POV-Ray : Newsgroups : moray.win : Impending Moray Source Release Server Time
29 Mar 2024 10:17:57 EDT (-0400)
  Impending Moray Source Release (Message 5 to 14 of 18)  
<<< Previous 4 Messages Goto Latest 10 Messages Next 4 Messages >>>
From: And
Subject: Re: Impending Moray Source Release
Date: 24 Nov 2022 07:45:00
Message: <web.637f6703b7839016e02c756aa81652d@news.povray.org>
As Bald Eagle said, meshifying primitives(and CSG combine) is a sugar feature in
my imagine in the past years when I use POV-Ray. But It can be regarded as an
extra- expect.


------------------------------------------------

My comments yesterday is like this:
If it can have a storage that collect "simple variable": like number, vector,
color, string, (not textures and objects) ..., then link to property of objects/
textures in the CSG-tree.

Just an imagine style of mine...


Post a reply to this message


Attachments:
Download 'variable table.png' (39 KB)

Preview of image 'variable table.png'
variable table.png


 

From: Bald Eagle
Subject: Re: Impending Moray Source Release
Date: 24 Nov 2022 08:35:00
Message: <web.637f7249b7839011f9dae3025979125@news.povray.org>
Thomas de Groot <tho### [at] degrootorg> wrote:

> Just for now, a very /basic/ wish for the new Moray, one that every
> Moray user will approve I am sure:
>
>
> * Make the GUI /left-handed/ instead of /right-handed/. *

I approve in the sense that I disapprove of locking the users into exactly the
opposite basis vectors.

If we would like to make the New Moray something useful and enduring, then I
would say that if all of the operations proceed from first principles, assuming
virtually nothing, then all of the operations can be generalized and it will be
as flexible and trouble-free as we can manage, possibly allowing things we might
not imagine when building it.

So I say, proceed from a configuration file or control panel, where the x, y,
and z vectors can be specified like in a matrix {}.  Then you can do whatever
you want.   Else, I can imagine the very first thing someone somewhere will
complain about is that they wanted to use Moray for something and - dammit - why
is it only LEFT handed???!  Can the developers make it so I can use it right
handed....?

Then we can see what Francois Le Coat does with it.  :D


Post a reply to this message

From: And
Subject: Re: Impending Moray Source Release
Date: 24 Nov 2022 09:00:00
Message: <web.637f77ceb7839016e02c756aa81652d@news.povray.org>
Then POV-Ray's function can be another table, I don't know how convenient such
type of program be implement, or very difficult? If check correct grammar is too
much demand...


Post a reply to this message


Attachments:
Download 'function table.png' (61 KB)

Preview of image 'function table.png'
function table.png


 

From: Bald Eagle
Subject: Re: Impending Moray Source Release
Date: 24 Nov 2022 09:00:00
Message: <web.637f784db7839011f9dae3025979125@news.povray.org>
"Convinced" <der### [at] gmailcom> wrote:

> As for features, I recall Moray had some parametric capabilities, I could build
> assemblies with it that were dimensionally correct. This capability would be a
> must as would 3D printing.

I would concur, as specifying exact parameters is crucial for precision work,
and using POV-Ray to model or design real and/or important things that are not
simply arbitrary and artistic in purpose.

With regard to 3D printing, I was flirting with the idea of working for a local
rapid prototyping company doing CNC, and for me at least, designing things in
POV-Ray or OpenSCAD seemed only one small step removed from writing G-code for
CNC machining.  A module to translate G-code commands to difference {} or even
animation would be a possibility to be aware of and leave open.
I say that, not to be ridiculous, but because one never knows who will become
interested in Moray, and by extension, POV-Ray - that could lead to more
widespread use, along with technical and financial support.

> The UI was OK and usable and intuitive as such. Ease of use is essential without
> gimmicks, all the benifits of CAD sftware is important.
> I've ressurected a few old programs and managed to run them in WIN 10 64bit and
> I am amazed at how marvellous they were, TrueSpace,Hexagon, Carrara to name a
> few.
> Please whatever you decide to do don't lose the magic of these programs.

I would love to see some examples / screenshots of what those programs do.

I have no idea how Moray was put together, but I think that the Linux users here
would probably agree that emulating the "Unix mindset" - writing a standalone
tool that does one thing, and one thing well - and then daisy-chaining those
tools together might be the best approach.

I'm reminded of analyses of what NASA did with the space shuttle and other
projects, where they transitioned from designing modules with individual
specifications, developed by separate teams assigned to those parts, and those
were then interfaced with each other to make a working whole.   Then they
switched to that whole "everything is integrated as a whole" and there was no
localized control or separability.  Every time something at one end got changed
- it affected what some other team was doing somewhere else. (Boy, does this
bring to mind Bezier splines and NURBS)

So I think that just as important as what "we" _should_ do, is what we _should
not_ do, if at all possible.   I recall clipka going through sections of code
and not having any idea what the person who wrote that part was thinking.  (uv
mapping modes)
Separable tasks were integrated into the parser, making them difficult to factor
out, if not impossible without a full from-the-ground-up rewrite.

And lastly, just have a text file associated with any modules or files or
functions or macros for the developer to just scribble in.   "Here's what I'm
thinking and WHY - here's where these numbers / formulas / methods come from,
here are the references and excerpts, WIP code snippets, etc....


Post a reply to this message


Attachments:
Download 'prototekpart.png' (174 KB)

Preview of image 'prototekpart.png'
prototekpart.png


 

From: And
Subject: Re: Impending Moray Source Release
Date: 24 Nov 2022 09:05:00
Message: <web.637f797fb7839016e02c756aa81652d@news.povray.org>
"Bald Eagle" <cre### [at] netscapenet> wrote:
> Thomas de Groot <tho### [at] degrootorg> wrote:
>
> > Just for now, a very /basic/ wish for the new Moray, one that every
> > Moray user will approve I am sure:
> >
> >
> > * Make the GUI /left-handed/ instead of /right-handed/. *
>
> I approve in the sense that I disapprove of locking the users into exactly the
> opposite basis vectors.
>
> If we would like to make the New Moray something useful and enduring, then I
> would say that if all of the operations proceed from first principles, assuming
> virtually nothing, then all of the operations can be generalized and it will be
> as flexible and trouble-free as we can manage, possibly allowing things we might
> not imagine when building it.
>
> So I say, proceed from a configuration file or control panel, where the x, y,
> and z vectors can be specified like in a matrix {}.  Then you can do whatever
> you want.   Else, I can imagine the very first thing someone somewhere will
> complain about is that they wanted to use Moray for something and - dammit - why
> is it only LEFT handed???!  Can the developers make it so I can use it right
> handed....?
>
> Then we can see what Francois Le Coat does with it.  :D

Reserve left-hand/ right-hand is better indeed.

I cannot figure out what is "where the x, y, and z vectors can be specified like
in a matrix {}"


Post a reply to this message

From: Bald Eagle
Subject: Re: Impending Moray Source Release
Date: 24 Nov 2022 09:30:00
Message: <web.637f7f27b7839011f9dae3025979125@news.povray.org>
"And" <49341109@ntnu.edu.tw> wrote:

> I cannot figure out what is "where the x, y, and z vectors can be specified like
> in a matrix {}"

When we use POV-Ray, we assume that the basis vectors for the Euclidian space we
are modeling in, are defined as:
x = <1, 0, 0>
y = <0, 1, 0>
z = <0, 0, 1>

TdG points out that a "fix" or workaround for scenes made in Moray, which is
right-handed, is to scale the camera "scale <-1, 0, 0>" in order to "flip" the
whole scene and make it look correct.

But fundamentally, what you are doing is changing the x basis vector (relatively
speaking), which affects every point in the 3D space that you are working in,
and therefore every object built using those coordinates.

Making those basis vectors user-definable at their core would add a maximum
flexibility.
In fact, if one really wanted to go nuts, rather than just having scalar values
in the matrix, one could define some sort of function to warp the entire
coordinate space.   I know that sounds crazy, and most people will ask, "WHY
would you want to do that???!", but watch Grant's excellent videos, and you
might get some ideas.


http://www.f-lohmueller.de/pov_tut/trans/trans_400e.htm

Also check out Grant Sanderson's (3Blue1Brown) excellent series on linear
algebra.
https://www.youtube.com/watch?v=k7RM-ot2NWY


Post a reply to this message

From: And
Subject: Re: Impending Moray Source Release
Date: 24 Nov 2022 10:10:00
Message: <web.637f87e8b7839016e02c756aa81652d@news.povray.org>
"Bald Eagle" <cre### [at] netscapenet> wrote:
>
> When we use POV-Ray, we assume that the basis vectors for the Euclidian space we
> are modeling in, are defined as:
> x = <1, 0, 0>
> y = <0, 1, 0>
> z = <0, 0, 1>
>
> TdG points out that a "fix" or workaround for scenes made in Moray, which is
> right-handed, is to scale the camera "scale <-1, 0, 0>" in order to "flip" the
> whole scene and make it look correct.
>
> But fundamentally, what you are doing is changing the x basis vector (relatively
> speaking), which affects every point in the 3D space that you are working in,
> and therefore every object built using those coordinates.
>
> Making those basis vectors user-definable at their core would add a maximum
> flexibility.
> In fact, if one really wanted to go nuts, rather than just having scalar values
> in the matrix, one could define some sort of function to warp the entire
> coordinate space.   I know that sounds crazy, and most people will ask, "WHY
> would you want to do that???!", but watch Grant's excellent videos, and you
> might get some ideas.
>
>
> http://www.f-lohmueller.de/pov_tut/trans/trans_400e.htm
>
> Also check out Grant Sanderson's (3Blue1Brown) excellent series on linear
> algebra.
> https://www.youtube.com/watch?v=k7RM-ot2NWY


Oh, I know what thing do you said now, thank you.
What you said is about the transform associate with camera before cast to
screen...

I have basic capability with vector and matrix. But fail to understand
eigenvalue/ eigenvector / diagonal matrix

I ever derive/coding method of axis rotate by myself.


Post a reply to this message


Attachments:
Download 'vector3.png' (65 KB)

Preview of image 'vector3.png'
vector3.png


 

From: Bald Eagle
Subject: Re: Impending Moray Source Release
Date: 24 Nov 2022 11:25:00
Message: <web.637f9a09b7839011f9dae3025979125@news.povray.org>
"And" <49341109@ntnu.edu.tw> wrote:

> Oh, I know what thing do you said now, thank you.
> What you said is about the transform associate with camera before cast to
> screen...

Yes, but it can affect more than the camera.  You can use a plain camera and
apply those types of transforms to objects or even patterns.   This is sort of
how some pigment patterns are generated, and it's also very much like
uv-mapping, parametrics, Bezier patches, etc.

One of our vectors might not be "x", but "East" as we travel around a globe.

You have your normal x, y, and z as a starting point - the standard orthogonal
unit vectors - but then traveling in the new directions does not result in
travel along the cardinal axes, or even straight lines, because you're following
the path of a different vector or parametric function.

> I have basic capability with vector and matrix. But fail to understand
> eigenvalue/ eigenvector / diagonal matrix

https://www.youtube.com/watch?v=PFDu9oVAE-g
https://www.youtube.com/watch?v=e50Bj7jn9IQ

> I ever derive/coding method of axis rotate by myself.

Yes, that's a good thing to do - the more you understand the basics, the less
mysterious the more complicated stuff seems to be.


Post a reply to this message

From: Thomas de Groot
Subject: Re: Impending Moray Source Release
Date: 25 Nov 2022 02:48:13
Message: <6380733d$1@news.povray.org>
Op 24/11/2022 om 14:31 schreef Bald Eagle:
> Thomas de Groot <tho### [at] degrootorg> wrote:
> 
>> * Make the GUI /left-handed/ instead of /right-handed/. *
> 
> [snip]
> 
> So I say, proceed from a configuration file or control panel, where the x, y,
> and z vectors can be specified like in a matrix {}.  Then you can do whatever
> you want.   Else, I can imagine the very first thing someone somewhere will
> complain about is that they wanted to use Moray for something and - dammit - why
> is it only LEFT handed???!  Can the developers make it so I can use it right
> handed....?
> 
> Then we can see what Francois Le Coat does with it.  :D
> 

Much better than my suggestion. I fully agree with this principle!

-- 
Thomas


Post a reply to this message

From: And
Subject: Re: Impending Moray Source Release
Date: 6 Dec 2022 09:30:00
Message: <web.638f517db7839013d32f071aa81652d@news.povray.org>
"Bald Eagle" <cre### [at] netscapenet> wrote:
> "And" <49341109@ntnu.edu.tw> wrote:
>
> > I have basic capability with vector and matrix. But fail to understand
> > eigenvalue/ eigenvector / diagonal matrix
>
> https://www.youtube.com/watch?v=PFDu9oVAE-g
> https://www.youtube.com/watch?v=e50Bj7jn9IQ
>
> > I ever derive/coding method of axis rotate by myself.
>
> Yes, that's a good thing to do - the more you understand the basics, the less
> mysterious the more complicated stuff seems to be.

Thanks the youtube link. I will think them one day.


Post a reply to this message

<<< Previous 4 Messages Goto Latest 10 Messages Next 4 Messages >>>

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.