[ImageJ-devel] Nonlinear axes work nearly complete

[Curtis Rueden]

Hi ImageJ / ImgLib2 / SCIFIO developers (mostly core developers),

We recently merged the 'nonlinear-goodies' branch of imglib.git to master.
This adds support for nonlinear calibrations in addition to linear ones:

https://github.com/imagej/imglib/commit/3ffc0bec00bb7e71d3395a75caaa9c3460b02d75

We tried to maintain backwards compatibility as much as possible,
deprecating obsolete methods and classes. They will stay there through the
next release (2.0.0-beta-17) and then get removed afterwards, to make
migrating to the new API easier.

We are working on updating SCIFIO and ImageJ2 to use the new API as well.
Work can be seen at:
https://github.com/imagej/imagej/compare/nonlinear-axes
https://github.com/imagej/imagej/compare/nonlinear-goodies
https://github.com/scifio/scifio/compare/nonlinear-calibration-update

Unfortunately I will be out of the office tomorrow, so completion of this
work will have to wait till Monday.

For those of you who were in the office physically or the #imagejdev chat
room: we finally fixed the bug with duplicate X and Y axes:
https://github.com/imagej/imagej/commit/0876338d24b2a4ffea805def141d36edb5acdf1b

So Mark and Barry: no need to dig into anything further here.

I verified that ImageJ2 is now behaving reasonably well against ImgLib2
master now. That means the next steps (for me on Monday) are:

1) Update SCIFIO master to work with latest ImgLib2 code.

2) Cut the 2.0.0-beta-17 release of ImgLib2.

3) Finalize SCIFIO's nonlinear-calibration-update branch to use the beta-17
release and merge it to master once working.

4) Cut a new release of SCIFIO.

4) Update ImageJ2's nonlinear-axes branch to use the ImgLib2 beta-17
release and new SICFIO release.

5) Cherry-pick relevant changes from nonlinear-goodies branch to
nonlinear-axes branch (dividing changes into conceptually related commits).

6) Merge nonlinear-axes branch to ImageJ2 master once fully complete.

In the meantime, please note that the duplicate X/Y axis bug lingers on
ImageJ2 master as things stand. It will not be fixed until ImageJ2 is
pinned against a new release of ImgLib2. (Even if you cherry-picked that
fix to AbstractData above, it does not fully fix the problem without
another fix to ImgLib2's Axes.get method too.)

If anyone has any questions about this, feel free to reply. But mostly this
is just a status update to let everyone know that I'll take care of it, and
asking for others to just let things lie until Monday (though unrelated
pushes to master, in the meantime, are fine).

Thanks,
Curtis
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://imagej.net/pipermail/imagej-devel/attachments/20131003/2fdd804a/attachment.html>