Re-categorizing to a feature request.  Octave no longer errors on the fifth argument, but it also isn't processed in any way.

Avinoam's suggestion looks useful as long as the options struct hasn't been implemented. Just changing L.70 into

narginchk (1, 5);
if (nargin > 4)
warning ("inputdlg: 5th 'options' argument ignored");
endif

will do.

While trying Matlab's example 4 on the page Mike indicated in comment #2, I hit a Matlab incompatibility:

>> answer = inputdlg (prompt, name, [1 40], defaultans, options);
error: inputdlg: ROWSCOLS vector does not match size of PROMPT
error: called from
    inputdlg at line 124 column 7

Apparently Octave does not recognize a singleton dimension. Should be an easy fix that may need generalizing into the matrix form as well but I have no time now.

Let's tackle one bug at a time. Retitling to request addition of the fifth input argument to inputdlg.

Description of the options argument is here:

http://www.mathworks.com/help/matlab/ref/inputdlg.html

The mathworks iFDTD toolbox can be downloaded from this page ...

http://www.problemsinelectrodynamics.com/tools/interactive-fdtd-toolbox

I tried to run it with Octave and got an error msg,

inputdlg command not implemented.

I anticipate that there will be more user I/O commands not implemented.

This is a fantastic toolbox, and I think interactive programs are the wave of the future, so these would be great features to add to Octave.