[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Fwd: Re: Passing Functions in Inputs



PureBytes Links

Trading Reference Links

-- [ From: Larry McBride * EMC.Ver #2.5.3 ] --

An example of what I'd like to be able to do is the following:

System initialization has -    Inputs: in1(0),in2(0);

I'd like to pass a function in through the inputs by setting in the
workspace say -
   in1 = momentum(C,in2),     in2 = 0...100,steps of 5 in the optimizer.

This will only work if you set, for example,  in1 = momentum(C,6)  (i.e.
you have to pick a specific number and can not therefore use the
optimizer on the input in2.)

I'm wondering whether there is any sneaky way to get around this that I
haven't found, whether you also pass the function through an include
statement later or not.

Thanks for any ideas, 
Larry

 
> Subject: Re: Passing Functions in Inputs
> 
> Larry,
> can you give a short EL example how you would pass a function ? Thanks
> Frank
> 
> (LARRY MCBRIDE) wrote:
> 
> > -- [ From: Larry McBride * EMC.Ver #2.5.3 ] --
> >
> > In a message from Bill Brower a while back, he wrote:
> >
> > >3) The biggest and most drastic action, without using GV's is to
use a
> > >heretofor undisclosed trick of feeding a large function in as an
input
> > and
> > >then passing the input to an IncludeSystem which can then perform
> > futher
> > >analysis.  The Calling system can be right near the 64K barrier and
so
> > can
> > >the IncludeSystem.  This way you can build systems much larger than
64K
> > ..
> >
> > Passing functions like this seems to work well, whether or not you
also
> > use an Include statement, but I don't think I can do what I'd really
> > like to do.  That would be to pass in a function as an input with
> > arguements that are parameters that can also be passed in as inputs.
If
> > this were possible, then the optimizer could be used to vary the
> > parameters of the function.  What it seems like, however, is that if
one
> > passes a function in as an input it  can only utilize arguements
that
> > can be fully evaluated when the input is edited.  These arguements
can
> > not reference another input also being passed into the system.  Is
this
> > correct?  Is there any way around it?  The additional flexibililty
would
> > be very powerful.
> >
> > Thanks for any insights.
> >
> > Larry
> 
> 
> 
> 

-------- REPLY, End of original message --------


------- FORWARD, End of original message -------