BUG: Added custom SetMultiGroupDataInformation(). A multi-group data object should have a valid information object at all times.

STYLE: Improved documentation and comments.

BUG: Re-architectured part of class to avoid calling Modified() during RequestData(). Although Kdtree is an internal variable the class was exposing it through a Get method. Changed such that Kdtree is not exposed but there is a way of setting the tree through SetCuts

BUG: Replaced this->SetPointArrayStatus() with this->MetaData->SetPointArrayStatus(). Called method on self causes Modified(). Algorithms are not supposed to call Modified() during RequestData().
STYLE: Fixed various style issues.

BUG: Algorithms should not modify themselves in RequestData(). Added method to modify private data member (ProgressText) without calling Modified()

BUG: Algorithms should not modify themselves in RequestData(). Added method to modify private data member (ScalarLut) without calling Modified()

BUG: ungroup was segfaulting because from did not yet have complete information. Addresses bug# 4784

BUG: Temporarily removed the check for the existence of data information. This is a temporary fix. We should check if the algorithm
produced what was requested and produce an error if it didn't.
However, when such a check is added, all algorithms that do this:
 internalAlg->Update();
 myOutput->ShallowCopy(internalAlg->GetOutput());
will break when running in parallel because ShallowCopy() will
copy that piece related keys which will be 0 of 1.

     catching this.

ENH: Removed the DSP filtering cruft from the ExodusII reader; it belongs
     in a filter of its own now that the pipeline has time.
ENH: Added an (untested) HasModeShapes member to keep up with the Morelands

BUG: Loop over correct number/type of items.

BUG:   Still chasing the memory leak.  Needed to loop through all of NumberOfLabelsAllocated instead of just the NumberOfLabels during the deletion of the text mappers to really get them all.

BUG: Interactor was asking for the bounding box of vertex -1 whenever the pointer was outside the actual tree map.

BUG: Labels were getting clipped incorrectly.  The culprit was an outside-window test against global screen coordinates when we really wanted to compare against the window's size directly.  This was in ConvertToDC and AnalyseLabel.

BUG: NumberOfLabelsAllocated was not getting updated when a new vtkTextMapper was created.  This led to memory leaks.

COMP:Removed the windows and systime includes. These were left over from the old StreamTracer class and I don't need them.
PERF:Moved a check for bad particle location outside the main loop until after integration finishes.