As I recall, I was "rewriting from scratch" a 1000-VI totally-undocumented LabVIEW Real-Time routine used to study sound localizations in medium-sized behaving (or mis-behaving, at times) primates of the BME student variety. I had two State machines running, one on the Host that basically ran the UI, which open an Excel Workbook and ran the various test sequences (which involved moving a speaker mounted on a 2-DOF robotic arm that positioned the speaker 2 meters from the "test subject" with radial arm that could be positioned from -30 to +30 degrees of azimuth and at a vertical height corresponding (if I remember correctly) -20 to +20 elevation with respect to the subject's ears. The Robotic Arm was driven by an Industrial Controller (Ladder Logic, anyone?), but was mostly used to position the speaker at various positions.
A PXI system (also running a State Machine) handled the delivery of the various sound patterns, the "light" cues (the subject was in an unlit anechoic chamber, with a laser target fixation spot that "went out" when he/she was supposed to point a laser (which became illuminated) at the perceived Target location, with results saved in the Excel Workbook.
During my rewrite, I wanted to "Track" the two State Machines and have some "structure" to the record to help me figure out "bugs in the code" (I must confess there were a few). I hit on XML to write a record each time a State was entered, noting whether from Host or Target, the time (based on a millisecond clock started when the Experiment started), and any "parameter" passed into the State.
This code was heavily used for about 5 years, but the lab has been shut down for slightly longer than that, so I don't remember all the details. But I was clearly writing an "Array of State Clusters", but wanted to do this one cluster at a time. I started with EasyXML, and figured out (Jim may have helped me here) and managed to write the Array one element at a time, but it was awkward.
Then I discovered someone at NI had put on the Tools Network something called NI-GXML, which included the Source code. So I "extended" the functionality and added a few additional functions and had what I wanted -- a system where I would open an XML output file, write a bunch of stuff, open an Array of Clusters, write them one-at-a-time, close the Array, write a few more things, then close the XML file. Worked fine.
I've used it a few times more for logging/debugging, but sometimes a plain text file created by Write Delimited Spreadsheet works just as well for simple logging-during-development ...
Bob Schor
