VTK 5.6 Release Planning: Difference between revisions

From KitwarePublic
Jump to navigationJump to search
Line 25: Line 25:
== API Changes ==
== API Changes ==


?
* the use of <tt>vtkMath::Random()</tt>, <tt>vtkMath::RandomSeed()</tt>, <tt>vtkMath::GetSeed()</tt>, <tt>vtkMath::Gaussian()</tt> is deprecated (there are error-prone and not deterministic (which is important for reproducibility of regression testing) as they use global static information. Instead, you should create an object of class <tt>vtkMinimalStandardRandomSequence</tt> to get a sequence of pseudo-random number following a uniform distribution. You should create an object of class <tt>vtkBoxMuellerRandomSequence</tt> to get a sequence of pseudo-random number following a Gaussian distribution.


== OpenGL requirements ==
== OpenGL requirements ==

Revision as of 18:35, 16 March 2010

Planned Schedule

What's New in VTK 5.6?

New features in VTK since the 5.4 release include:

  • Improved chart support
  • Classes moved from vtkEdge (vtkGPUVolumeRayCastMapper & associated classes. What else?)
  • 211 new C++ classes.
  • 91 new C++ tests running on most dashboards
  • 17 new C++ examples that demonstrate how to use VTK effectively.
  • Zhanping
    • More robust EnSight reader.
    • Addition of Tecplot reader.
    • Accurate anti-jagginess streamline integrator
    • Incremental / dynamic octree-based point locator for fast point location and insertion of non-uniformly distributed points (3 times faster than the previous version).
    • Integration of robust cell location to vector evaluation on complex grids.
    • High-performance table-based clipper for various kinds of grids (one to two orders of magnitude faster than the previous version).
    • Addition of GPU-based real time Line Integral Convolution (LIC) with enhanced image contrast and LIC+geometry compositing for dense vector field / flow visualization.

API Changes

  • the use of vtkMath::Random(), vtkMath::RandomSeed(), vtkMath::GetSeed(), vtkMath::Gaussian() is deprecated (there are error-prone and not deterministic (which is important for reproducibility of regression testing) as they use global static information. Instead, you should create an object of class vtkMinimalStandardRandomSequence to get a sequence of pseudo-random number following a uniform distribution. You should create an object of class vtkBoxMuellerRandomSequence to get a sequence of pseudo-random number following a Gaussian distribution.

OpenGL requirements

Same as VTK 5.2 (?): See this section in the Frequently asked questions (FAQ)