Rube Goldberg Code

[thoult]

---

@JÞB wrote:

That being said, if the optomization is known to exist at a lower level there is no need to duplicate it at a higher one.  Or, is there? - it would depend on the return in performance for each specific implementation I THINK.

---

Philsophical questions here:

- If the compiler will optimise without you seeing, do you care?

- If the gain in self-optimisation is small relative to the cost of refining your method, do you care?

- Does the optimisation preclude expansion, or at least make it difficult? Do you care?

- Does optimisation result in obfuscation? Do you care?

The first one's fairly specific to coding here, but the last three are the same for any system you work on, mechanical, electrical etc.

I'm a process developer - I can improve my process specifically for one variant at the cost of a negative effect somewhere else. I can improve my process unilaterally, but for such a small marginal gain that it's not worth it. I can change my process so that it's more efficient, but harder to maintain.

That last point is relevant again to software - we specifically avoid LVOOP in many situations where it would be perfect for implementation, purely because we have to make sure the code is supportable by a team which is less well versed in its use, and the 'cost' of maintenance is higher than other implementations.

I'm not saying that we purposefully implement RGs, but there are times when I've chosen more long winded solutions so that the working is more obvious, or because the most efficient method for my current problem is to ignore adaptation for future ones.

[JÞB]

---

@thoult wrote:

---

@JÞB wrote:

That being said, if the optomization is known to exist at a lower level there is no need to duplicate it at a higher one.  Or, is there? - it would depend on the return in performance for each specific implementation I THINK.

---

Philsophical questions here:

- If the compiler will optimise without you seeing, do you care?

- If the gain in self-optimisation is small relative to the cost of refining your method, do you care?

- Does the optimisation preclude expansion, or at least make it difficult? Do you care?

- Does optimisation result in obfuscation? Do you care?

The first one's fairly specific to coding here, but the last three are the same for any system you work on, mechanical, electrical etc.

I'm a process developer - I can improve my process specifically for one variant at the cost of a negative effect somewhere else. I can improve my process unilaterally, but for such a small marginal gain that it's not worth it. I can change my process so that it's more efficient, but harder to maintain.

That last point is relevant again to software - we specifically avoid LVOOP in many situations where it would be perfect for implementation, purely because we have to make sure the code is supportable by a team which is less well versed in its use, and the 'cost' of maintenance is higher than other implementations.

I'm not saying that we purposefully implement RGs, but there are times when I've chosen more long winded solutions so that the working is more obvious, or because the most efficient method for my current problem is to ignore adaptation for future ones.

---

Probably worth a dedicated thread for dicussion.  A bit OT for this mega thread

I guess the point of my last post was "It probably doesn't matter to modern processors but, since I'm blissfully ignorant of those low level details that LabVIEW and other layers abstract away, I'll fall back on my early training and avoid operating on the value of data that I'm going to throw out anyway, because its a good habit."

[altenbach]

Slicing and dicing an array of waveforms (seen here).

[JÞB]

---

@altenbach wrote:

Slicing and dicing an array of waveforms (seen here).

---

---

---

 Hacking out a bunch more useless code in following post.  The vi actually cleans up nice if you shave off the unneeded stuff

[altenbach]

No code shown, but VI Analyzer seems to complain that:

"This diagram contains 410 write global variables, which is more than the user-specified maximum number(5)"

I would say it is a good candidate for this thread. Just guessing! 😄

(seen here)

[JÞB]

Well, THERE'S the problem!

And I cannot even read the poster's native language:)

[altenbach]

---

@JÞB wrote:

Well, THERE'S the problem!

---

Next thing we'll see an "idea" to raise the default limit to 500. 😄

---

@JÞB wrote:

And I cannot even read the poster's native language:)

---

Here's what I use. Works great! 😉

[Former NI Employee MS]

This one is a good catch

A single for loop does the job

[altenbach]

---

@Former NI Employee MS wrote:

A single for loop does the job

---

This is a very long thread. Which post are you referring to?

[Former NI Employee MS]

You're absolutely right !

I had hoped that by clicking "Reply" it would somehow link the post I was trying to answer (or I made a mistake). I am myself currently unable to find back the post I answered