[vtkusers] c# port of VTK

Sean McInerney seanm at nmr.mgh.harvard.edu
Tue Aug 3 00:11:20 EDT 2004


Hi Prabhu,

   I didn't address the issue of virtuals and inheritance within the 
wrapper language since I would most likely implement a subclass in C++ 
and then wrap the derived class. I would be willing to guess that this 
is not everyone's preference and that such users might prefer to extend 
VTK in their non-C++ language of choice.

   Am I understanding the problem?

-Sean

Prabhu Ramachandran wrote:
>>>>>>"SMI" == Sean McInerney <seanm at nmr.mgh.harvard.edu> writes:
> 
> 
>     SMI> Andy,
>     SMI>    I can only conclude that C# is evil. My advice is to
>     SMI>    forswear all
>     SMI> things Microsoft and move to a mountain commune. Save your
>     SMI> soul and use the Python wrappers if you want the advantages
>     SMI> of a "managing" language.  .Net is just another of many MS
>     SMI> products designed to "cut off the air supply" of existing
>     SMI> solutions.
> 
> He is right about the virtuals and the issue is much the same with
> Python or any wrapped language.  Checking to see if a virtual function
> has been created in the target language can be expensive.  How
> expensive depends but it can be significant in some cases.  I don't
> know how it compares to the C# <-> C++ interface though.
> 
> BTW, the VTK Python wrappers do not allow one to override virtuals in
> the target language and still be seen by the C++ side, i.e. if you
> override a virtual in Python and pass an instance of the class to a
> VTK object in C++, it cannot access the overridden function.
> 
> Also, non-MS implementations of the .net environment are available in
> the form of DotGNU and Mono.  Whether they are also nono's I am not
> sure about. :-D
> 
> cheers,
> prabhu



More information about the vtkusers mailing list