Drawing planes in 3D with TikZ

Question

In drawing sets of planes like the figure below

one frequently see solutions that involve drawing each visual piece of each plane in the order from the back to the front, like the code included here, and for another example see here (intersecting planes).

Question: Is it possible to draw the planes using 3D-coordinates and choose a viewpoint within TikZ, without having to calculate the view before hand?

 \documentclass{standalone}
 \usepackage{tikz}
 \usetikzlibrary{positioning,calc}
 \usetikzlibrary{intersections}

 \begin{document}
 \pagecolor{blue!30}
 \pagestyle{empty}

 \begin{tikzpicture}[scale=1.6]

 \definecolor{bg}{RGB}{246,202,203}

 \coordinate (A) at (0.95,3.41);
 \coordinate (B) at (1.95,0.23);
 \coordinate (C) at (3.95,1.23);
 \coordinate (D) at (2.95,4.41);

 \coordinate (E) at (1.90,3.30);
 \coordinate (F) at (0.25,0.45);
 \coordinate (G) at (2.25,1.45);
 \coordinate (H) at (3.90,4.30);

 \coordinate (I) at (-0.2,1.80);
 \coordinate (J) at (2.78,1.00);
 \coordinate (K) at (4.78,2.00);
 \coordinate (L) at (1.80,2.80);

  \path[name path=AB] (A) -- (B);
  \path[name path=CD] (C) -- (D);
  \path[name path=EF] (E) -- (F);
  \path[name path=IJ] (I) -- (J);
  \path[name path=KL] (K) -- (L);
  \path[name path=HG] (H) -- (G);
  \path[name path=IL] (I) -- (L);
    \path [name intersections={of=AB and EF,by=M}];
    \path [name intersections={of=EF and IJ,by=N}];
    \path [name intersections={of=AB and IJ,by=O}];
    \path [name intersections={of=AB and IL,by=P}];
    \path [name intersections={of=CD and KL,by=Q}];
    \path [name intersections={of=CD and HG,by=R}];
    \path [name intersections={of=KL and HG,by=S}];
  \path[name path=NS] (N) -- (S);
  \path[name path=FG] (F) -- (G);
    \path [name intersections={of=NS and AB,by=T}];
    \path [name intersections={of=FG and AB,by=U}];

 \draw[thick, color=white, fill=bg] (A) -- (B) -- (C) -- (D) -- cycle;
 %\draw[thick, color=white, fill=bg] (E) -- (F) -- (G) -- (H) -- cycle;
 %\draw[thick, color=white, fill=bg] (I) -- (J) -- (K) -- (L) -- cycle;

 \draw[thick, color=white, fill=gray!80] (P) -- (O) -- (I) -- cycle;
 \draw[thick, color=white, fill=gray!80] (O) -- (J) -- (K) -- (Q) -- cycle;
 \draw[thick, color=white, fill=gray!40] (H) -- (E) -- (M) -- (R) -- cycle;
 \draw[thick, color=white, fill=gray!40] (M) -- (N) -- (T) -- cycle;
 \draw[thick, color=white, fill=gray!40] (N) -- (F) -- (U) -- (O) -- cycle;

 \end{tikzpicture}
 \end{document}

Answer 1

This is more a fun answer with the message that it can be done but requires some patience. Moreover, I allow the latitude angle theta only to be in the range above 90 degrees. In this case, there are only two cases that have to be distinguished, i.e. this is not the general case. There are four cases, which are distinguished by two binary numbers

• the sign of the projection of the 3D x-axis on the x-direction of the screen, \xproj.
• the sign of cos(theta), \zproj (in the conventions of tikz-3dplot theta is between 0 and 180, and the equator is at theta=90). This sign tells one if one is in the southern or northern hemisphere.

That is, we will distinguish the cases \xproj and \zproj positive or negative. Depending on these signs, the order in which the planes are drawn changes. In order to retain a bit of clarity this answer comes with a macro \DrawSinglePlane{<plane number>, such that a change of the drawing order just corresponds to a permutation of the list of plane numbers.

\documentclass[tikz,border=3.14mm]{standalone}
\usepackage{tikz}
\usepackage{tikz-3dplot}
\usetikzlibrary{calc}
\newcommand{\DrawPlane}[3][]{\draw[#1] 
(-1*\PlaneScale,{\PlaneScale*cos(#2)},{\PlaneScale*sin(#2)})
 -- ++  (2*\PlaneScale,0,0)
 -- ++ (0,{sqrt(3)*\PlaneScale*cos(#3)},{sqrt(3)*\PlaneScale*sin(#3)})
 -- ++  (-2*\PlaneScale,0,0) -- cycle;}
\newcommand{\DrawSinglePlane}[2][]{\ifcase#2
\or
\DrawPlane[fill=blue,#1]{210}{240} %left bottom 
\or
\DrawPlane[fill=red,#1]{-30}{-60} % right bottom
\or
\DrawPlane[fill=purple,#1]{210}{180} % bottom left
\or
\DrawPlane[fill=purple,#1]{210}{0} % bottom middle
\or
\DrawPlane[fill=purple,#1]{-30}{0} % bottom right
\or
\DrawPlane[fill=blue,#1]{90}{240} % left top
\or
\DrawPlane[fill=red,#1]{90}{-60} % right middle
\or
\DrawPlane[fill=red,#1]{90}{120} % right top
\or
\DrawPlane[fill=blue,#1]{90}{60} % left top
\fi
} 
\begin{document}
\foreach \X in {0,5,...,355}
{\tdplotsetmaincoords{90+40*sin(\X)}{\X} % the first argument cannot be larger than 90
\pgfmathsetmacro{\PlaneScale}{1}
\begin{tikzpicture}
\path[use as bounding box] (-4*\PlaneScale,-4*\PlaneScale) rectangle (4*\PlaneScale,4*\PlaneScale);
\begin{scope}[tdplot_main_coords]
% \draw[thick,->] (0,0,0) -- (2,0,0) node[anchor=north east]{$x$};
% \draw[thick,->] (0,0,0) -- (0,2,0) node[anchor=north west]{$y$};
% \draw[thick,->] (0,0,0) -- (0,0,1.5) node[anchor=south]{$z$};
\path let \p1=(1,0,0)  in 
\pgfextra{\pgfmathtruncatemacro{\xproj}{sign(\x1)}\xdef\xproj{\xproj}};
\pgfmathtruncatemacro{\zproj}{sign(cos(\tdplotmaintheta))}
\xdef\zproj{\zproj}
% \node[anchor=north west] at (current bounding box.north west)
% {\tdplotmaintheta,\tdplotmainphi,\xproj,\zproj};
\ifnum\zproj=1
  \ifnum\xproj=1
   \foreach \X in {2,1,5,4,3,7,6,9,8}
    {\DrawSinglePlane{\X}}
  \else
   \foreach \X in {1,...,9}
    {\DrawSinglePlane{\X}}
  \fi  
\else
  \ifnum\xproj=1
   \foreach \X in {9,8,7,6,3,4,5,2,1}
    {\DrawSinglePlane{\X}}
  \else
   \foreach \X in {8,9,6,7,3,5,4,1,2}
    {\DrawSinglePlane{\X}}
  \fi  
\fi  
\end{scope}
\end{tikzpicture}}
\end{document}

And for \tdplotsetmaincoords{90+40*cos(\X)}{\X}

A potentially important comment concerns pgfplots. In principle one can use patchplots to do the same. pfplots comes, as discussed in the comments, with some means to do the ordering.

UDPATE: Covers now the full range.

IMPORTANT NOTE: No duck or marmot were harmed in these animations. ;-)

Answer 2

Is Tikz a strong requirement? I've found Asymptote (included with TeXLive) to be an excellent tool for such tasks. Below is a very lightly edited example from the Asymptote Gallery.

You can change the viewpoint simply by changing the line that begins with currentprojection.

size(6cm,0);
import bsp;
real u=2.5;
real v=1;
currentprojection = oblique;
path3 y=plane((2u,0,0),(0,2v,0),(-u,-v,0));
path3 a=rotate(45,X)*y;
path3 l=rotate(-45,Z)*rotate(45,Y)*rotate(45,Z)*y;
path3 g=rotate(45,X)*rotate(45,Y)*rotate(45,Z)*y;
face[] faces;
filldraw(faces.push(a),project(a),gray);
filldraw(faces.push(l),project(l),blue);
filldraw(faces.push(g),project(g),pink);
add(faces);

This produces the following figure:

While the one line change to currentprojection = perspective(5,2,3); produces this figure:

An excellent Asymptote Tutorial has been written by Charles Staats, a PhD student at the University of Chicago.