Re: why still use C? [long]

To add to Francis Glassborow's excellent reply:

> [4] Would the application needs be met by a simpler, more direct
> approach? If rounding is a problem, for example, could not logic
> be added to current (FP-2,FP-16) FPUs to produce the desired
> behaviors?

No, because it is a mathematical impossibility to produce the
desired behavior (other than by using the BFP numbers as
integers, and handling the exponent separately).

See examples at google: decimal arithmetic FAQ

One could, possibly, build a dual unit which can switch the
base for calculation -- but it would be unlikely to be an
optimal design (for example, for BFP it's best if the
coefficient is a binary integer; for decimal it's better to have
a decimal coefficient (e.g., in BCD).

However, it is certainly possible to share the FP register
file, and that is a significant saving.

Mike Cowlishaw
Google: decimal arithmetic
-- 
comp.lang.c.moderated - moderation address: [EMAIL PROTECTED]