On 3 May 2010 11:42:47 -0700, in bit.listserv.ibm-main you wrote: >On Mon, 3 May 2010 08:58:16 -0700, Ed Gould wrote: > >>This might be of interest to those wanting to do floating point arithmetic. >>Please *NOTE* I do NOT know if this pertains to IBM or not. >> >>http://floating-point-gui.de/ >> > >Common decimal numbers such as 0.1 can not be accurately >represented in binary. If you divide X'1' by X'A', you will get >X'0.19999999....'. Or try this with your favorite hexadecimal >calculator. Divide X'10000000' by X'A' and you will get >X'1999999'. Multiply that by X'A' and see what you get. >The windoze hex calculator gives X'FFFFFFA'. > >Decimal Floating Point doesn't have this problem. IIRC, DFP >was first implemented on the z9 and on the z10 it is >implemented entirely in hardware.
And the IBM COBOL people don't feel there is a need to implement it despite IBM having spent a bundle to help define it and then implement it in hardware. If they do, they intend to use the same USAGE most people would use to mean IEEE binary floating point. ARRGH ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
