Slicer3:MiniRetreat February 7-8, 2007

Jim Miller and Dan Blezek will meet with Steve Pieper and company on Feb. 7-8, 2007 for a mini-retreat. This will be a detailed work session focusing on the following topics:

• Python integration - We'll be looking at a Python package for providing publication quality 2D plots. At this meeting, we'll determine the feasibility of this approach and develop an initial design. Using this Python package, opens the door for incorporating additional aspects of NumPy and SciPy as well. This could provide a MATLAB style interface to images for rapid prototyping.
• Other viewers - Taking the aforementioned 2D plotting capability as an additional viewer of data that can be added to the Slicer display, there are requests for other types of viewers as well. One that GE has interest in for an internal project is a CSV viewer to display Comma Separated Value files. Other viewers could be plain text viewers, HTML, XML.
• Coordinate frames for plugins - Providing multiple datasets to a plugin implies we are also providing the plugin with the relative coordinate frames, either explictly or implicitly. We need to finalize a design whereby volumes and models are handled uniformly with respect to coordinate frames, in particular local verses world and LPS verses RAS.
• Binary File Descriptor Library - We would like to use the Binary File Descriptor Library for peeking into command line executable modules for global symbols. Libbfd can provide a 50X decrease in module discovery time. However, there are some issues with 64bit Linux installations. Also, libbfd is not installed by default on all Linux and Mac configurations.
  • We now have a TRY_COMPILE that mimics how Slicer uses libbfd, so if the TRY_COMPILE fails, Slicer will not be built with libbfd support.
  • Still need to figure out how to get these Linux and Mac hosts to have the proper libbfd support.
• DWI/DTI Plugins - We would like to finalize the command line module handling on DWI and DTI data. This should be the last mile integration step.
• Usability - Addressing Ron's gentle reminders to make Slicer easy to use.
• Editor - We can take a look at the editor module and discuss strategies to improve performance.
• Testing - Develop plans for incorporating more Slicer Regression Tests (SRT) which start Slicer, load data, change a module parameter, run a module, regression test result. We also need a test presents the CMake environment (Cache, etc.) as a note.