[ome-devel] OME 2.6.0 Release

[Ilya Goldberg]

The OME development team is proud to announce the official release of  
OME 2.6.0!
Download for generic UNIX:
http://cvs.openmicroscopy.org.uk/snapshots/ome-2.6.0.tar.gz
Or, a graphical installation wizard for OS X (universal native PPC/ 
Intel):
http://cvs.openmicroscopy.org.uk/snapshots/ome-installer-2.6.0.dmg
The OS X installer includes all software dependencies for both  
architectures, so its a large (~40 Mb) download.

Thanks to all who contributed, and helped us test the release  
candidates!
We will be updating our website (http://openmicroscopy.org) in the  
next day or so to reflect this new official release.

The major changes in this release (with respect to OME 2.4.0) were in  
the web interface, access control, and connectivity with other  
applications.  Reports indicate that this release is also much easier  
to install than OME 2.4.0.

The web interface was revamped to make it entirely HTML template- 
based.  The templates are indexed by the DB, so each user can have  
their own collection of templates to view specific types or  
collections of information.  Since templates are mostly just HTML,  
the user interface can be changed without having to program.  The UI  
allows switching templates on the fly with a drop-down menu.

Access to objects is now controlled through provenance instead of  
containership.  That's a fancy way to say that objects are now owned  
by those who crated them rather than by those who own the objects  
they're contained in.  The other way was not only more convoluted to  
say, it lead to many more inconsistencies.

Two-way integration with Excel (OME-Excel.xls) allows populating  
spreadsheets directly from a live OME server using a VisualBasic  
library included in src/Excel/OME-Excel.xls.  A spreadsheet importer  
allows bulk annotation of images as well as populating generic  
structured data - either from native Excel documents or tab-delimited  
text files.

Connectivity with Matlab is much improved and more robust.  Images  
managed by OME can be accessed from matlab using matlab's imread()  
function.  Extensive testing and validation of OME Analysis Modules  
implemented as functions in matlab (i.e. ".m" files).  Most matlab  
functions can be used unchanged both in "pure matlab" and in OME  
(with appropriate XML wrappers).

Among other features included:
* Better VisBio support.
* Guest access.
* Encryption of passwords in transit over http.
* Graphical OS X Installation "Wizard".
* More robust, easier installation on supported platforms.
* Substantially improved latency and performance for remote access.
* A simplified URL-based API to get OME data as tab-delimited text  
tables (basis for OME-Excel.xls).
* Improved import format support
* Simplified and more robust data curation: Deletion of images,  
datasets, users and module execution results with the ome commander
* New web and command-line wizards make it easy to convert  
spreadsheets into structured OME annotation
* More useable and flexible 3+4 D segmentation with FindSpots.
* Integrated on-line help!

Anticipated major features in our next release - 2.6.1:
* Integration with LDAP (Thanks to Mike McCaughey of Vanderbilt)
* Integration with LOCI's BioFormats (adds support for tons of new  
image types)
* Improvements to integration with LOCI's VisBio
* Better (slightly) security over http to prevent some forms of  
session stealing.
* Allow use of pre-compiled MATLAB modules (i.e. without a MATLAB  
license).
* Robust distributed image analysis over many network-connected "nodes".
Much of this functionality is already in CVS (HEAD), so we anticipate  
this release will be in early 2007.

Enjoy!