List just returns the tree. I know ECJ does this but I was looking for a means of 
getting the equivalent of a depth 1st traversal of the leafs without the nodes. I 
found that eventually in gpsemantics. 

One of the problems I had was in setting up a function to return the leafs. In 
gpsemantics, SemanticJ.java defined a join. I had already given up on the idea 
of the three list functions in favour of one join function. SemanticNode.java 
doesn't look at the children, which was one of the things I found confusing 
from the other examples.

I haven't gotten to testing this yet, but I think that part should work well 
enough now. 

Not what I was looking for, but passing a list of data between nodes sounds 
interesting...

What has that been used for?

-- ray

On Mon, 3 Mar 2014 16:37:05 -0500, Sean Luke <[log in to unmask]> wrote:

>On Mar 1, 2014, at 1:05 AM, Raymond Shpeley <[log in to unmask]> 
wrote:
>
>> This is probably painfully obvious to most... How do I create a Koza list 
function?
>> Koza had Listx, with x being the arity of the function. I'll use this for testing 
only
>> and at this time it will be fed by ERC terminals as Sean previously 
suggested.
>>
>> I'll create 3 list functions, List16, List17, List18, and an ERC with an integer 
range
>> of 1-26, to create trees which return either 16, 17, or 18 ints in a list. That 
list
>> will be checked for conflict constraints, ie if 6 then not 7, producing 32 
unique
>> individuals. I realise Koza needed the Listx function because he was using 
Lisp,
>> I'm just wondering how this is normally done.
>
>I don't know what you're referring to by a listx function -- this just be a Koza 
III thing; at any rate, not normally used in GP.
>
>If you want to pass a list of data between nodes, this can be done, just set 
up a pointer for it in your GPData object.
>
>Sean