Influence of order of parameters in procedure 1

[Hermanator]

I hope that somebody is able to solve this little mystery:

I am working on a numerical simulation program. It calculates a certain device current and puts this in an output file. To keep track of the changes I make to the program, I compared the output files of different versions. After making some small changes that should not influence the calculation algorithm, I observed a small change in the output values. After a long search, I have managed to track it back to the following:

In the program a procedure was declared:

PROCEDURE Calculate(VAR x: vector; y, z: vector);

Where "vector" is a declared type of an ARRAY of DOUBLE. Obviously, since all parameters of the procedure are vectors, I changed that line by:

PROCEDURE Calculate(VAR x, y, z: vector);

This change causes a small change in the output values, while it should not at all influence the calulation algorithm.

Do the 2 different options signify something different?

 

[Glenn9999]

The only influence on results for "var" is that it enables the direct passing of a variable, which may be changed. For the first procedure definition, X is saved/changed out of the procedure, while Y and Z are not. In the second one, all three are saved/changed.

To simplify things, it really doesn't matter for input, but output of a procedure requires "var". But the best practice is to not pass var for input variables and pass it for output. The first procedure type tells me that X is an output variable, and Y and Z are input variables.

But overall, whether a result is true or not should not be influenced by defining procedure variables one way or the other.

Dilbert is not a fictional cartoon. It is a documentary.

 

[Hermanator]

Of course, that's it. I failed to realize the significance of "VAR".
It turns out that both y & z get changed inside the procedure, but that this change shouldn't be passed to the main loop.

Thanks for your help.