Encodeing of compressed intervals
On Thu, 26 Jun 2014 21:23:39 -0400 I had some typos (bad cut&paste):
> atoff("nan(8)") 7F80 0008 7FC4 8000
> atoff("nan(16)") 7F80 0010 7FC2 4000
This should have been:
atoff("nan(8)") 7F80 0008 7FC4 0000
atoff("nan(16)") 7F80 0010 7FC2 0000
(I had also mis-spelled PA-RISC as PA-RISK -- no slight intended!)
Finally, about those messy-looking DPD encodings: if the decorations
are encoded with quantum 1 (as would be usual for small integers), the
hex dump would actually look quite familiar as decimal numbers. That's
because 0 through 79 are encoded such that the bottom two hexits of the
hexadecimal reprentation can be interpreted as if it were decimal:
DFP32(79) = 0x22500079 (in DPD encoding)
(Normalized to full precision looks ugly: DFP32(79.00000) = 0x3E02300)
So much for vicarious entertainment...
Michel.
---Sent: 2014-06-27 03:32:28 UTC