1

I can't seem to get consistent behavior with the proof environment and QED symbol in the latest SIAM document class, in particular with the eqnarray and equation environments. My MWE requires some interactivity, so please read it. My question is only partially answered by this post.

\documentclass[]{siamart1116}
\begin{document}
I would really love a QED symbol to show up at the end, without having to abandon the proof environment or do something annoying by hand every time.
\begin{proof}
    \begin{eqnarray*}
        A
        & = & B \\
        & = & C \\
        & = & D. \mbox{ end of proof!}
    \end{eqnarray*}
\end{proof}
Some words and thoughts on that earth-shattering result.  Oops, but as a clueless reader who skipped the proof, I have no idea where the post-proof discussion begins! \\
The equation environment seems ok, but sometimes the QED symbol shows, and sometimes it doesn't.  Right now it seems to show.
\begin{proof}
    Since $A = B = C = D$, then it's probably also true that
    \begin{equation}
    A = E.
    \end{equation}
\end{proof}
Now comment out the middle proof, and just run with this one.  We're back to weirdness.
\begin{proof}
Hell, I bet
    \begin{equation}
    E = F = G = H...
    \end{equation}
    i.e., that the whole flippin' alphabet is equal.
\end{proof}
\end{document}

1 Answers1

2

The main problem is using eqnarray, which is not suitable for real math typesetting, see eqnarray vs align.

On the other hand, the class is based on ntheorem which is buggy when its thmmarks option for automatically adding the end-of-proof symbol is used. For instance, using align* instead of eqnarray* in the first alignment and removing the middle proof yields

enter image description here

However the wrong end-of-proof symbol disappears upon a new compilation:

enter image description here

Removing the middle proof initially confuses an internal counter, but a new compilation brings things in sync.

\documentclass{siamart1116}

\begin{document}

I would really love a QED symbol to show up at the end, 
without having to abandon the proof environment or do 
something annoying by hand every time.
\begin{proof}
\begin{align*}
  A
  &= B \\
  &= C \\
  &= D. 
\end{align*}
\end{proof}
Some words and thoughts on that earth-shattering result.  
Oops, but as a clueless reader who skipped the proof, 
I have no idea where the post-proof discussion begins!
The equation environment seems ok, but sometimes the QED 
symbol shows, and sometimes it doesn't.  Right now it seems to show.
\begin{proof}
Since $A = B = C = D$, then it's probably also true that
\begin{equation}
A = E.
\end{equation}
\end{proof}
Now comment out the middle proof, and just run with this one.  
We're back to weirdness.
\begin{proof}
Hell, I bet
\begin{equation}
E = F = G = H...
\end{equation}
i.e., that the whole flippin' alphabet is equal.
\end{proof}

\end{document}

enter image description here

Unfortunately, ntheorem doesn't set up a warning mechanism when end-of-proof symbols are misplaced and a new LaTeX run is needed. Be sure to run LaTeX once more when your document is in final form.

egreg
  • 1,121,712