How are editors supposed to know the difference between transpose operators and character arrays?

For example, with

B = A1 + A2' + A3 + A4'

I am seeing the second half of the RHS highlighted oddly because the editor thinks it's forming a character array.

Wouldn't it be very helpful (and obvious) to allow a different operator to do transposes?? You could always do the PITA way of defining separate variables for each transpose you want to do, but that defeats the point of defining the operator in the first place and you might as well just use transpose() then.

[Matt J]

Don't know what you mean. The Matlab Editor is giving me know so such issues:

[leka0024]

Here is a screenshot of what I see:

That's in Jupyter Lab, and is my issue now.

However, in VS Code with MATLAB extension it is smart enough recognize the difference:

So I guess the issue is with Jupyter Lab (or the jupyter-matlab-proxy ).

However I think the general gripe is still valid, in that the apostrophe character for transpose operator is not the best. Guessing it would take some monumental work that no one at MathWorks has any time for to be able to change it or define a new one, so I guess this thread can be closed.

[Steven Lord]

Guessing it would take some monumental work that no one at MathWorks has any time for to be able to change it or define a new one, so I guess this thread can be closed.

It likely wouldn't be that much work to change the operator that MATLAB associates with the ctranspose function. Honestly, if we chose to make that change most of this part of the work would probably be getting people to agree what character or characters to use to replace it.

However, it would be far, far, FAR more work (both inside MathWorks and for our users) to update nearly 40 years worth of MATLAB code (MathWorks was founded in 1984) to replace the ' operator with the new one. The scale of the backwards incompatibility would absolutely make this a non-starter as far as I'm concerned. It would be at best minimal bang for a Jupiter (the planet) sized pile of buck.

[leka0024]

@Steven Lord even better point, thanks. (Although I wasn't proposing to do away with ' as transpose, just to allow users to define an alternate.)

The MATLAB extension for VS Code gets the highlighting correct, so perhaps somehow the jupyter-matlab-proxy group could see how the extension team was able to do it. The ideal thing would be for these two projects to merge into one ultimate tool for MATLAB coding: Notebooks in VS Code. :)

[James Tursa]

I haven't found the apostrophe ' operator to be an issue particularly. But there can be ambiguities with the .' operator with regards to where the period belongs. E.g., in the following:

2i^3.'

Does the period belong with the 3 making the apostrophe ' a ctranspose? Or does the period belong with the .' making it a straight transpose operator? You can run it to see ...

2i^3.'

ans = 0.0000 - 8.0000i

2i^(3.)'

ans = 0.0000 + 8.0000i

(2i^3).'

ans = 0.0000 - 8.0000i

So MATLAB pairs the period with the operator and not the number in this case. Of course, in this particular case if you wanted the ctranspose there is no reason to have the period at all:

2i^3'

ans = 0.0000 + 8.0000i

But my point is you sometimes may need to be careful how you write code with the dot operators such as .' to make the code intent clear to the reader.

Answer 1

[Matt J]

A cheezy, but compact way you could define an alternative ctranspose operator (temporarily, within a particular workspace) is with a user-defined object that overloads mpower,

classdef myclass
   methods
       function out=mpower(base,obj)
           out=ctranspose(base);
       end
   end
end % class

>> T=myclass;

>> A=reshape(1:12,3,4)

A =

1 4 7 10

2 5 8 11

3 6 9 12

>>A^T

ans =

1 2 3

4 5 6

7 8 9

10 11 12

Answer 2

[Matt J]

This seems to offer Jupyter Lab integration in a way that preserves Matlab syntax highlighting,

https://github.com/mathworks/jupyter-matlab-proxy

though I cannot find an example there of what happens with A' and A.' expressions.

[leka0024]

Thanks @Matt J , I am indeed using jupyter-matlab-proxy, that is the first screenshot above (with Jupyter Lab 3.6, they mention on the repo that highlighting is not yet supported on JL v4).