In the course of my research, I have worked extensively with a variety of instrumentation linked to computers for data acquisition and instrument control using the IEEE 488/488.2 standard interface (a.k.a. the General Purpose Interface Bus [GPIB] or Hewlett-Packard Interface Bus [HPIB]). There are a variety of programming environments available to interact with the necessary device drivers to achieve this, from simple BASIC interface libraries compatible with most BASIC compilers to visual programming languages such as LabView and HP-VEE which can be used to generate complex "virtual test laboratories".
When I started work in the Chemical Sensors Group at the University of Toronto, I found that although LabView 1.0 allowed me to very rapidly create and debug data acquisition software, it did not have the interface tools to allow the level of data analysis required. After experiencing growing frustration with transfering the resulting large text files from application to application, I got frustrated and ended up writing my own integrated data collection and analysis program in C. Since then, both LabView and data analysis software (as well as the hardware that runs both) have improved dramatically so that I might well not follow the same course today (funds permitting). Then again, my software has improved enormously after ten years, too!

Lots of goodies, multiple simultaneous measurements from different instruments, test functions, variable delay.....

Data search and display, transformation, scaling, graph overlays, expansion and rescaling, baseline subtraction....

Did I mention custom curve fitting? Not so easy to do in most spreadsheet programs for the types of functions relevant to chemical sensors, although it is possible to program such fits in data exploration packages like MathCad, Matlab, Origin, etc. (But then there's the whole data import/export issue again...)

Individual peak analysis too (not to mention digital filtering). As you can see, the experimental situation imposes a large number of requirements beyond simple trigger/read/record operations. There is always more than one way to solve a particular data acquisition/analysis problem: custom software is one of them.