Wiring tunnels, default data....

[tartan5]

Pincpanter, that's a nice trick!  I'll definitely keep that in mind!

Centerbolt, what are the advantages / disadvantages of using a functional global?  Coming from a programming background, I tend to not use globals by habit, but perhaps in Labview they lend themselves to certian applications?

[altenbach]

---

@centerbolt wrote:
In this case why not use a functional global instead of the shift register?

---

Sorry, I don't understand you comment in the context of this thread. Which post are you referring to?

[PJS]

I generally use a cluster to pass data like this.  In this way, if I need to add or remove data going to cases, it is as simple as adding or deleting from the cluster.  and then bundle and unbundle the specific variables I need in the cases I need them.  Voila, only one wire ever transcends every case.  As for functional globals.  it may be an easy implementation, but my thought on that is globals (however you implement them) are just that, and meant to be used when data needs to be set or read outside of one VI and not used as a local implementation.  But heck, if it works for you.. do as you see fit.

If you are using LabVIEW 8.5 on windows, feel free to download my queued message handler template from my samples, it shows this cluster implementation in action.
Good Luck,
Paul

[tst]

I would also go with Lynn's suggestion of placing all your wires into a typedef cluster which you pass through the SR. That's what I usually do, but that isn't really that practical for small loops where you're doing a simple thing. Then again, for small simple loops you usually have many cases.

Paolo, nice trick.

[Wayne.C]

@ altenbach

Instead of using the shift register in the original post, the contents of which are only used in one of the cases and require all the extra wiring we all hate to do, why not store the data using a functional global?  This gets rid of a lot of clutter on the block diagram.

[MichaelAivaliotis]

---

@smercurio_fc wrote:
As tst mentioned, the Tunnel Wiring Wizard can be very handy in these cases. There's a thread on this here.

---

Here's the original thread on the Tunnel Wiring Wizard back in 2004. The 8.0 version is here.

You can make more of your own scripting tools like this by poking around the VI Scripting Forums here.

[Henrik_Volkers]

Did I miss something?  If data is changed only in one of n cases LV offers the feedback node !

Or is the use of the feedback node a NONO?

However, since in most of my cases I already have (at least) one cluster running through, I seldom use the feedback node.

[tst]

---

@Henrik Volkers wrote:

Did I miss something?  If data is changed only in one of n cases LV offers the feedback node !

---

At least in 7.x feedback nodes can not be inside structures. They have to be in the loop itself.

8.5 now gives the options of having a feedback node without any loop.

The main reason I rarely use FNs is that they are harder to understand than SRs. A minor reason is that they have some performance issues.

[Gabi1]

some time ago i had the same question: why keep only default values, and not having option for "last value registered".

at the time Jhoskins said it would be hell to debug.

i guess he is right. although, that would make my codes so much nicer

cant find the thread anymore...