Would it be possible to also include an IEEE bit-exact encoding for float and double types, in addition to the base-10 encodings in the recommendation? I mean, some binary or base64 version that matches the 4-byte IEEE standard for floats, and the 8-byte IEEE standard for doubles? Every conversion from binary to base-10 and back causes rounding errors... Perhaps the exact value could be encoded using binary or base64, and an "alt", human-readable value could be shown in base-10 for informational purposes only... Thanks, Peter Zingg CTO IMSI 75 Rowland Way Novato, CA 94945 tel:415-878-4000 fax:415-897-2544 mailto:pzingg@imsisoft.comReceived on Thursday, 2 November 2000 12:06:04 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 23 October 2007 06:13:19 GMT