Hello all
I just completed the work on extended floating point precision in matrix
multiplication and inversion. This is new work compared to what we had
in Geotk. I added in the package javadoc [1] a small section, namely
"Extended floating point precision", which try to explain why we care
about the accuracy of those operations.
Now the last work needed is to port existing Geotk code for handling of
NaN values and some special kind of non-square matrix, and I should be
done with matrix. The next work will be on geodetic datum.
Martin
[1]
https://builds.apache.org/job/sis-jdk7/site/apidocs/org/apache/sis/referencing/operation/matrix/package-summary.html
