Which is a better choice to data flow?

[ACarpio]

Hi everyone,

      just in case that is not possible to made flow data naturally, Which is a better option, use sequences or artificial data dependency, like a trigger signal?

thanks for your answers!  

[RavensFan]

Can you give us an example of what you mean by trigger signal?

An external signal?  A notifier?  Something else?

[johnsold · Accepted solution]

Either can be used.  Usually a single frame sequence with an error wire or other data wire is sufficient. I cannot think of a case in many years where a multiple case sequence structure was needed.

If you use artificial data dependency, add a note to the diagram explaining why the unused signal is there.  Otherwise someone may remove it in the future because it appeared to be unnecessary and then the data dependency would be missing.

Another option, often a good one, is to create a subVI of the portion of the code which does not have a natural data dependency and add error clusters.  The subVI then allows the data dependency and may have the bonus of reducing the space required on the diagram.

Lynn 

[F._Schubert]

In addition to Lynn's post: using the error clusters allows you to make a case structure that the code is executed only if no error occured. Imagin a test sequence with several waits, you normally don't want to have all the waits executed on error. I was using my own 'delay.vi' encapsulating the wait primitive, until I started using OpenG.

Felix