When working with clusters and SubVIs, pass only the data the SubVI needs. Do not let the SubVIs consume the entire cluster.

Do NOT use variant attributes for data storage unless their high performance is definitely needed.

• Attributes hide away any information about the datatypes at edit time

Wrap VIs in classes (.lvclass) or libraries (.lvlib) for name spacing and generally improved encapsulation / cohesion / coupling

• See Organising Code for details

If a technology supports multiple threads (more than one instance in memory at the same time), all the VIs of the corresponding driver usually need to be set to reentrant execution.

Do not use “First Call?”, use an FGV instead

Use Compound Arithmetic instead of boolean logic gates

Do NOT use invert options on boolean functions and operators

• use a separate not operator instead

For constants that are reused throughout the project, create a subVI that only contains a single constant on the block diagram and one indicator returning the constant value

• append –constant to the VI name
• See 02 Layout & Design for VI icon design

Use “Use Default If Unwired” appropriately

If no default value should ever be used, enable “Linked Input Tunnel”

When using CD Structures, put them into the project tree:

• as far up as possible (ie. direct at the project root level, as opposed to the targets)
• as far down as necessary (eg to distinguish different FPGA targets)

For each CD Symbol in the project, add a log entry or similar to your project so you can identify at runtime which values were defined for the symbols

• Put these into a subVI directly after starting the application
• When using HSE libs, put this subVI directly after initialising the HSE-Logger

Transferring data from one loop to another loop (VI-local)

• Using VI server (reference or property node) to update the value of a control or indicator is the least efficient mechanism by far.
• Local variables come with only a slight overhead to directly writing to the terminal.
• Global variable would be the fastest with the additional caveats of race conditions and issues when global variables are not handled correctly.
• RT FIFO or a User Event add context and encapsulation.
• See the LabVIEW Craftsmen blog

Wrap each query in a separate SubVI

• this allows for easy reuse
• add documentation to describe the returned values
• optional: convert the returned 2d-array of strings into a 1d-array of typedef'ed cluster

The Array to Cluster node has the number of elements hard-coded

As an array is dynamic, we need to check the size during runtime

To convert the generic cluster to a typedef-ed one, use the Coerce To Type function