Rube Goldberg Code

[altenbach]

To get two columns from a 2D array we could (seen here)

1. transpose the array
2. get the second row (was second column before the transpose)
3. measure the size of the second row (i.e. the size of the original first column)
4. reshape the entire 2D array to a 1D array with the same size as the second row to get the first row, (which originally was the first column).

.... or we could just get the two columns directly (note the zero wired to the lower index). 😄

 

[altenbach]

Famous last words, completely redefining the word "simple" (seen here😞

"Hi, I'm working on a simple code with bunch of flat sequence structures and for loops."

[Yamaeda]

---

wiebe@CARYA wrote::

---

Goto is a syntactically correct statement in C++...

Probably the fastest way to getting fired if you use it, but still: syntactically correct.

---

It's the Local Variable of C/C++, it's easy to abuse but have it's uses. 🙂

[rolfk]

---

@Yamaeda wrote:

---

---

It's the Local Variable of C/C++, it's easy to abuse but have it's uses. 🙂

---

Did you mean to say Global Variable? Use of locale variables in LabVIEW is definitely a valid use scenario when programming more complex UIs.

As to goto in C, yes it is a bad fellow to use, but I have done so myself quite a few times. Without structured exception handling it is the next best thing for error handling if you do not want to have deeply indented code. Of course one counter argument could be that once you end up having more than 3 or so indentions, you should probably have refactored the code into sub functions.

[santo_13]

Here is a code by an intern with many Rude Goldberg, for instance, the following cast structure implementation is just to get a subset of the DDT wires

Seen here - https://forums.ni.com/t5/LabVIEW/error-209836-shows-after-program-runs-for-12-hours/td-p/4252717

[wiebe@CARYA]

---

@santo_13 wrote:

Here is a code by an intern with many Rude Goldberg, for instance, the following cast structure implementation is just to get a subset of the DDT wires

---

And the double wired to a case selector is bad enough by itself...

[BowenM]

I feel bad for that intern. He was given a difficult task (for a newbie to LabVIEW) in something he wasn't trained to do with virtually no oversight and given an extremely tight deadline to do it. The fact he got as far as he did given the circumstances is impressive. Not only that, his post is courteous and shows that he's already learned a bunch from the start of the project but just doesn't have time to refactor his code.

[santo_13]

---

@BowenM wrote:
I feel bad for that intern. He was given a difficult task (for a newbie to LabVIEW) in something he wasn't trained to do with virtually no oversight and given an extremely tight deadline to do it. The fact he got as far as he did given the circumstances is impressive. Not only that, his post is courteous and shows that he's already learned a bunch from the start of the project but just doesn't have time to refactor his code.

---

It is indeed impressive for an intern to develop something that large without any formal training.

[altenbach]

The IPE (index array) is of course useful if you care about the existing value. If we don't care about the existing value, it's just a glorified "replace array subset". 😄 (seen here)

(Yes, that entire thread is weird in general, trying to solve a very poorly defined problem)

[altenbach]

So the simple task is to build xy data for an xy graph in a loop, one point at a time.

(seen here)

Apparently, we need:

• four height=1 "built array" nodes
• two "insert into array" nodes. (Yes, they are actually wired correctly. No way to tell from the picture that the wires are actually crossed behind the primitive!!!)
• two shift registers
• two "array size" nodes (they will both always show the same value)
• a bundle node
• etc.

Seems a bit much! I have the nagging feeling that it could be simplified. 😄