topic Re: Rube Goldberg Code in BreakPoint

Rube Goldberg Code

altenbach — Thu, 03 Aug 2006 15:09:09 GMT

A Rube Goldberg Machine is defined as an "exceedingly complex devices that perform simple tasks in very indirect and convoluted ways"

http://en.wikipedia.org/wiki/Rube_Goldberg

One example would the the simplfied pencil sharpener.

Let's define Rube Goldberg Code as ""exceedingly complex LabVIEW code that perform simple tasks in very indirect and convoluted ways"

For example, let's have a look at the code to test if two booleans are equal. Just beautiful! 🙂

Unfortunately, the beauty of this is lost to most of you. 😞 To test this point, I have recently posted an obvious joke to see if anyone actually pays attention.... Apparently not!!! Nobody commented on my example that I snuck into the tic-tac-toe discussion. For comparison, here's my actual code detail compared to the posted image. 😄 Duuhhh!!!

Anyway, I am sure that most of you have some great examples of such code. This thread is the place to share the fun! 🙂

I am not really looking for intentional constructs (anybody can do that!). The focus should be more on existing serious code where you suddenly noticed that there is a much simpler way.

(Many code examples posted in the forum contain mini versions, e.g. the use of "delete from array" with "deleted portion" as the only output wired, used as a substutute for "array subset" seems quite popular ;))

Message Edited by altenbach on 08-03-2006 08:10 AM

Re: Rube Goldberg Code

Darren — Thu, 03 Aug 2006 18:03:15 GMT

I'm not making this up...the following was in a VI I was doing a code review on several years ago:

You'll all be happy to know this person no longer works at NI.

-D

Message Edited by Darren on 08-03-2006 01:03 PM

Re: Rube Goldberg Code

Ben — Thu, 03 Aug 2006 18:46:52 GMT

When I was "wee-G-Babe" I wanted to sort folders based their size.

I did not know about the "Cluster Sort" method so I coded up a variation on a bubble sort.

Greg McKaskle said he was stumped for a while but eventually figured it out.

This was one of the first "Bad" from the "Good the Bad and the Ugly" series.

Ben

Message Edited by Ben on 08-03-2006 01:47 PM

Re: Rube Goldberg Code

jasonhill — Fri, 04 Aug 2006 14:59:21 GMT

I am one of those sick bastards that actually enjoys straightening other people's spaggetti. I have seen some impressive examples of Goldburg constructs. Recently I ran across the gem in the top illustration. To the guy's credit, in the original code the selects were nowhere near eachother. The equality, however, is inexcusable.

Disclaimer: I am guilty of this sort of thing too. In fact, some examples I have posted have glaring Goldbergs.

Altenbach, you have such an established guru reputation that I am sure many people assumed there was some obscure reason why you set the example up that way. Especially in the highly optimized enviornment of a coding challenge.

Message Edited by jasonhill on 08-04-2006 10:00 AM

Re: Rube Goldberg Code

altenbach — Fri, 04 Aug 2006 15:15:39 GMT

I think I saw a variation of this in PRIME.vi posted yesterday. As a bonus, it also contains Darren's construct inside the case structure. 😉

So how do we go about constructive criticism in these cases? I am sure a few nudges in the right direction can make a big difference. Still I am impressed that people with no obvious LabVIEW or general programming background can get a fully working program together. 🙂

Re: Rube Goldberg Code

TCPlomp — Fri, 04 Aug 2006 19:18:57 GMT

Hi there:

Here is a golberg code about activating a boolean after an enter,

Sorry Altenbach..

Ton

PS which code was faster (first post)?

Message Edited by TonP on 08-04-2006 09:36 PM

Re: Rube Goldberg Code

JeanPierre — Fri, 04 Aug 2006 23:03:27 GMT

Search for the numerous implementations of exclusive radio buttons (before the native ones)
For example here chutla's implementation and the elegant solution of mine.

Do they still teach how to make truth tables?

Message Edité par JeanPierre le 08-04-2006 07:04 PM

Re: Rube Goldberg Code

JeffBoettcher — Mon, 07 Aug 2006 16:02:50 GMT

Ok, this is too good to pass up. I do not mean to offend the original poster(s) but check out the following:

Launch LabVIEW 8.0 (7.1 doesn't allow browsing web examples).
Launch Example Finder.
(You may need to click "Setup", click the Web tab, and increase your number of web hits to 50 or more.)
Include ni.com examples.
Browse into Fundamentals - Arrays and Clusters.
Scroll to the bottom of the list and you'll see a few examples with globe icons and an orange user icon which indicate they were submitted by external users.

Among the 10 or so user-submitted examples in this category are two that are particularly noteworthy.

2D Array to number of Rows and Columns.vi
2D_Array_to_number_of_Rows_and_Columns - revised.vi

Before opening them, think about how you'd write a VI with this name... For me, I was thinking Array Size followed by Index Array if you wanted them as scalars. Both of the above VIs are equally effective, but I'm not sure that's how I'd teach someone to do it. 🙂 Definitely not too efficient.

Re: Rube Goldberg Code

Ben — Thu, 04 Jan 2007 13:54:33 GMT

In this thread

http://forums.ni.com/ni/board/message?board.id=170&message.id=222491&jump=true

I posted a method to copy the contents of a cluster to another cluster.

Ben

Re: Rube Goldberg Code

tst — Thu, 04 Jan 2007 15:58:48 GMT

@Ben wrote:

In this thread I posted a method to copy the contents of a cluster to another cluster.

And here I was thinking that this code was supposed to copy an array of variants into the cluster...

Re: Rube Goldberg Code

Ray.R — Thu, 04 Jan 2007 17:55:03 GMT

I see code as shown by Jason (#4) and Darren (#2) all the time... Especially Darren's example.

Then again, I've been corrected by much simpler code while posting answers to this forum, too..

where's the smiley-guilty icon??? :womanwink:

Re: Rube Goldberg Code

RayFarmer — Thu, 04 Jan 2007 21:54:23 GMT

we are probably all guilt of these bits of code sometime or another.

watch out the person whom professes to be innocent and posts example code on these types of forums.

Re: Rube Goldberg Code

Ben — Tue, 23 Jan 2007 19:27:40 GMT

Here is another one.

I included both states just incase you thought there may have been a goode reason hiding in the other state.

Here is another in the same VI

Ben

Message Edited by Ben on 01-23-2007 01:28 PM

Message Edited by Ben on 01-23-2007 01:31 PM

Re: Rube Goldberg Code

Darren — Tue, 23 Jan 2007 19:50:52 GMT

I've seen the array size/indexing thing a LOT before...I wonder if there's an easier way we could teach new users how auto-indexing works...

-D

Re: Rube Goldberg Code

shoneill — Tue, 23 Jan 2007 21:49:16 GMT

Ben,

I have them all over some code I inherited from a previous employer at our company.

I honestly studied it for a while thinking ther HAS to be some significance to that code (Regarding the error cluster). I mean, come on.....

In the end I convinced myself I WAS actually sane (Something I feel I have to do on an almost daily basis) and it had no purpose whatsoever.

Shane.

Message Edited by shoneill on 01-23-2007 10:49 PM

Re: Rube Goldberg Code

altenbach — Tue, 23 Jan 2007 22:11:24 GMT

@Ben wrote:

Here is another one.

We defined Rube Goldberg code as "exceedingly complex LabVIEW code that performs simple tasks in very indirect and convoluted ways" (see first post).

This (and some of the examples shown earlier) is now actually a Rube Goldberg NOP (or Rube Goldberg NOOP), freshly defined as:

"exceedingly complex LabVIEW code that performs absolutely nothing in very indirect and convoluted ways". 😄

(We could also define the construct as a WEQ (wire equivalent code), code that can be replaced with a plain wire without any change in functionality. ;))

Re: Rube Goldberg Code

Ben — Wed, 24 Jan 2007 13:50:00 GMT

First the legal stuff:

Warning: The code posted previously was written by "proffesionals". Do NOT attempt to write this code yourself.

Excellent point Christian. (Leave ti to one our academic types to recognize when we need a new term ).

Just to help me understand the differences...

WEQ's are identity operations and satisfy the condition

A = B

where:

A = Input

B =Output

Can we say the same about RG-NOP's?

Shane wrote;

That construct is all over this code I recieved. I wonder if it was the same developer or if they both came from the same LV school of thought.

After stepping outside the "Good Developer Box" for a second it hit me what that code was for!

It is a very handy way to make sure the sub-VI does not return an error. If the sub-VI is failing, just switch the boolean constant to false!

The the very thought alone gave me the sensation of "finger-nails on the chalk board".

If this is indeed the case, then the code I posted is not RGC or RG-NOP or WEQ...

It's a hack!

Ben

Re: Rube Goldberg Code

Ray.R — Wed, 24 Jan 2007 16:41:38 GMT

LOL !!! @ Ben!!!!

Whoever coded that must have worked here (customer) too!!!

Now talk about useless code!!!! If it's No Error, then nothing is done to the error cluster values. If there was an error, the status was ALREADY TRUE!!, so again no change!!!

I'm spending hours upon hours at changing code like this..

Re: Rube Goldberg Code

altenbach — Wed, 24 Jan 2007 21:31:00 GMT

Here's another WEQ, straight from the forum :):

See: http://forums.ni.com/ni/board/message?board.id=170&message.id=225862#M225862

Message Edited by altenbach on 01-24-2007 01:31 PM

Re: Rube Goldberg Code

shoneill — Wed, 24 Jan 2007 22:25:46 GMT

Hey Ben,

"It's a hack!"

You're forcing me back to another round of sanity checks.....

Somehow I always knew LabVIEW was referring to ME when it complained about "Insane Objects"......

Shane