EJ's points are well taken as usual. An inference from his position, one that he politely omits to draw himself, is clear.
There is little excuse for the continuing existence and use of RMODE(24) applications. They should long since have been converted to RMODE(31) [and, where necessary, AMODE(31)]. There is, of course, an obvious, if not perhaps finally very important, exception to this stricture: RMODE(24) applications that are 1) very small and 2) do no I/O at all may remain so. John Gilmore, Ashland, MA 01721 - USA
