1D matrix addition weirdness

>> ones(5,1)+ones(1,5)

ans =

     2     2     2     2     2
     2     2     2     2     2
     2     2     2     2     2
     2     2     2     2     2
     2     2     2     2     2
% what?????     
     


% but
>> ones(2,5)+ones(5,2)
Matrix dimensions must agree.
% as expected

% 2014 ver produce error in both cases

Why is this unexpected vecor addition behaviour in recent MATLABs (2017,2018)??????

Answer 1

[madhan ravi]

It's not wierd see https://www.mathworks.com/help/matlab/matlab_prog/compatible-array-sizes-for-basic-operations.html . Clearly stated here.

[Alex Pedcenko]

Thanks! (its still weird from mathematics point of view [wrong actually])

[madhan ravi]

Anytime :)

[Stephen23]

"its still weird from mathematics point of view [wrong actually])"

Note that implicit expansion is really just an extension of scalar expansion, which has been part of MATLAB for a very long time: is scalar expansion weird/wrong as well?? A strict language would require all operations to be performed only on exactly compatible array sizes (i.e. the same size for array operations), but not only would this be inconvenient, using repmat all the time would likely be much less efficient than scalar expansion and implicit expansion.

It would certainly be nice if it was possible to select expansion or no expansion.

https://blogs.mathworks.com/loren/2016/11/10/more_thoughts_about_implicit_expansion/

https://se.mathworks.com/matlabcentral/answers/318495-deactivation-switch-for-implicit-expansion

https://www.mathworks.com/matlabcentral/answers/383533-how-do-i-avoid-getting-fooled-by-implicit-expansion

https://www.mathworks.com/matlabcentral/answers/1680-what-parts-of-matlab-would-you-not-miss#answer_305200

https://undocumentedmatlab.com/blog/afterthoughts-on-implicit-expansion