>     Martin
>
>
>
>[1] http://issues.apache.org/jira/browse/SIS-68
>[2] http://issues.apache.org/jira/browse/SIS-69
>[3] 
>http://builds.apache.org/job/sis-jdk7/site/apidocs/org/apache/sis/geometry
>/Envelope2D.html#toRectangles%28%29
>

Reply via email to