[vtk-developers] changes in temporal support in VTK

Pat Marion pat.marion at kitware.com
Fri Apr 27 14:37:16 EDT 2012


It sounds like a reasonable change to me (assuming the community is willing
to accept the backward incompatibility.)  I've never used
vtkTemporalDataSet, but I have used CONTINUE_EXECUTION.  Is it correct to
say that this change doesn't remove any core functionality- by leveraging
CONTINUE_EXECUTION a filter can accumulate multiple timesteps internally,
with the same efficiency and memory usage as was possible before the
change?  I like that removing vtkTemporalDataSet simplifies things since
there is no longer two different ways to accomplish the same task.

Pat


On Fri, Apr 27, 2012 at 1:40 PM, Yuanxin Liu <leo.liu at kitware.com> wrote:

> Hi,
>   We are currently considering a few backward-incompatible changes to the
> temporal support in the VTK pipeline and would like to hear comments.  We
> are also interested in hearing other suggestions to improve handling of
> time-dependent data.
>
>   The main change is to remove the use of vtkTemporalDataSet and to limit
> the time request to be a single time step, i.e.
>
>    - Currently, a filter can request objects from multiple time steps by
>    setting the key UPDATE_TIME_STEPS to a vector of doubles; the resulting
>    objects get wrapped by the pipeline into a single vtkTemporalDataSet object.
>    - After the change, a filter can only request a single time step by
>    setting the key UPDATE_TIME_STEP to a single double.  The "wrapping" no
>    longer happens.
>
>   For existing filters that depend on multiple time steps,  we will modify
> them to use the CONTINUE_EXECUTION key to iteratively execute the pipeline
> and store the objects from the iterations.   There will be super
> classes/utility classes to make this easy.
>
>   The main motivation is that, by pushing the temporal functionality from
> the core pipeline to the filters, the execution pipeline becomes simpler
> and easier to maintain. Some bugs, e.g.( ParaView bug Id=6662<http://www.paraview.org/Bug/view.php?id=6662>)
> will go away as a result.
>
> Leo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20120427/b37662c5/attachment.html>


More information about the vtk-developers mailing list