Compatibility
Despite best efforts by Mathworks to make matlab useful, some functionalities still only exist in octave. This page is for tricks that will help in writing code that will run in both octave and matlab.
Why?
Consider the following case. You have found some useful code that is currently matlab specific and want to port it into octave. You may need, or want, to make changes, either due to missing functions in octave or to use simpler/faster/etc octave specific functions. If you make these changes while keeping it matlab compatibility you can send the modified code, now both matlab and octave compatible, upstream. This will make contributions between the octave and matlab community easier. It will be easier to apply changes/bug fixes that both you and the upstream developer makes.
Are we running octave?
The most commonly used trick is to create a subfunction that will check whether we are running octave, and use it as condition to run selected block of code that are incompatible. Due to the persistent variable it can be called repeatedly without a heavy performance hit.
Code: use and creation of is_octave() 
function foo
%% fancy code that works in both
if (is_octave)
%% use octave super_powers
else
%% do it matlab way
end
%% fancy code that works in both
end
%% subfunction that checks if we are in octave
function r = is_octave ()
persistent x;
if (isempty (x))
x = exist ('OCTAVE_VERSION', 'builtin');
end
r = x;
end

Calling an internal octave function (even though you shouldn't)
Octave internal functions start with underscores (ie __delaunayn__ ). Normally, user programs shouldn't call these, but sometimes, for testing, you may want to do that. The problem is that matlab can't accept function names with underscores. The trick is to use feval
Code: code with underscores 
if isoctave ()
T = feval('__delaunayn__', opts)
else
% matlab code here
end
% Code to check the results are compatible
