There is a chunk of Xt code in VTK and I’m wondering if people still use VTK within an Xt based app. Or has everything migrated to xcb or Qt etc?
Open Inventor still has Xt-based API, so there might be a reason, since it represents a maintenance cost to its vendor.
Hearing no response…now that 9.0 is branched, maybe removal for 10.0 would be something to consider?
Here at work we identified a single application that uses VTK with Motif, which is based on Xt.
Ah, sorry. I hadn’t realized Open Inventor was VTK-based.
No! Where did I say that? OpenInventor is another 3D API. I cited it because they keep a Xt interface for some reason (e.g. applications that use the legacy UI).
Oh. Why would their support of Xt affect our decision to use it? I don’t think we know of any customers using the Xt APIs through VTK (the majority probably use ParaView and therefore the Qt stuff).
I just cited an example.
“that uses VTK”
So does OpenInventor use VTK or not?
Why are you asking that? My goodness…
I think you’re trying to stir confusion.
I gave you two reasons to justify Xt:
- Another 3D API uses Xt.
- I cited a client application using VTK with Xt.
Now you can make a better informed decision. If you decide to end Xt support, the consequences are:
- Potential users may adopt OIV (or other 3D APIs) instead of VTK;
- Orphaned legacy applications may demand support for old versions of VTK.
Here’s what I was reading in this thread:
- Anyone using VTK+Xt?
- OpenInventor uses Xt (no indication of relation to VTK)
- I suggest removing it
- You reply that you identified an application
- I assume that it is the OpenInventor you mentioned earlier and reply to that effect
Basically, I was assuming OpenInventor was your application because of the way your “we identified a single application” and my reading that your tone seemed indicate that you had already mentioned it before.
I don’t think we’d make a decision based on such things. If others still supported, e.g., SGI, I think we’d still drop it (if it was a decision to make today).
Oh… I thought OIV was fairly known int the 3D developer community. Sorry about that.
I don’t think we’d make a decision based on such things.
So, what’s the point of starting a discussion? Simply take it off and be done with it.
I appreciate the feedback as it helped me gauge the risks of removing Xt.
I have removed the dependence on Xt from the XRenderWindowInteractor. The class now only relies on Xlib. It made the support for Vulkan a bit easier to implement and cleaner. I’m not sure if that will even be a problem for Xt based apps. They may have to change how they parent a vtk window or add a bit of glue code themselves to do it like the old Xt based interactor style did. It might be a really simple change or nothing at all. But it is out there now. If it turns out to be a big issue we can always bring back the old XRenderWindowInteractor as a XtRenderWindowInteractor. Thanks!
We can stick to the newest VTK version with Xt for a while. Thanks.
I’m a build systems guy; I only use VTK and ParaView in as much as I need to debug things usually .
I meant that just because FrobNarRender supports Xt wouldn’t be that influential (to me) for VTK’s suggested support.