On 10/05/2018 21:18, bartc wrote:
On 10/05/2018 19:51, Chris Angelico wrote:
On Fri, May 11, 2018 at 4:31 AM, bartc <b...@freeuk.com> wrote:
2x100 (4) Binary
3x100 (9) Ternary
4x100 (16) Quaternary
5x100 (25) etc
6x100 (36)
7x100 (49)
8x100 (64) Octal
9x100 (81)
... (Not implemented 11x to 15x, nor 10x or 16x)
0x100 (256) Hex
YAGNI much? How often do you need a base-9 literal in your code??
I've just found out these also work for floating point. So that:
a := 8x100.5
print a
gives 64.625 in decimal (not 64.5 as I expected, because .5 is 5/8 not
5/10!). Exponent values are octal too, scaling by powers of 8.
I tried it in Python 3 (0o100.5 - I find that prefix fiddly to type
actually as I have to stop and think), and it seems to be illegal.
Based floating point literals may be unusual, but bear in mind that in
decimal, some values may not be represented exactly (eg 0.1). I believe
that in base 2, 4, 8 or 16, any floating point literal can be
represented exactly, at least up the precision available.
--
bartc
--
https://mail.python.org/mailman/listinfo/python-list
