Re: [Swarm-Support] Re: trouble with index

[Scott Christley]

emcleskey wrote:

> Scott,
> I did not have Flowfield.h #include 'ed in!  That was it.  I actually had 
> defunct -getXPos and -getYPos functions in the BrownTrout.m file as well (the 
> same place -step was) that I was initially going to use but ended up not 
> needing and then neglected to take out.  The complier was looking at those 
> functions and giving me the error message.  Thank you!
>
> So, the header file of the class the list contains must be included in the 
> implementation file where the method referencing that list is declared?  It 
> wasn't clear to me from the swarm user guide that I had to include that file.  
> Like I said, I'm relatively new to swarm and programming in general.  So if I 
> can ask one more question, same topic...
Yeah, this is more of an ObjC thing then anything to do with Swarm.  The 
point is that the compiler must see the declaration of the methods so 
that it knows the return data type(and parameters as well); it cannot 
assume that the return type is the same as your variable data type, 
because they are allowed to be different if one can be "cast" to the other.

> My code (called in the brownTrout class) indexes through one list created in 
> and containing objects of the flowfield class and then (as specified in the 
> 'while loop')adds them to another list CREATED in the brownTrout class BUT 
> containing flowfield objects.  If I then want to index through that list 
> (created in one class but containing items from another) and call functions, 
> do I put them in the brownTrout.m or flowfield.m file?
>
> It seems like the functions should be declared in the implementation file of 
> the class contained in the list.  Is that correct?

It is not exactly clear to me what you are doing.  If you are asking if 
BrownTrout should be allowed construct a list of FlowField objects, then 
this sounds reasonable to me.  The opposite is just as reasonable where 
a FlowField constructs a list of BrownTrout.  Which way you implement is 
more of a conceptual question for how you perceive your model.  Maybe 
BrownTrout are sensing local FlowField and performing actions based upon 
that, or maybe FlowField is detecting what BrownTrout are around and 
adjusting itself; or maybe both is occuring.  It is not uncommon for 
environment and agents to have a circular relationship with each other 
so that both viewpoints are reasonable.


cheers
Scott