Building a method with more than one result data

Question

Hi, everyone:
I'm a little shy to ask this question, however, it's been hanging in my mind for so long, so I think I'd rather make a confession on it. You may laugh at me if you want, I'm ready for that, but I more look forward to that someone can really give me the light, or even the link, or some hint....
For your ease of reading, I give the question first, and my whole story behind:
When I need a method which can provide more than one result( in other words, multiple outputs), how can I do it in Java? As I know, either you pass and object, or the result of the function is an object will do , for the object contains the datas you want, but that means your needs for those data have to be defined in object format in advance, won't that be inconvinient? Or Java has a better solution for that?
//And here's the whole story....
I began my career as a programmer by starting with LabVIEW, it's a graphical programming language made by National Instrument, and it's powerful on DAQ, and industrial field. One of the most important issues on design is to devide your system into multiple functions( in its own term: subVI), I think it's just like applying structured analysis method.
When we dealing with functions in LabVIEW, a programmer can define his own function with mulitiple inputs and outputs, for example, I can design a function called SumAndDevide, which accepts two input ( two variables to be summed and devided) and gives two results( result of summing and that of deviding).
The methodology has its power, at least it provide the functional decomposition, and you can compose a suitable solution in certain circumstance even they are not the smallest unit function. And testing is easy. It affects me so large that I look the trail of it when I come to other programming languages. In COBOL( well, that is a VERY old COBOL version ), I was scared to find there is no protection to the inner data on the performed sections, while making a outside subroutine to be called is cubersome and really a hard work. When I came to Delphi, I knew that using the result of a function cannot satisfy me, for it give only one output, even you can define it as variant, but I think it's vague to realize. So I use the difference of called by value and called by reference to handle the problem, that is: a value parameter for the input, and a variable paramter for the output.
Well, when I came to Java, I am stunned again, now there is no passing by reference mechanism in Java, otherwise you have to pass it as an object, but that means when you need multiple outputs, the output has to be defined in object form in advance. And that will be very inconvinient!! I tried to find some solutions, but I can't. So is there any way that in Java you can define a method with multiple output? or Java handles the problem in totally different way?
Any comments will be appreciated!!
Thanks!!
aQunx from Taiwan

Answer 1

You missed the most common OO solution - separation of concerns and implementation hiding.
If you have a function which returns a string, that is one method of the object that provides the service.
If you have a function which returns a real, that is a different method of the object.
If both functions require common code, move that into a private method which is called by both. If the method is costly, cache the result.
eg an aerodynamics properties class, which could be done as a multivalued return of (lift, drag), refactored to independent lift() and drag() methods, which delegate to an interpolate() method, which caches the interpolated value and uses mach, pressureHeight and _alpha to determine whether it should recalculate:  /**
   * Calculates the aerodynamic drag force at a given mach, alpha and pressure height.
  public double drag (final double aMach, final double aPressureHeight, final double aAlpha) {
    interpolate(aMach, aPressureHeight, aAlpha);
    return _drag;
   * Calculates the aerodynamic lift force at a given mach, alpha and pressure height.
  public double lift (final double aMach, final double aPressureHeight, final double aAlpha) {
    interpolate(aMach, aPressureHeight, aAlpha);
    return _lift;
  private void interpolate (final double aMach, final double aPressureHeight, final double aAlpha) {
    if (aMach != _mach) {
      setMach(aMach);
      _pressureHeight = Double.NaN;
    if (aPressureHeight != _pressureHeight) {
      setPressureHeight(aPressureHeight);
      _alpha = Double.NaN;
    if (aAlpha != _alpha) {
      setAlpha(aAlpha);
... actual interpolation happens in the private setXXX methods.