LabVIEW Idea Exchange

cancel
Showing results for 
Search instead for 
Did you mean: 
0 Kudos
Taylorh140

A new type of object reference thing

Status: Declined

Any idea that has not received any kudos within a year after posting will be automatically declined. 

So I often find myself wanting "pointers" Smiley Frustrated But they are weird and not flowish in the labview world. Most of the time I don't want them for concurrency i want them for the ability to create complex data structures like link lists and such. And yes we can make them using DVR or priority Queues but it just seems like unnecessary overhead. I mean Pointers are really Really fast. So what options do i have.Options.png so who wins above? fastest to slowest:

 

Flow Count

ArrayRef

Variant Count 2

DVR Count

Queue

Variant Count

 

So I guess i could use a array and use indexes for reference but that would be tedious! Plus I can't remove an element from an array without ruining all the other references! So I suppose i would makes something a mix between a ARRAY a Variant and a DVR the data flow would be like the array and the references would be typed like a DVR wire. I Imagine it looking something like this:

 

I call it a SCOPE because i honestly don't know what to call it.

 

SCOPEIDEA.png

 

Well what do you think would this be useful to you?

2 Comments
AristosQueue (NI)
NI Employee (retired)

You are working on a problem I have worked on off-and-on for 15 years. In a dataflow-consistent manner, how do you create a data structure that has internal references?

 

Your picture matches match some of mine. Essentially, we need a by-value macro structure that converts to a whole bunch of by-reference micro structures at the border of an IPE. The trick is figuring out where that "end of flow" is -- you don't want to define it at every FPTerminal or you eliminate the ability to create subVIs.

 

Whether or not scope refs should allow recursion is a separate discussion. There are pros and cons there. But that's a conversation to have if the main grouping problem gets solved.

 

You aren't by any chance going to be at the CLA Summit? NIWeek? Or in Austin for any other reason some reason? This is one of those topics that is much easier to talk about in person than over forums.

Darren
Proven Zealot
Status changed to: Declined

Any idea that has not received any kudos within a year after posting will be automatically declined.