mike@BRL.ARPA (Mike Muuss) (04/25/88)
Just as a "heads up", I'd like you to know that I am currently writing an RFC that documents BRL's Standard for Network Floating Point Numbers. Because we chose to use the IEEE 754 representation, sent in Big-Endian order, I suspect that there should not be a whole lot of debate about such a standard. We also have library routines (in C) that convert between network order and local host order for hosts using IEEE, IBM, DEC VAX, and Cray formats, which we are placing into the public domain. The efficient transmission of floating point data is very significant to the scientific computing community, so I think the time is ripe for defining such a format. I hope to have this RFC to you in 2-4 weeks. Best, -Mike
ark@alice.UUCP (04/25/88)
In article <13098@brl-adm.ARPA>, mike@BRL.ARPA writes: > Because we chose to use the IEEE 754 representation, sent in Big-Endian > order, I suspect that there should not be a whole lot of debate about > such a standard. IEEE 754 says nothing about the sequence of bits within a word. Even a machine that stores all the even-numbered bits and then all the odd-numbered bits is permitted.