2

I need some serious help getting this commutative diagram to look presentable. I've been messing around with it for literally weeks and this is the best I've got.

cd

The only suppression that I could accept is swapping $H^{n-p}(F)$ for a suppression like $\mathcal{F}$. The code is:

\[\begin{tikzpicture}[scale=1.5]
\node (sa) at (0,5) {$\ldots$};
\node (sb) at (4,5) {$\ldots$};

\node (a1) at (0,4.3) {$\mathop\bigoplus\limits_{p = 0}^n [H^p(U\cup V) \otimes H^{n-p}(F)]$};
\node (a2-1) at (-.12,3.3) {$\mathop\bigoplus\limits_{p = 0}^n [(H^p(U)\otimes H^{n-p}(F)$};
\node (a2-2) at (0,3) {$\oplus$};
\node (a2-3) at (.16,2.7) {$(H^p(V)\otimes H^{n-p}(F))]$};
\node (a3) at (0,1.7) {$\mathop\bigoplus\limits_{p = 0}^n [H^p(U\cap V)\otimes H^{n-p}(F)]$};
\node (a4) at (0,.7) {$\mathop\bigoplus\limits_{p = 0}^n [H^{p+1}(U\cup V) \otimes H^{n-p}(F)]$};

\node (b1) at (4,4.3) {$H^n(\pi^{-1}(U\cup V))$};
\node (b2-1) at (4,3.3) {$H^n(U\times F)$};
\node (b2-2) at (4,3) {$\oplus$};
\node (b2-3) at (4,2.7) {$H^n(V\times F)$};
\node (b3) at (4,1.7) {$H^n((U\cap V\times F))$};
\node (b4) at (4,.7) {$H^{n+1}(\pi^{-1}(U\cup V))$};

\node (ea) at (0,0) {$\ldots$};
\node (eb) at (4,0) {$\ldots$};

\path[->,font=\scriptsize,>=latex]
([xshift=5pt]sa.east) edge node[above]{$\Psi$} ([xshift=-5pt]sb.west)
(a1) edge node[above]{$\Psi$} (b1)
([xshift= 36pt]a2-2.east) edge node[above]{$\Psi$} ([xshift= -20pt]b2-2.west)
(a3) edge node[above]{$\Psi$} (b3)
([xshift=5pt]ea.east) edge node[above]{$\Psi$} ([xshift=-5pt]eb.west)
(a4) edge node[above]{$\Psi$} (b4)

(sa) edge ([yshift= -3.6pt]a1.north)
([yshift=5pt]a1.south) edge node[left]{$\delta$} ([xshift= 3.43pt,yshift=-4pt]a2-1.north)
([xshift= -4.55pt]a2-3.south) edge node[left]{$\gamma$} ([yshift=-3.6pt]a3.north)
([yshift= 5pt]a3.south) edge node[left]{$d^*$} ([yshift= -3.6pt]a4.north)
([yshift= 5pt]a4.south) edge (ea)

(sb) edge (b1)
(b1) edge node[right]{$\delta$} (b2-1)
(b2-3) edge node[right]{$\gamma$} (b3)
(b3) edge node[right]{$d^*$} (b4)
(b4) edge (eb);
\end{tikzpicture}\]
1
  • My only advice is: turn the page 90 degrees and write it vertically
    – Ottavio
    Oct 29, 2019 at 20:16

1 Answer 1

2

Here's a possibility with tikz-cd; the Psi labels would better be aligned, though; there should be a way!

\documentclass{article}
\usepackage{amsmath,mathtools}
\usepackage{tikz-cd}

\begin{document}

\[
\newcommand{\sbop}{\mathop{\textstyle\bigoplus}\limits}
\begin{tikzcd}[column sep=6em,row sep=large]
\cdots \arrow[r,"\Psi"] \arrow[d] & \cdots \arrow[d]
\\
\smash{\sbop_{p = 0}^n}\, [H^p(U\cup V) \otimes H^{n-p}(F)]
  \arrow[r,"\Psi"] \arrow[d,swap,"\delta"]
  & H^n(\pi^{-1}(U\cup V)) \arrow[d,"\delta"]
\\
\begin{multlined}
  \smash[t]{\sbop_{p = 0}^n}\, [(H^p(U)\otimes H^{n-p}(F) \\
  \oplus (H^p(V)\otimes H^{n-p}(F))]
\end{multlined}
  \arrow[r,"\Psi"] \arrow[d,swap,"\gamma"]
  & H^n(U\times F) \oplus H^n(V\times F) \arrow[d,"\gamma"]
\\
\smash{\sbop_{p = 0}^n}\, [H^p(U\cap V)\otimes H^{n-p}(F)]
  \arrow[r,"\Psi"] \arrow[d,swap,"d^*"]
  & H^n((U\cap V\times F)) \arrow[d,"d^*"]
\\
\smash{\sbop_{p = 0}^n}\, [H^{p+1}(U\cup V) \otimes H^{n-p}(F)]
  \arrow[r,"\Psi"] \arrow[d]
  & H^{n+1}(\pi^{-1}(U\cup V)) \arrow[d]
\\
\cdots \arrow[r,"\Psi"] & \cdots
\end{tikzcd}
\]

\end{document}

enter image description here

3
  • I appreciate the effort. I think this it is simply fate that this diagram be forever unattractive.
    – aaron
    Feb 24, 2016 at 8:52
  • @aaron Is your question answered by this post? If not, please edit your question. If yes, please accept this answer here by clicking the check mark to its left and I will delete my comment here. Thank you.
    – LaRiFaRi
    May 3, 2016 at 11:38
  • @LaRiFaRi It doesn't answer my question. But, my paper is already in printing so it doesn't matter anymore. I would delete the question myself but I'm unable to.
    – aaron
    May 4, 2016 at 2:09

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .