Feedback Thread: WPILib

[Nadav Zingerman]

I haven't seen any LabVIEW feedback yet, so I guess I'll be the first.

• The RefNum Registry - I think that this is a bad idea. While it does simplify development by reducing the number of advanced concepts programmers need to know (TypeDefs mostly), it also violates key programming best-practices, such as not using globals unnecessarily. It also discourages a logical grouping of RefNums, such as putting all RefNums for a sub-system in the same cluster.
• Many VIs in the library are not sufficiently documneted (are they even ready for use?) such as the DMA and Interrupt VIs.
• It would be nice if there were support for additional features of joysticks and gamepads, such as tactile feedback (many gamepads have a vibration feature).
• Beta screen shots showed an "Experiment Framework" in addition to the Robot Framework. It would be nice it this were made available again (I'm assuming this is something similar to 2009's Basic Framework) for testing sensors and the like. Right now we open example files and modify them for this purpose.

I don't know where bug reports belong, so I'll report here:

• The DBL Array version of the polymorphic PID block does not work out of the box. It seems to be looking for some DLL. A workaround is to use the LabVIEW implementation by setting the "PIDimplementation" symbol to "G", but this probably hurts performance, and regerdless, not many teams will know to do this.
• The PID Autotuning VI is broken and fails at the autotuning proccess, also searching for a missing file.
• The PID palette exist twice. Once as in the main palette list and once under Control Design & Simulation. The latter contains more functions, some of them quite useful, so it's a shame to have them hidden away in a submenu. Besides, duplication is evil.

Overall, great job with WPILib