smith@umn-cs.UUCP (06/06/83)
#R:ihnss:-151400:umn-cs:6900010:000:546 umn-cs!smith Apr 26 09:39:00 1983 I used to work for a company that did speech recognition work, and our primary debugging tool for these dreadful programs was a scope driven by an A/D converter. All of your favorite bugs (overflow, underflow, etc) showed up as spikes or other weird effects that contrasted against the otherwise recognizable waveforms. I suspect this trick was first used on the Whirlwind I computer at MIT, possibly in the late 1940s. There's a photo of them using a scope driven by the computer in the 1951 IRE (nee IEEE) conference preceedings. Rick.