Automatically insert appropriate node when 2 sources are wired together

[Jaegen]

I want to be able to wire one error cluster wire into another, and have the "Merge Errors" node be inserted automatically.  Like this:

I also want to be able to do the same thing with 2 wires of the same data type, and have the "Build Array" node be inserted automatically.  Like this:

Of course, we should be able to disable this if we want.

Jaegen

[elset191]

I'm intrigued.  However, I think there are just too many viable options for most datatypes.

[dthor]

For most datatypes, yes, there are many many diferent options when merging two wires. But with the Error Wire, I think having it auto-insert the Merge Error node is a great idea.

[altenbach]

We already have that functionality for wires containing dynamic data where it automatically inserts a "merge signals". Try it. 😉

Overall, this idea is probably too ambiguous, but for error wires (and maybe some others) I see some usefulness with this.

[altenbach]

Also, your first image is a it ancient, because we now have a resizeable merge error function. So, what should happen if a merge error already exists elsewhere in the wire? Should it (a) simple create another one or (b) find the existing one and extend it by one terminal?

[Jaegen]

I agree - this suggestion, at least for non-error wires, is probably too ambiguous.  I tried to think of configuration options that could be put in place to give us tighter control of what happens, but couldn't really think of anything that wasn't overly complicated.  I hoped smart people would chip in here with ideas

Christian: Wow - I did not know that about the dynamic data wire.  I figured I probably wasn't the first person to think of this idea.  Also, LV 2011 has only been out for a few months - I'm not sure it's fair to say "ancient" .

[altenbach]

It's more than a few months. A growable "merge errors" node was introduced with LabVIEW 2010. 😉

[Jaegen]

Shoot - you're right.  Working with 4 different versions at once scrambles my brain.  Though I'm not sure 14 months old counts as "ancient" either.

[AristosQueue (NI)]

> Though I'm not sure 14 months old counts as "ancient" either.

Try telling someone that a feature they want is at least 14 months away...

[Christina_R]

The Merge Signals auto-insertion for Dynamic Data was something we hoped would alleviate some problems we saw during usability testing of the Express feature set. (This was LabVIEW 7, which introduced Express VIs and the dynamic data type, among other things).

It had mixed results, which is why we kept it for dynamic data but didn't expand on the idea. If the auto-insertion does the "right" thing for what you're trying to do, it's wonderful. If it places something "wrong," however, it can cause even worse problems than a broken wire, especially for new users who are confused about how things work. There's the assumption that, if the product did something automatically for me, it must be the right thing to do, and can lead people to build erroneous mental models of how the system works.

So, in summary, I don't think the efficiency gain of this suggestion is worth the potential confusion.

[altenbach]

> If it places something "wrong," however, it can cause even worse problems than a broken wire, especially for new users who are confused about how things work

Christina, I fully agree, and that's why we also need this idea 😄