Scary Code

[altenbach]

Sometime we need to clear a string indicator by writing an empty string to its terminal. Here's the Rube Goldberg way.

• Of course if the content of the string was concatenated from 5 substrings, we need to concatenate five empty strings to match it 😮
• We also cannot really trust a single wire, so we need to branch it and write to the indicator terminal AND to a local variable of same.
• Now can we be double sure that that sucker gets really cleared. 😄
• For government applications, we should actually add a few more copies of the local variable to bring it up to specs.

(from http://forums.ni.com/ni/board/message?board.id=170&view=by_date_ascending&message.id=294413#M294413)

Message Edited by altenbach on 01-11-2008 11:14 AM

[Ben]

Christian wrote

For government applications, we should actually add a few more copies of the local variable to bring it up to specs.

You had me scared there for a second.   I do a fair amount of gov. work and need to keep up on req's.

 

Ben

[Ray.R]

LOL! @ Altenbach..

I can almost not believe that it was true..  Wow!  LOL!! 

LOL!!  😄 😄

Now that's a funny RG code!

LOL!!

[altenbach]

If monitor resolution was truly unlimited, we could easily manage a diagram that is 14285 pixels wide and 3354 pixeld high.

Here's how it looks if I scale it down to 5% (!!) of the original size in a graphics program: 😮

Note: each of the inner stacked sequences has 10 frames. Maybe we should convert those to flat sequences. 😄

Suddenly, a state machine does not sound like such a bad idea. 🙂

(from http://forums.ni.com/ni/board/message?board.id=170&view=by_date_ascending&message.id=298939#M298939)

Message Edited by altenbach on 02-01-2008 10:59 AM

[Ray.R]

Altenbach,

I admire your tenacity at looking at such diagrams.. 

It takes courage.... and probably some strong beverage at the end of the day.. 😄

[RavensFan]

Sometimes it may be better if the code does take up more than one screen!

This is the most densely packed code I've seen.

And what is the purpose of feedback nodes on error clusters that aren't even part of a loop?  It makes it hard to tell which is the first function that executes.  Is that so you can hold on to errors from the last time you ran the program?  "I had an error and stopped the program, but everytime I restart it won't work and I get the same error!  I can't figure out why!!!"

http://forums.ni.com/ni/board/message?board.id=170&view=by_date_ascending&message.id=299635#M299635

Message Edited by Ravens Fan on 02-05-2008 12:08 AM

[RayFarmer]

[And what is the purpose of feedback nodes on error clusters that aren't even part of a loop?  ]

maybe they are being used as a direction indicator

Regards

Ray

[altenbach]

Most likely the author of this code wire things randomly together, accidentally creating loops and these feedback nodes were inserted automatically (the "auto inser feedback nodes in cycles" option is enabled by default). For example you can delete the rightmost feedback node and simply wire the remaining ends together and the feedback node will magically reappear.

I wonder why all calculations are done in DBL, but displayed as SGL. 😮

[tst]

---

Ravens Fan wrote:

This is the most densely packed code I've seen.

---

I guess you weren't looking hard enough. I title this post "Scary Code" (note the zoom window):

[RavensFan]

You're right that is scary. And worse.  I especially like the amount of overlapping and parallel wires in the right hand side of the window of the top picture.

I was impressed with how in my example, the programmer worked hard to really compact everything, and by doing so was able to squeeze it into much less than one screen, so it was like they were trying to follow one style rule!