what does these three dots mean in this equation...

diff_im = diff_im + ...
                  delta_t*(...
                  (1/(dy^2))*cN.*nablaN + (1/(dy^2))*cS.*nablaS + ...
                  (1/(dx^2))*cW.*nablaW + (1/(dx^2))*cE.*nablaE + ...
                  (1/(dd^2))*cNE.*nablaNE + (1/(dd^2))*cSE.*nablaSE + ...
                  (1/(dd^2))*cSW.*nablaSW + (1/(dd^2))*cNW.*nablaNW );

Answer 1

[Walter Roberson]

The three dots mean line continuation. MATLAB expressions normally end at the end of the line unless they are specifically continued (exception: within the [] list building operator.)

[Peter Vranken]

The documented meaning of the three dots is the line continuation. However, I saw a script using the token as a comment opening character at the same time. Evidently, all the rest of the line behind the dots is ignored by the parser and can be used for comments.

My question: Is this a well-defined, documented, officially supported feature or is it just a deprecated side-effect of the token?

[Stephen23]

This clearly recommends using the ellipses for commenting out:

https://www.mathworks.com/help/matlab/matlab_prog/comments.html

"To comment out part of a statement that spans multiple lines, use an ellipsis (...) instead of a percent sign."

[Steven Lord]

Yes, this is a documented feature.

"Three or more periods before the end of a line cause the MATLAB® software to ignore the remaining text on the current line and continue the function on the next line. This effectively makes a comment out of anything on the current line that follows the three periods"

[Jan]

@Stephen: Using the ellipsis in such cases avoids the ambiguity for the line breaks:

header = [1 ...
          ... 2 
          3]

This is [1, 3], but:

header = [1 ...
          % 2 
          3]

is [1; 3], a column vector, because the implicite line break is taken as separator of lines. This even happens, if comma is used as separator:

header = [1, ...
          % 2 
          3]

replies [1; 3]. Brrr, not intuitive.

You see, this is important. Please post this as additional answer.

[Adam]

Certainly the lack of consistency is not intuitive, but

header = [1 ...
          ... 2 
          3]

producing [1, 3] is not overly intuitive either!

I often wanted an option like C++ /* */ commenting in Matlab, but now that I see it it is so ghastly I don't think I would ever dream of doing such a thing in code. Even if it is documented, using the line break operator as a comment tool is horrible!

[Walter Roberson]

Historically, /**/ is C style comments, C++ recommended style is // commenting to end of line (but /**/ is supported for backwards compatibility.)

[Jan]

@Adam:

 using the line break operator as a comment tool is horrible!

Do you mean the "line continuation operator" or that the linebreak in the text is interpreted as row break in the matrix? I agree with both.

[Adam]

Well, I was meaning the line continuation operator. I have been known to have functions in C++ where I comment out one argument temporarily from the middle of the list (either formatted with the full function signature on one line or, as I often do in Matlab, with one argument per line), but at least there it has more of a look of being commented out, being surrounded by operators that would not otherwise be anywhere near.

The idea of having 1 ... on one line and ... 2 on the next line meaning that only the 1 is actually part of the code is not programmer-friendly. Having just tried it for the first time though I notice the '2' does at least turn green to show it is commented out.

Answer 2

[Jan]

I summarize the important comments of Stephen Cobeldick and Steven Lord above to make this more prominent:

Beside the line continuation, the ellipsis '...' starts a comment also:

https://www.mathworks.com/help/matlab/matlab_prog/comments.html:

 To comment out part of a statement that spans multiple lines, use an ellipsis (...) 
 instead of a percent sign.

See also: special characters.

Using a % is interpreted as line break and therefore as an implicit row break of the data:

header = [1 ...
          ... 2       % line break commented away
          3]
>> [1, 3]

header = [1 ...
          % 2          % line break interpreted as row break
          3]
>> [1; 3]

header = [1, ...
          % 2 
          3]
>> [1; 3]              % ???

header = [1; ...
          ... 2 
          3]
>> [1; 3]              % As intented obviously

Only the last version looks intuitive for me. I'm still convinced that it was a bad design idea to allow spaces as column-separators and linebreaks as row-separators because this leads to ambiguities.