'Scaled', 'scaling', and their cognates have a long history in computing. Packed-decimal arithmetic results are scaled programmatically in compiled COBOL code. In assembly language, on the other hand, all packed-decimal arithmetic is integer arithmetic. The programmer who uses it for fixed-point computations must scale outcomes himself. He must, that is, keep track of the decimal points in results himself. The same things are tgrue mutatis mutandis for binary points in binary-fixed arithmetic.
Fletcher Jones, the founder of the Computer Science Corporation and its leader when it was an important technical force in the industry, had the habit of delivering a well-honed exposition of what he called "the lost art of scaling" with little or no encouragement. (I did of course mean 2^10 = 1024 and10^3 = 1000.) John Gilmore, Ashland, MA 01721 - USA ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
