[Paraview] Comparison of Visit and ParaView development

Wed Jun 15 10:18:06 EDT 2016

I will leave it to the community to address some of these concerns since I
am obviously biased.

However, Sven made some misguided and unfair statements and I would like to
address those.

Books: Utkarsh & the ParaView community have put a lot of effort in
developing a User's Guide for ParaView, which is excellent. This is a
community developed effort, which is maintained very regularly and is
available in print and download form in addition to the LaTeX source code.
Yes, this is focused on users rather than developers but the same would be
true for any visualization tool out there. In addition, we are in the
process of open sourcing both VTK books, something which will take some
effort because they need to be converted from Framemaker to LaTeX. The goal
is to combine them in some way and open the development to the community.
In addition, ParaView will soon leverage VTK-m for shared-memory parallel
processing (http://m.vtk.org/) and we are developing two open source books
for that.

Community: First of all, ParaView is a natural extension of VTK. Therefore,
a lot of development happens at VTK level. I encourage everyone to take a
look at the archives of the VTK mailing lists including the developer list.
You will find a very vibrant community that goes well beyond Kitware there.
In fact, we are a minority on those lists very often. There are well over
50 developers who have contributed to VTK over the last few years, from
academia, industry and government. Almost all algorithmic work is done in
VTK. It is true that most of the ParaView application is developed by folks
at Kitware. Only a few folks from the ParaView development team regularly
responds to the mailing lists (kudos to them). Our team is now 43 people
about 20 of whom are focused on VTK and ParaView development.

Applications: There are a number of ParaView based applications out there,
clearly demonstrating ParaView as a viable application platform. Not
counting VTK based applications because there are too many of those to list
here.

* Tomviz: http://www.tomviz.org/
* CMB: http://www.computationalmodelbuilder.org/
* VeloView: http://www.paraview.org/veloview/
* Mantid 3D viewer: http://www.mantidproject.org/

In addition to a number of commercial cloud based platforms based on
ParaViewWeb such as SimScale (https://www.simscale.com/). Some of these are
developed at Kitware but some of the have been almost entirely developed
without very little help from us.

Features: There are a lot of huge features that are making their way into
ParaView. To be fair, these will also make their way into VisIt because
they come from VTK so these are really not differentiators between the two.
The two major ones are the new rendering engine including the use of much
more modern OpenGL features with huge performance increase and a focus on
next generation architectures such as GPUs and Xeon Phis though the
development of VTK-m. In addition, we have a strong focus on in situ
analysis through the development of Catalyst, which I would argue is the
premier in situ library with lots of advanced features not available
anywhere else. Something we demonstrated working very well on 1 million MPI
ranks on the Argonne Blue Gene (Mira). Other more research focused areas
such as Cinema (http://cinemascience.org/) are being pushed in
VTK/ParaView/Catalyst quite heavily right now. All of these focus on
exascale computing, which is US Department of Energy push right now.

Issues with ParaView as a development platform:

* It is true that our developer documentation is lacking. ParaView as a
development platform has never been the main driver so making it more
accessible to the larger community is something that has not been pushed as
much as we would like. Not true for VTK though.

* The ParaView code architecture is fairly complicated. This is because it
has evolved to address the needs of various kinds of applications. A much
broader set of needs that tools like VisIt tend to focus on. This is
because our community targets more than HPC Visualization, which were the
original drivers for both VisIt and ParaView. The kind of applications I
listed above is a good demonstration of this.

I believe that the main differentiator between ParaView and other vis tools
out there is the broad functionality _and_ the code quality. Having the two
together is really tough but our community managed this with a heavy
emphasis on code review and code testing. I strongly recommend that folks
look at the software processes used to develop VTK & ParaView as well as
the huge amount of testing (both test quantity and platform coverage) that
we do before every single commit in addition to nightly. There is a very
good overlap between the CMake, CTest & CDash communities and the
VTK/ParaView development communities and there is very good reason behind
this.

Best,
-berk

On Wed, Jun 15, 2016 at 6:54 AM, Sven Kramer <svenkramer40 at gmail.com> wrote:

> Dear ParaView programmers,
> I would like to have some opinions on the suitability of either ParaView
> or Visit for large scalable visualization applications. When I had the
> first contact to visualization more than 5 years ago, Visit was not quite
> scalable and ParaView seemed the only open source alternative.
> Now I am trying to get into some serious developments, and although most
> papers in the field still cite ParaView as their development framework,
> there seems to be no active developers community except for two or three
> kitware employees. Questions on this list are at the most basic level, and
> nobody seems to be able to answer, who isn't working for kitware.
> My impression is that ParaView has grown so complicated over the years
> that it is no longer usable as a framework for parallel visualization
> development. Only the ParaView application itself is still frequently used.
> On the other hand, Visit seems to have overtaken ParaView in all aspects.
> Most important is the high quality documentation and very active community,
> scalability is now as good as ParaView, and collaborative web visualization
> is coming rapidly.
>
> What is your opinion? Are there any points, where ParaView is clearly
> advantageous over Visit? Or is it used by most people only out of
> convenience, because they have always worked with ParaView?
> Given the horribly out of date books on ParaView and parallel VTK
> development, web documentation mixing three major versions without stating
> which version is referred to, and the dead development discussions I don't
> see any possibility of learning ParaView programming today. Which material
> do you use for tutorials and documentation?
>
> Thank you
> Sven
>
> _______________________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Please keep messages on-topic and check the ParaView Wiki at:
> http://paraview.org/Wiki/ParaView
>
> Search the list archives at: http://markmail.org/search/?q=ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://public.kitware.com/mailman/listinfo/paraview
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/paraview/attachments/20160615/386943dd/attachment.html>