the "Bypass" feature in imaging
David Gobbi
dgobbi at irus.rri.on.ca
Thu Dec 16 18:30:23 EST 1999
Hi Lisa,
I never use the Bypass method myself. Many of the imaging filters
I use do geometrical operations (like vtkImagePermute,
vtkImageClip, etc) and a Bypass for these filters really makes
no sense. Same with vtkImageMapToRGBA, and pretty much any
dual-input filter.
Perhaps it could be implemented on a class-by-class basis for
the filters where it really is useful.
- David
--
David Gobbi, MSc dgobbi at irus.rri.on.ca
Advanced Imaging Research Group
Robarts Research Institute, University of Western Ontario
On Thu, 16 Dec 1999, Lisa Sobierajski Avila wrote:
> Hello everyone,
>
> As you may know, the VTK pipeline architecture is being reworked for VTK
> 3.0. The main goal of this rework is to simplify the pipeline update
> mechanism. This is being done by making the update process occur mostly at
> the superclass level (vtkSource and vtkDataObject) and clearly defining
> what parts of the update process could / should be overridden. A byproduct
> should be that new filters will be easier to write - and writing a graphics
> filter won't be too much different than writing an imaging filter.
>
> This leads to my question - how many people out there are using the Bypass
> feature in the imaging filter? And, if you are, how awful would it be if it
> goes away? Here are the pros and cons:
>
> Why we should keep the bypass feature: This feature allows you to easily
> turn off a filter in the middle of an imaging pipeline without having to
> keep track of all the other filters.
>
> Why we should eliminate the bypass feature: It complicates the execute
> method in the imaging filters. It is not implemented in every imaging
> filter now (sometimes it is just ignored). If you use it without thought to
> what two filters you would be connecting if you turn Bypass on, then you
> can cause an error or even a segmentation fault in your pipeline.
>
> It is possible to argue that instead of taking it out of the imaging
> filters, we should make it a more general feature of the pipeline (for
> example vtkDataSetToDataSetFilter subclass could bypass, as could
> vtkPolyDataToPolyDataFilter subclasses). Of course, this would then
> complicate even more execute methods, and would mean that many filters
> would have to produce an error if you try to turn Bypass on.
>
> My opinion is that it should be removed. It is a nice feature, but it does
> complicate the execute methods and therefore would make understanding /
> writing filters easier if it is gone. Also, you can simply reconnect your
> pipeline to achieve the same effect. Of course, I don't use this feature so
> I may not have a full understanding of why this feature is used and what
> impact its removal will have.
>
> If you have an opinion, please express it in email to the user's list (and
> do it soon please - since I want to check in all these changes in the next
> few days) If enough people express a strong opinion that it should stay,
> then I will probably leave it in pretty much as is right now, with the goal
> of making it more general in the future.
>
> Thanks,
>
> Lisa
>
>
>
> -----------------------------------------------------------------------------
> This is the private VTK discussion list. Please keep messages on-topic.
> Check the FAQ at: <http://www.automatrix.com/cgi-bin/vtkfaq>
> To UNSUBSCRIBE, send message body containing "unsubscribe vtkusers" to
> <majordomo at gsao.med.ge.com>. For help, send message body containing
> "info vtkusers" to the same address. Live long and prosper.
> -----------------------------------------------------------------------------
>
-----------------------------------------------------------------------------
This is the private VTK discussion list. Please keep messages on-topic.
Check the FAQ at: <http://www.automatrix.com/cgi-bin/vtkfaq>
To UNSUBSCRIBE, send message body containing "unsubscribe vtkusers" to
<majordomo at gsao.med.ge.com>. For help, send message body containing
"info vtkusers" to the same address. Live long and prosper.
-----------------------------------------------------------------------------
More information about the vtkusers
mailing list