[vtk-developers] vtkGL2PSExporter: need feedback.

[Jeff Lee]

more comments...

Prabhu Ramachandran wrote:

>Hi,
>
>Thanks again for the feedback.  Comments follow.
>
>  
>
>>>>>>"JL" == Jeff Lee <jeff at cdnorthamerica.com> writes:
>>>>>>            
>>>>>>
>
>    >>  I beg to differ.  Very few classes are used in every
>    >> situation.  If vtkXYPlotActor belongs in VTK, a gl2ps writer
>    >> would also make a good addition.
>    >> 
>    JL> I don't understand your comparison with vtkXYPlotActor here,
>    JL> but it looks like you misunderstood my argument.  If someone
>
>Very likely I misunderstood.  I thought your argument was that gl2ps
>is not useful all the time for the different types of visualizations.
>Mainly transparent actors and complex scenes.  So my argument was that
>not all classes are useful all the time anyway and cited the example
>vtkXYPlotActor.
>
>    JL> has a scene and wants a hardcopy and uses your class, there is
>    JL> a good chance that the resulting picture will be unacceptable.
>    JL> There is no adjustable parameter to make your class produce an
>    JL> acceptable picture if the user has transparency.  You will
>
>So what?  There are limitations to everything.
>
Good point.  Agreed.

>
>    JL> also need to provide hooks for the sorting and culling
>    JL> options.  In contrast, vtkWindowToImageFilter does work well
>    JL> in all situations, with an adjustable parameter of
>
>Unfortunately, it is not good enough in all situations.  Just imagine
>generating 20 graphs (even simple 2D ones) and then realize that you
>have to rescale the plots and the images show up badly.  Its worse if
>its a paper that you sent a year ago and then they want nicer pictures
>now and you need to figure out how to re-generate those 20 images.  If
>the images were postscript there are no such issues.
>
These are new requirements.  It sounds like you have found gl2ps and are 
now coming up with requirements on-the-fly to justify it.  Please, just 
state why we need this.  Every time we argue, a new requirement appears.  

>
>I'll admit that this is a contrived example but I'm not buying the
>argument that vector graphics are useless and can be replaced by
>bitmapped images.
>
>    JL> magnification, and produces quality hardcopy without trying to
>    JL> turn your printer's postscript rendering engine into a gl JL>
>    rendering engine.
>
>Well, its unfortunate that the printer is treated like a rendering
>engine but vector graphics has its uses.
>
Postscript is a very powerful language.  Why don't we write a 
vtkPostscriptRenderWindow?  How about vtkPostscriptActor/Mapper.

>
>[snip]
>
>    JL> There is no such thing as gl to ps.  I feel like I'm arguing
>    JL> about the principle of using postscript as a rendering engine.
>
>If this argument holds then what do you say about VRML output?  Do you
>think it is correct to use a web browser as a rendering engine?
>
Please!  There is no comparison between vrml and postscript.  If you 
could describe the complexity of a gl scene in postscript the way vrml 
can, then there would be no argument, and we'd all be doing it.  All I 
was saying is that this discussion doesn't seem to be about anything 
constructive, because we are not working with a list of requirements.  I 
don't like arguing requirements and implementation in the same breath - 
it is confusing.

>
>Honestly, I dont care if it ends up that PS is being used as a
>rendering engine.  My point is absolutely simple.  Vector graphics
>output is useful and desired by quite a few folks.  VTK does not
>generate vector graphics output at all even for the simplest of
>scenes.  By adding vtkGL2PSExporter we allow people to try out gl2ps.
>If it works for them fine if not let them go back to using images or
>whatever is convenient.  Why deny someone the opportunity to use the
>class.  Its not like this is the first time someone has asked for
>this.  Besides if gl2ps was so evil why has it gotten to where it is
>today (version 0.71)?
>
It's not evil, just inherently limited, and perhaps not appropriate to 
have in vtk.  I'm not advocating denying the vtk community of vector 
postscript hardcopy.  Honestly, just state the requirements that you are 
trying to satisfy, don't just come up with all the features of gl2ps and 
say that if we don't have them that we are denying someone something. 
 It would be nice to summarize all the reasons you came up with as a 
result of our arguing, because some of them are good ones.  I do think 
that the reason we don't already have this capability in vtk is because 
of the limitations of converting gl to postscript - nothing more.  If 
you are arguing that the limitations are worth it because we can then 
save simple scenes and 2d plots and rescale them at a later date, then 
fine.  It just wasn't clear to me that you were aware of the 
limitations, and it initially sounded like you were advertising 
something that isn't possible.

Respectfully,
-Jeff

-- 
Jeff Lee <jeff at cdnorthamerica.com>
Software Engineer
CD North America
21 Lafayette Street, Suite 230
Lebanon, NH 03766
Tel:    (603) 643-9993 ext. 109
Fax:    (603) 643-9994