XNode Usage

[hooovahh]

Okay so I've never really dipped my feet in when it comes to XNodes.  At the moment I'm not looking to make XNodes just understand using them, and really just get others to tell me it's safe.

 

The ones for discussion are the Array XNodes, and the OpenG Array XNode.  I see the functions this code has and I love the idea of using it.  But I know that there is no support from NI for using XNodes, and I've heard NI say that using them can be quite dangerous (without specifically saying what it could do or how to avoid it).

 

So I want to pose this question.  Is the Array XNodes in the Code Repository as safe to use as the normal OpenG array functions?  Is there a compelling reason I should not be using XNodes of any kind in an application?  I assume Lava would not certify code that was unstable, and I also see that people are using these functions (over 3000 downloads combined).  So is there nothing to worry about here?  And if these are so great and stable, should I exclusively use these instead of the real OpenG ones?

[Rolf Kalbermatter]

Okay so I've never really dipped my feet in when it comes to XNodes.  At the moment I'm not looking to make XNodes just understand using them, and really just get others to tell me it's safe.

 

The ones for discussion are the Array XNodes, and the OpenG Array XNode.  I see the functions this code has and I love the idea of using it.  But I know that there is no support from NI for using XNodes, and I've heard NI say that using them can be quite dangerous (without specifically saying what it could do or how to avoid it).

 

So I want to pose this question.  Is the Array XNodes in the Code Repository as safe to use as the normal OpenG array functions?  Is there a compelling reason I should not be using XNodes of any kind in an application?  I assume Lava would not certify code that was unstable, and I also see that people are using these functions (over 3000 downloads combined).  So is there nothing to worry about here?  And if these are so great and stable, should I exclusively use these instead of the real OpenG ones?

 

You would like to receive absolution to use the XNodes, despite all the well known comments out there.

 

The only people who can really do that, most likely won't as they are not allowed to do that and we can't other than the few who tried it. Working on them seems a rather crash intense affair, using them seems a bit more safe, but the mileage may vary greatly depending on LabVIEW version, OS and what else, including the position of the moon. What I can safely say is, that there is absolutely no guarantee, that XNodes will not work better or worse in future versions of LabVIEW. They may be improved, left to code rot that will cause more crashes in newer versions, or eventually discontinued entirely and removed from future LabVIEW releases.

 

As such I would consider it a totally irresponsible decision to use them for anything but private experiments.

[gb119]

I would add "in-house code which the developer is in full control of" to the class "private experiments" but otherwise not disagree with Rolf. That said, they work for me and I trust my $200k cryostat to them, but you wouldn't find me trusting a heart monitor to them !It appears NI would like to have a mechanism to let users write generic, type adapting code, but they have stated here that XNodes will not be the solution.

[hooovahh]

Not the answer I was hoping for, but I'm not surprised.  As tempting as it may be I will try to stay away from XNodes, and stick with 100 polymorphic instances, and OpenG code that can't adapt to every thing I need.

[gb119]

I guess what one could (in theory and I might think about trying at some stage) do is to have an XNode that responds to a right click menu option by replacing itself with a static subvi that includes the same code - thus one could have the benefits of XNode type adaptation during development but with the ability to convert to static vis when preparing a production version of code.