[BioC] EBImage (>= 4.0.0) - browser-dependency display() issue

[Robert Baer]

On 10/7/2012 4:43 PM, Andrzej Oleś wrote:
> We are happy to announce that with Bioconductor release 2.11 the
> package EBImage (>= 4.0.0) can be installed without any of the system
> dependencies that were required by previous versions (ImageMagick,
> GTK+).
>
> Further functionality upgrades include:
> - Perreault's constant time median filter,
> - self-complementary top-hat,
> - streamlined object feature computations,
> - improved 'rotate' / 'translate' functions and a new 'transpose' function,
> - alpha channel support,
> - and performance enhancements.
>
> See the package NEWS file and documentation for details.
>
>
> Andrzej Oles
> EBImage Development Team
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor
It used to take forever to get GTK+ and ImageMagick installed and 
working properly, so I really appreciate your new approach. However, 
there is a browser-dependency with the new display() function.

 > nuc = readImage(system.file('images', 'nuclei.tif', package='EBImage'))
image 510 x 510 x 0, tiles 0 x 0, bps = 8, spp = 1 (output 1), config = 
1, colormap = no
image 510 x 510 x 0, tiles 0 x 0, bps = 8, spp = 1 (output 1), config = 
1, colormap = no
image 510 x 510 x 0, tiles 0 x 0, bps = 8, spp = 1 (output 1), config = 
1, colormap = no
image 510 x 510 x 0, tiles 0 x 0, bps = 8, spp = 1 (output 1), config = 
1, colormap = no
 > display(nuc)

On Windows 7 64-bit with IE 9.0 as my default browser, I get the browser 
starting with the message, "Internet Explorer restricted this web page 
from running scripts or ActiveX controls".  If I click the button,"Allow 
Blocked Content", I still get no display of the images.  The display 
command therefore appears to fail. If I do to use the built-in raster 
display:

display(nuc, method = 'raster'),

or better, display(nuc, method = 'raster', all = TRUE)

the images display quite nicely.

I'm not quite sure why you picked "browser-based display" as the 
default, but it seems there should be checking to see if the appropriate 
browser controls are available if the browser display is to be the 
default.  Just for completeness, if Chrome (my more common default 
browser) or Mozilla Firefox are my defaults, the browser based display 
works fine.  Note here though, that I have a current and 
well-functioning Java that didn't exist by default on the Windows 
platform.  Do you have code that should install an ActiveX control in IE 
that is not functioning properly?  The issue with browser-based display 
as the default is "reliability" not "flexibility".

I do not have Safari or Opera installed on this computer so I have not 
tried them.

Let me finish with a wishlist item.  I would be terrific if there were a 
way to directly tie the OME data model 
http://www.openmicroscopy.org/site/support/file-formats to EBImage 
processing in R.  Being able to deal with multidimensional data and 
manage the mass of metadata associated with images seems destined to be 
the future.  I know this is a BIG WISH not easily fulfilled, but I 
thought I'd express it anyway.

Again, I'm grateful for what you've already contributed with the 
existing EBImage!  Kudos to you on a terrific package!

Rob Baer


Thanks,
Rob Baer