Twenty five years ago, in august 1976, a specialist workshop gathered in Cornell University, at John Silcox's invitation, scientists covering a broad spectrum of interests to assess the potential of analytical electron microscopy, as to instrumentation, fundamental limits, general level of data processing and current theoretical understanding. If the STEM instrument had already been existing for a few years in Crewe's laboratory, its major emphasis, in spite of the existence of an EELS spectrometer, had focused on new modes of high resolution imaging. At the first Cornell workshop, the STEM instrument was for the first time recognized as a potentially formidable analytical instrument because of the possibility of extracting all available signals simultaneously. Furthermore it was directly suitable for digital computer processing and therefore for providing quantitative information. It was also pointed out that a major advantage of the STEM would be its potential to record EELS spectra from every point in the field of view as one scans an area to form an image, thus offering the capability of "chemical" mapping beyond "elemental" mapping.