When Quick Drop inserts a division operator: wire up the numerator, and leave the denominator unwired.

[D*]

When you use Quick Drop to insert a division operation onto a wire: you almost always want the numerator to be automatically wired, and the divisor to be unwired.  Yet, Quick Drop always wires up the denominator.

[User002]

Here, here! What's the struggle in gaining acceptance? Seems like if you're working on wire X and you want to scale it you're going to insert a divide to get X divided by Y. The insertion as it is now actually inverts wire X. Surely that's the vastly minor use case.

[AristosQueue (NI)]

D* wrote:

> There is already special coding for specific nodes 

There is? Can you give me file name and line number in the C++ code? I'm happy to take a look at how that's implemented, but I don't know of any such thing.

> (And such a standard wouldn't make life easier anyway).

It would make life easier in general because nodes can (and do) behave consistently. Having nodes that respond in specialized ways can help in some situations but hurt in others, and it isn't always clear which one is the majority use case (or that the minority case is sufficiently minor as to be worth the tradeoff), so we tend to avoid such specialness. But if specialness already exists, I can take a look at it and its design history to re-evaluate this idea.

[D*]

"In Range and Coerce" wires up the middle of the three inputs.

"Select" wires up the bottom of the three inputs.

"Replace Array Subset" wires up the array terminal.

"Insert into Array" wires up the index terminal.

and so on....

So, I infer there exists specialization.  I can also tell there is no standard like "it always wires up the bottom terminal".  No need to be prickly about it, asking me for file names and line numbers, what does that accomplish except to stifle discussion.

Quick Drop Insert is a shortcut.  Shortcut are made to be faster.  If you'd really like a standard for Quick Drop Insert, discuss what you think it should be.  How about "when a node is inserted, in which that node intends to modify the data wire, the insertion should wire up the node's terminals so that the data wire is on the input and output terminals suitable for modification".

[wiebe@CARYA]

>"Select" wires up the bottom of the three inputs.

Not if the wire is a Boolean. If so the middle.

>"Replace Array Subset" wires up the array terminal.

Not if the wire is an integer. If so the index.

"Insert into Array" wires up the index terminal.

Not if the wire is an array. If so the array.

> I can also tell there is no standard like "it always wires up the bottom terminal".

There probably is, but it's a standard that might not make sense to you.

Something like "wire the first input that makes sense".

[Darren]

Based on the feedback in the thread, I've decided to re-open the idea. I'm starting to agree with the idea that the majority of the time, when you insert a function on a wire, you want the top-most appropriate input to be wired, not the bottom-most. So here are the changes I think we could get most everybody to agree on. Let me know what y'all think:

Wire up the top-most valid input in Quick Drop Insert for:

Binary numeric, Binary Boolean, Binary comparison, Select, Search and Replace String, Match Pattern, Search/Split String, Insert Into Array, Delete From Array

The way Quick Drop Insert is implemented, I maintain a white-list of special cases where I do post-insert scripting to fix stuff up. I would be adding special case code for the items listed above. Let me know what y'all think about that plan and if I'm missing anything.

[Darren]

Available in LabVIEW 2020 and later. Quick Drop now inserts binary functions with the top input wired instead of the bottom input.

[User002]

Much good karma channeled your direction!