Harvey S. Liszt
NRAO, Charlottesville, VA
In December 1985, the NRAO ended its longterm use of IBM mainframes for most computing needs, in order better to support UNIX and AIPS. Some babies were thrown out with this bathwater and non-AIPS computing needs were severely impoverished after the fact. I saw two problems in attempting to use the heavily-burdened mini and super-mini computers which replaced the mainframe; recoding in FORTRAN both my own works and the many orphaned NRAO-written utilities on which I depended (all done in, you guessed it, PL/1), and lack of hardware support, as the Observatory allowed me only 1-2 MBytes of disk storage, on machines of near-legendary sluggishness. By contrast, an 8 MHZ 8086 IBM-XT clone with a 20 MByte disk and a Turbo Pascal compiler offered quicker response, more storage, and better graphics. I proceeded to become infatuated with this device, and spent the years 1986-1990 coding a family of programs which support analysis of singledish data and FITS images interactively. To validate my effort, I published this work with an easy-to-use interface, and it is now in fairly wide use in the singledish community. In 1991, I returned to research and now support the programs part-time, adding code as needed or requested, but with a very high priority on bug-fixing. Being on both sides of the fence taught me many lessons about how to create usuable and useful programs, which I hope to communicate to the AIPS++ project as its advisor on singledish programming and to the audience at ADASS96.