Fa3ien  wrote in message <4707ba6b$1@news.povray.org>:
> With built-in functions, i can write :
> light_source {... area_light ... circular orient adaptive 2}
> or, as well
> light_source {... area_light ... adaptive 2 orient circular}
> 
> I know that almost no programming language allows that kind
> of parsing feature in user-defined functions, but I regret it.

If you do not mind naming all the parameters, a lot of languages can do
that.

Objective Caml can do something like that:

light_source ... ~area_light:true ~shape:`Circular ~adaptative:2 ()

Python has a similar feature, I think.

Perl can emulate that very well using hashes:

light_source ... area_light => 1, shape => "circular", adaptative => 2;

And in fact, almost any language that supports variable number of arguments
can emulate it. For example in C:

light_source(..., "area_light", "circular", "orient", "adaptative", 2, NULL);

This requires a little bit more work inside the user-defined function, but
not much, and some of it can be eased by common helper functions.

Some libraries, like Gtk+, use similar schemes.

Post a reply to this message

> Some aspects I would like the new syntax to provide are:
> 
>   Not from scratch, reuse and inherit notational features from current
> syntax
>   Allow non-programmers to make reasonably good scenes
>   Provide the language with some main features we know as being very useful
> or necessary but diffucult to code (complicated algotithm, advanced maths
> ....).
>   Let the gurus code these

The key, IMO, is : how many gurus will there be.

Currently, the maximum a user can do is a complex macro.  There is a 
good proportion of users who can do that, but these are "small gurus",
since the system doesn't allow them to go very far.

Currently, if someone wants to do more, he have to either write an
external utility (which might be cumbersome to use, hardly portable...)
or do a POV-Ray patch, which requires even more competence.  These
are "big gurus", but represents a smallish fraction of users.

If the new SDL allows enough power while keeping a smooth learning
curve, there can be many "medium-to-big gurus", and everyone would
benefit from these developments.

Fabien.

Post a reply to this message

> However, it has always been stressed that povray -- like Renderman, Mental
> Ray and other renderers -- deals exclusively with the rendering process.

That's a fundamental misconception.  POV-Ray ISN'T a "renderer".
POV-Ray is, and has alwas been, a "rendering system", complete out of 
the box, with the ability to generate scenes through a scripting
system (until POV 2.2), then a more complete SDL (from POV 3).

With Mental Ray, you can't do that, you need a modeler with an interface 
to control it.

Even with Renderman, you can't created decent scenes with a text file,
it totally lacks any feature that would allow text-based modelling.

POV-Ray allows you to do that, not as a side-effect of the SDL, but
precisely BY DESIGN.  Drop that, and you kill POV-Ray (there are
already many nice and free pure renderers out there).

> It was never supposed to be a modeller, yet efforts like yours and many
(snip)

It always had a modelling system, albeit not a graphical one.  CSG,
loops, macros... are there to allow modelling tasks.

> others are only proof to the great strength and flexibility of povray's
> SDL.  And a SDL is a good thing even if regulars users don't use it.

Fabien.

Post a reply to this message

>> (along with the "magical commercial plug-in for lazy GUI users" that 
>> creates fabulous pics out of the box just by pressing a key...).
> 
> Everybody I know who has Bryce has made images by putting a primitive 
> with one of the thousands of pre-made textures, or using the "water and 
> island wizard" or whatever...

And how have this ever produced "fabulous pics" ?

Fabien.

Post a reply to this message

Nicolas Alvarez <nic### [at] gmailisthebestcom> wrote:

> > (along with the "magical commercial plug-in
> > for lazy GUI users" that creates fabulous pics out of the box just by
> > pressing a key...).
>
> Everybody I know who has Bryce has made images by putting a primitive
> with one of the thousands of pre-made textures, or using the "water and
> island wizard" or whatever...

I think the attitude Gilles is complaining of is better expressed as the
idea that commercial apps have a "Make Art" button. I don't think it was
ever a particularly useful insight that commercial apps would have tools
not found in a free app like POVRay, but we are not in the 90s any more and
these days, nobody has any excuse at all - there are plenty of free apps out
there with extremely strong tools similar to those in commercial apps.

People using exclusively POVRay now are doing so by choice, not because of
poverty, and for anyone to complain that other tools make some things
easier would be verging on the ridiculous. Although I haven't seen anyone
mention the Magic Plugin syndrome or a make art button for a while now :)

In Bryce's case it was designed as a landscape tool aimed at home users
IIRC, so I think it's pretty clear why it has wizards for common tasks. I
think they released some recent versions of Bryce free of charge as well,
so again there's nothing I can see that's worth complaining about. I too
wonder how many of the wizard-produced images actually have any
individuality, and I also note that POVRay comes with a substantial library
of predefined textures, scenes and primitives, even if some of them are
getting on a bit these days. And is rather faster, if I remember rightly.

Tom

Post a reply to this message

Fa3ien <fab### [at] yourshoesskynetbe> wrote:

> > However, it has always been stressed that povray -- like Renderman, Mental
> > Ray and other renderers -- deals exclusively with the rendering process.
>
> That's a fundamental misconception.  POV-Ray ISN'T a "renderer".
> POV-Ray is, and has alwas been, a "rendering system", complete out of
> the box, with the ability to generate scenes through a scripting
> system (until POV 2.2), then a more complete SDL (from POV 3).
>
> With Mental Ray, you can't do that, you need a modeler with an interface
> to control it.
>
> Even with Renderman, you can't created decent scenes with a text file,
> it totally lacks any feature that would allow text-based modelling.

renderer, rendering system... what's the difference?  They accept an image
model as input and produce an image as output, that's all.

What's the difference if that image model is binary mesh and textures
format, a compiled C/C++ program or an ASCII textual SDL script model?
Each has their pros and cons... certainly the SDL appeals to programmers,
though not much to people used to visual modelling tools.  And its more
true today than at povray's heyday.

The point is that Renderman and Mental Ray don't come with their own VISUAL
modeller GUI.  Instead, they ship as plugins for the most common commercial
Visual Modellers, like Maya or 3DStudioMax... they too, despite backed by
heavy ammounts of money and developers, chose to do what they do best:
rendering, leaving visual modelling to others.  Why should povray with its
limited pool of developers do differently?

Post a reply to this message

>>   Let the gurus code these
> 
> The key, IMO, is : how many gurus will there be.

And how much free time they will have.

Post a reply to this message

> renderer, rendering system... what's the difference?  

A renderer comes at the end of the creative process, changing
tons of triangles and various directives into a good-looking
image.

A rendering system INCLUDES the creative process.

3DS Max PLUS Mental Ray is a rendering system.

POV-Ray is a rendering system on its own.

> They accept an image
> model as input and produce an image as output, that's all.
> 
> What's the difference if that image model is binary mesh and textures
> format, a compiled C/C++ program or an ASCII textual SDL script model?

The difference is that the SDL is a user-friendly method to design
a scene to render, with some special capabilities.

Some years ago, there was some hype about a new raytracer that
did caustics. I can't find its name anymore, please, if someone
remembers...  It was in fact a bunch of C++ libraries, and you
had to write a C++ program to setup a scene and get an image...
Of course, it attracted 1 or 2 geeks, allowed its programmer
to get a good job somewhere (maybe), and is now forgotten.

Attracting users with ease of use AND features is a requirement.

> Each has their pros and cons... certainly the SDL appeals to programmers,
> though not much to people used to visual modelling tools.  And its more
> true today than at povray's heyday.
> 
> The point is that Renderman and Mental Ray don't come with their own VISUAL
> modeller GUI.  Instead, they ship as plugins for the most common commercial
> Visual Modellers, like Maya or 3DStudioMax... they too, despite backed by
> heavy ammounts of money and developers, chose to do what they do best:
> rendering, leaving visual modelling to others.  Why should povray with its
> limited pool of developers do differently?

If POV-Ray becomes a simple renderer, just good enough to be the queue 
of Blender or whatever other stuff, it might as well be stopped. There
are plenty of good pure renderers out there, POV-Ray must be something
different.

Fabien.

Post a reply to this message

nemesis wrote:
> Fa3ien <fab### [at] yourshoesskynetbe> wrote:

>>> However, it has always been stressed that povray -- like Renderman, Mental
>>> Ray and other renderers -- deals exclusively with the rendering process.
>> That's a fundamental misconception.  POV-Ray ISN'T a "renderer".
>> POV-Ray is, and has alwas been, a "rendering system", complete out of
>> the box, with the ability to generate scenes through a scripting
>> system (until POV 2.2), then a more complete SDL (from POV 3).
>>
>> With Mental Ray, you can't do that, you need a modeler with an interface
>> to control it.
>>
>> Even with Renderman, you can't created decent scenes with a text file,
>> it totally lacks any feature that would allow text-based modelling.
> 
> renderer, rendering system... what's the difference?  They accept an image
> model as input and produce an image as output, that's all.
> 
> What's the difference if that image model is binary mesh and textures
> format, a compiled C/C++ program or an ASCII textual SDL script model?
> Each has their pros and cons... certainly the SDL appeals to programmers,
> though not much to people used to visual modelling tools.  
One of my applications for POV is in a script I wrote to generate 
transmembrane proteins. Input is a database with for a number of genes 
the transmembrane segments and a user specified number of aminoacids you 
want to highlight (possibly a database of all known mutations). Typical 
use is to create a piece of celmembrane with all known mutations sorted 
by disease in different colors (see attached). ATM I do the processing 
of the databases in Matlab, just because that is most easy for me, and I 
need a little more support for reading files, but I would love to 
convert this into a POV only tool. As you can imagine the fact that POV 
is a scripting language is essential for this application. Also because 
I only generate a gene specific include file, but there are a number of 
switches in the .pov file for various type of output (color, b/w, other 
shapes apart from spheres etc).
I know of no other program that can do this for me in this way. I know 
that I could generate meshes in stead of spheres, curved lines etc, but 
I would not have though about starting such a project. With POV it was 
incredibly easy to conceive and to execute this idea. A couple of people 
have admired my use of shade and reflections in this and other images. I 
am honest enough to tell them that it would be more complicated to 
remove those, but not everybody believes that.
Think about that, there is a whole field of applications that are not 
possible (or totally impractical) without POV. That goes much further 
than personal preferences for command line or GUI interfaces.


> And its more
> true today than at povray's heyday.
> 
> The point is that Renderman and Mental Ray don't come with their own VISUAL
> modeller GUI.  Instead, they ship as plugins for the most common commercial
> Visual Modellers, like Maya or 3DStudioMax... they too, despite backed by
> heavy ammounts of money and developers, chose to do what they do best:
> rendering, leaving visual modelling to others.  Why should povray with its
> limited pool of developers do differently?
> 
As shown above, because it has its own niche.

Post a reply to this message

Attachments:
Download 'gene.png' (52 KB)

Preview of image 'gene.png'

Fa3ien <fab### [at] yourshoesskynetbe> wrote:
> > renderer, rendering system... what's the difference?
>
> A renderer comes at the end of the creative process, changing
> tons of triangles and various directives into a good-looking
> image.
>
> A rendering system INCLUDES the creative process.
>
> 3DS Max PLUS Mental Ray is a rendering system.
>
> POV-Ray is a rendering system on its own.

Yes, you can take ages to produce a very high quality pure-SDL scene in
povray using nothing but CSG or isosurfaces and no characters.  Or, like
Gilles stressed, you can do it much more fluidly, easily and faster with an
interactive GUI-based modeller with real-time visualization.

In that sense, the SDL is a very poor modelling solution and it makes povray
a very limited "rendering system".  But it is a pretty good renderer anyway.

> Some years ago, there was some hype about a new raytracer that
> did caustics. I can't find its name anymore, please, if someone
> remembers...  It was in fact a bunch of C++ libraries, and you
> had to write a C++ program to setup a scene and get an image...
> Of course, it attracted 1 or 2 geeks, allowed its programmer
> to get a good job somewhere (maybe), and is now forgotten.

next you'll tell me the pure SDL scene writers out there are not geeks...
like the chemical geek rendering molecules out of matlab and SDL...

> If POV-Ray becomes a simple renderer, just good enough to be the queue
> of Blender or whatever other stuff, it might as well be stopped. There
> are plenty of good pure renderers out there, POV-Ray must be something
> different.

It is something different:  a renderer with the most simple, flexible and
accessible input format.

why stop something which works well?

Post a reply to this message