Replacing Arithmetic or Boolean operator with Compound Arithmetic should automatically set Mode to match replaced operation

[SteveP]

Replacing a Boolean AND, OR, or XOR with Compound Arithmetic results in its Mode being pre-selected to OR no matter which boolean operation is being replaced.  It would be convenient if the replace operation would recognize from where it is coming and set the mode accordingly.  Similarily, replacing an Addition or Multiplication with Compound Arithmetic results in its Mode being pre-selected to Add.   A multiply operation should be replaced with a multiply operation (not an addition).

[crossrulz]

This is interesting...The duplicate has more Kudos than the original.

[G-Money]

Available in LabVIEW 2012

[X.]

It is completed in the eye of the developers. I would argue that the following is still missing:

===>   

====> 

====>

But of course, this would possibly freak out some users...

[AristosQueue (NI)]

X. -- you're welcome to post as a new idea. No one here on the forums ever mentioned those nodes nor did anyone around here think of them when doing the feature.  I take it back... Jack did, but it got missed. Nonetheless, posting as a new idea is the right way to go at this point.

[X.]

I have tried my chances with the QD author here. If that gets any traction, I may.

[Darin.K]

Consider yourself forewarned:  The performance of the equivalent compound node can be significantly worse than the primitive, especially for arrays.  For example, achieving subtraction by inverting a terminal is not recognized as subtraction, rather two steps are performed:  negate the input and add.  The result is correct obviously, but the two steps slows the process.  Same for division.   Similarly, a dedicated squaring operation can be faster than self-multiplication, and it is inplace as well.

On the other hand, exanded compound nodes (3+ inputs) can be faster than the chain primitives.  Probably because LV does not guess about which buffers to reuse and just goes to work.   All very interesting benchmarking but obviously subject to version to version variations.

[X.]

That's fine with me. After all, we are not asking for algebraic simplification, just consistency:

Can you find the outlier?