Rube Goldberg Code

[altenbach]

I think you got a nonzero chance to get a 5 digit number once in a while. You probably should round to -infinity somewhere.

[JW-JnJ]

Rounding would ruin your uniform random distribution... probably not enough that you would care.

[crossrulz]

---

@JW-L3CE wrote:

Rounding would ruin your uniform random distribution... probably not enough that you would care.

---

That's why you should use the OpenG random number.

http://lavag.org/topic/14981-random-number-integer-numeric-package/page__view__findpost__p__90015

[altenbach]

Well, since we convert to integer, rounding will happen no matter what. The original "toI32" will round to the nearest integer, meaning that the first and last possible bin (0000, and 10000) only have half the typical chance of being populated, and some outcomes will have five digits (10000). Stuff like that could kill all casino profits! 😮

If we multiply by 10000 then round down to the next integer, we get equal probability from 0000 to 9999. A truly uniform random distribution for all possible 4 digit integers! That is the way to go and that's what I proposed earlier. And I do care! 😄

[JW-JnJ]

---

@altenbach wrote:

Well, since we convert to integer, rounding will happen no matter what. The original "toI32" will round to the nearest integer, meaning that the first and last possible bin (0000, and 10000) only have half the typical chance of being populated, and some outcomes will have five digits (10000). Stuff like that could kill all casino profits! 😮

If we multiply by 10000 then round down to the next integer, we get equal probability from 0000 to 9999. A truly uniform random distribution for all possible 4 digit integers! That is the way to go and that's what I proposed earlier. And I do care! 😄

---

My bad, I read the "0-1" and not "less than 1". I was picturing poor ole 10000 with a 1/10000 chance. I should know better, I used RNGs all the time in Matlab. Not much reason to use them in them in test systems, heh.

They need a forum prompt: "Are you SURE you want to correct Altenbach?" with button "I rethought it, NO" and "Your funeral".

[Ray.R]

Thanks Christian,

I only needed to generate two random numbers to create unique indexes. 

I do like what you suggested and will apply it as a good coding practice.  I also like the openG version.

Since this didn't turn out to be a Rube, it is more of a good coding practice example, maybe it is time to create yet another thread to discuss good coding practice.

[LV_Pro]

So I helped inspire you? At it early I see.

[Ray.R]

You always inspire me, Putnam 🙂

Especially with your travels..  🙂

[J.Mamakos]

Seems I made my own little bit of Rube recently...

That's what you get for reducing code from somewhere else and not rechecking the logic. (the shame)  

[altenbach]

How many primitves do you need to make a subVI that returns one row and one column from a 2D array?

A single "index array" primitive would probably be sufficient (bottom), but we could also do it using 8 primitives and 4 diagram constants (top)

 

(seen here)