I think Mike's writeup is a very good intro:
http://www2.hursley.ibm.com/decimal/
Birger Heede
IBM, GBS Denmark
Paul Gilmartin wrote:
On Wed, 9 Aug 2006 01:26:20 +0000, john gilmore wrote:
In the end pandering to clots, however compelling the arguments of the
marketing VP for doing so may seem to be, is always ill-advised.
In celebration of the unveiling of Decimal Floating Point, I'll return
to this, which I've been pondering for weeks.
Is Decimal FP mostly the product of wishful thinking that the magnitude
range of floating point can be combined with the accounting precision
of packed decimal? Has a marketing VP subscribed to a clot's belief that
it will enable him to take the price of a dozen eggs in dollars; convert
to euros, then shekels, then the price of a single egg in yuan, then
back to a dozen in dollars and be assured of getting the original
number exactly? Not so.
I glanced at the document and, in fairness, it provides numerous rounding
conventions that an accountant might wish to use. Will this be productive?
Will the control of rounding be the aspect more valuable than floating
point per se?
When will COBOL and PL/I support DFP?
--gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
