Yes, the startup VI should be your main VI. This could be your GUI VI, or perhaps a loader VI that initializes and starts up your main VI. Basicly, whatever VI should run first. You can, as you noticed, have more than one startup VI, though most applications don't require it.
Regular subVIs that are on the block diagram of your VI don't need to be included as dynamic VIs, because LabVIEW already knows to include them. You only need to include a VI explicitly as a Dynamic VI if you are calling dynamically it with VI Server calls (like Invoke Nodes) rather than as a regular subVI. The reason is that this VI is only being referenced by its path, so until run-time, LabVIEW has no idea whatsoever what VI to load. Consequently it can't guess which VIs to include as dynamic VIs when it builds an executable.
Here's a trick that makes it much easier to organize your dynamic VIs and have them included automatically in the build process. In your LabVIEW Project, create a virtual folder and call it something like My Dynamic VIs. The name isn't important, just make sure it makes sense to you. Then, whenever you have a VI that you will call dynamically in your application using Invoke Nodes or Call by Reference Nodes, add that VI to the virtual folder you created.
Now, when you go to create a build specification, add your one main VI as the startup Vi, but then add the
whole My Dynamic VIs virtual folder under the Dynamic VIs and Support Files category. When you select build, whatever happens to be in the virtual folder at that time gets added as a Dynamic Vi in the build. And if you add more dynamic Vis into that folder at a later time in your project and you want to rebuild, you don't need to change the Build Specification at all. Everything in that folder will always get added to the build automatically!
Jarrod S.
National Instruments
