Matlab 2016b Matrix Dimensions for addition and subtraction?

In previous version of Matlab if I type:

>> a = [1 2; 3 4]; >> b = [1;3];

and then attempt to execute:

>> a - b

I get an error (as I should) because the matrix dimensions don't agree.

However, if I attempt this in 2016b I get:

>> a-b

ans =

     0     1
     0     1

Is this a new feature? Is this seen as an improvement? If so, can I turn this off? I feel this is dangerous.

Thanks,

-mcd

Answer 1

[James Tursa]

Yes, this is a new feature and is intentional. See this link under Implicit Expansion:

http://www.mathworks.com/help/matlab/release-notes.html?rntext=&startrelease=R2016a&endrelease=R2016b&groupby=release&sortby=descending&searchHighlight=

I don't see anything in the release notes that indicates you can shut this off. It is basically bsxfun that has now been implemented directly into the language.

[Michael]

Thanks for the response. I can see how this could be convenient in some situations, but it can lead to very weird results that could be hard to debug.