02-05-2009 03:24 PM
"... 6-8 days would be enough ..."
So, in reality 12-16 days. Take your best, best, BEST estimate and double it. In my experience, you're usually cutting it close ... and frequently go over that. ![]()
02-05-2009 03:30 PM
Yes, that can be a problem for now. Once you get some of these things worked out and switch to a state machine architecture that will no longer be a problem because the program will not stay long in a condition where it cannot respond to user inputs. One of the earlier responses in this thread suggested going to an event driven method and I agree.
What I did was not intended to be a working version, but a demo of how the process could be started. I am quite sure that the Cycling loop can be reduced to a single while loop with a single case structure inside. A few cases may have small case structures, especially boolean controlled ones. Eventually you will have two loops, one with the event structure and the other with the big state machine.
Keep it up. As you are well aware you are dealing with a messy problem. There is no easy way to clean something like this up. Sometimes it will seem like two steps forward and one step back (or maybe three back), but you can do it.
Lynn
02-05-2009 03:44 PM - edited 02-05-2009 03:50 PM
hahaha, i can't ask my company to hire a pro to do it..
its just saying..
"hey guys, i can't do my job..."
i would love to, but i dont think thats a smart idea. and besides...my manager is happy with rev B.
just that i am not..
hahaha....just out of curioiusity..how would i go about getting someone to do this?
where can i find help?
how much do they cost?
02-05-2009 03:53 PM
johnsold...
i don't see how i can make this any smaller, into 1 state machine..
i do have a state machine in my first loop...
and i do have an event driven loop. (that's where i normally add new materials)
but the big while loop..i dont see how you can make that any different..
02-05-2009 04:35 PM
There is a Forum for LV Job Openings. People sometimes post there when they need a job like this done.
State machine... how do I describe what I am thinking?... In the Cycling loop (I am working from memory as I have shut LV down) one case structure (Call Modules?) has about 15-20 cases, each of which basically performs one task. This would be the core of the state machine. Each frame in the sequence structure might be a candidate for another case (state). All of the cases which are hooked to booleans: One or two states each. Once Cycling becomes a state machine, it can be merged with the other state machine. Similarly the contents of the other loops could become states.
In some ways it would be better to start over from scratch. Take the working (or mostly working) program you have now as the basis for defining what the program is supposed to do. Write down the inputs and the corresponding outputs. Draw up a flow chart or state diagram. Then design a clean program from the ground up which does the same thing. You might end up with three parallel loops: one with the user interface and the event structure, one for I/O to the hardware, and one for the control logic and data processing. Communicate between loops with queues rather than global or local variables.
Lynn
02-05-2009 04:40 PM
hahaha..
johnsold, are you accepting applications?
i have an opening for you..
this will be coming out of my own pocket so please be kind..hahaha.
02-05-2009 07:06 PM - edited 02-05-2009 07:10 PM
this is my current working version with the software i have..
it has alot of local variables..
it is working right now..
if a user decided to press any buttons, the locals for them will change..
i want to break away from that..
do you recommend me trying to do three parallel loop?
event
temp
press?
02-05-2009 07:21 PM
also, i want to change the cluster to typedef, but i am using labview 7.0.
how do i go to Edit and CUSTOMIZE CONTROL?
02-06-2009 07:10 AM
02-06-2009 07:10 AM