[cfp-interest 3909] Re: Trailing Decimal Exponent on a Hexadecimal standard floating type

[Joshua Cranmer]

On 4/21/2026 1:56, Paul Zimmermann wrote:
> I take the opportunity to ask whether one should standardize the way
> hexadecimal numbers are printed: 0x1.5abp+23 or 0x2.b56p+22 or
> 0x5.6acp+21 or 0xa.d58p+20. I prefer 0x1.5abp+23 since the magnitude
> is easier to guess.
I was working on a tool where my original criterion was "digit before 
the decimal point should be 1-radix", but then I found that also 
conflicted with my desire to indicate the exact number of digits in the 
significand, since not all the standard floating-point types line up 
nicely with hexadecimal digits with that rule. For my personal needs, I 
haven't entirely resolved which rule matters more to me.