From: erewhon@nowhere.uk   
      
   On Fri, 29 Apr 2005 11:42:33 GMT, "Raoul Watson"   
    wrote:   
      
   >   
   >"J French"  wrote in message   
   >news:4270a3a2.265496530@news.btclick.com...   
   >   
   >> I am afraid I don't fully agree with that   
   >> - the problem is down to the Institute of Electrical & Electronic   
   >> Engineers   
   >>   
   >> The problem is that IEEE is simply 'faulty'   
   >> - some numbers can only be stored as approximations   
   >>   
   >   
   >Agree with you a 100% on that buddy. MS implemention is my issue..   
   >   
   >Ever since Kahan brought his graduate students, the 754 floating point   
   >standard became a faulty committee . Only one representative from the   
   >computing field was there. When the standard was ratified, exactly two   
   >companies produced floating point arithmetic chips that attached to   
   >microprocessor CPU chips. Intel sold the 8087 and National Semiconductor   
   >sold a chip for the NS32 processors.   
   >   
   >The supposedly target audience had ignored it for over 20 years. It has   
   >never been used by Cray, makers of the premier scientific computers. It has   
   >never been used by IBM mainframes, which at the time did a great deal of   
   >scientific computing. It was never used by CDC. It was never used by DEC. No   
   >RISC microprocessor has ever implemented IEEE 754 floating point. Since the   
   >RISC processors were always by far the leaders in floating point   
   >performance, the fact that they for twenty years never implemented the   
   >standard is telling. (They generally use the IEEE format, and there is   
   >usually a software-fixup technique that allows them to claim complete   
   >conformance!) No one uses them in that mode, however. It's way too slow.   
   >   
   >In the early 80's, when Apple developped SANE (Standard Apple Numeric   
   >Environment) it supports all the requirements of the IEEE standard and yet   
   >went beyond it to "shield" the users from the deficiencies by developing   
   >libraries for high quality engineering, scientific, financial and accounting   
   >applications.   
   >   
   >MS could have done the same thing or stick to the BCD and improve on it.   
   >   
   >To have a sleuth of non-working functions and come back to the end user by   
   >explaining that's because of the IEEE standard and how numbers are   
   >manipukated internally is a pisspoor excuse.   
      
   Thanks for the info   
   I guess that SANE holds numbers internally in its own format, and only   
   allows the problems to occur when they hit the disk   
      
   I must confess that in the late 1980's we managed to improve on the MS   
   IEEE conversion routines that were in BASIC 7.1   
      
   The VB5 conversion routines are slightly less accurate than those in   
   B7.1 - which gave me a shock when porting code.   
      
   However, I still reckon that the IEEE standard is faulty   
   - the crazy thing is that BCD appears to be built into the   
   'co-processor'   
      
   Personally I would have added a new data type   
   IEEE would be a 'sort of optional extra' for Import/Export   
      
   Following stupid standards - is ... well stupid    
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)