I do not expect significant use of DFP in existing COBOL systems. The COBOL community is a highly conservative one that comes to terms with new technology only very slowly or not at all, witness the recent discussion here of the wide continuing use of AMODE(24).
I did think it useful to turn Frank's question around. What can be done at all well in packed decimal is a very small proper subset of what can be done very well indeed in DFP, and this needs to be understood. There is no technical rationale for any further use pf packed-decimal arithmetic in any context, although the data type itself does and will continue to have its uses in conversions from computational to display values asnd vice versa. These things said, inertia and reactionary managements will be barriers to significant use of DFP even in new applications. It is always possible to find that there is no economic justification for a technical innovation, to make a persuasive case of this sort to those who anyway want to hear it. I do expect that vendors and ISVs will make much use of DFP under the covers, where the advantages it confers can be exploited without disturbing customers. This indeed is already the case with one very celebrated RDBM, which has used BFP (sic) instead of packed decimal internally for many years but does of course externalize values as packed-decimal ones on demand. 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
