In 1988, at the second Computer Supported Cooperative Work conference, I presented a paper discussing why software designed to support groups was generally failing. It was intended to be of immediate assistance to my fellow software engineers who in those days outnumbered academics at the conference. It just received the first CSCW Lasting Impact Award, to be presented at CSCW 2014 in February. I will describe how it came to be written, who helped along the way, the combination of timing and good fortune that led to its influence, and which findings still seem useful and which less so. The experience suggests some guidance that diverges a bit from conventional advice.