Make decimal floating point be default, remove binary floating point

[Retro]

BPO	7389
Nosy	@gvanrossum, @mdickinson

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2009-11-24.11:24:32.961>
created_at = <Date 2009-11-24.11:05:58.305>
labels = ['interpreter-core', 'type-feature']
title = 'Make decimal floating point be default, remove binary floating point'
updated_at = <Date 2009-11-24.11:24:32.793>
user = 'https://bugs.python.org/Retro'

bugs.python.org fields:

activity = <Date 2009-11-24.11:24:32.793>
actor = 'mark.dickinson'
assignee = 'none'
closed = True
closed_date = <Date 2009-11-24.11:24:32.961>
closer = 'mark.dickinson'
components = ['Interpreter Core']
creation = <Date 2009-11-24.11:05:58.305>
creator = 'Retro'
dependencies = []
files = []
hgrepos = []
issue_num = 7389
keywords = []
message_count = 2.0
messages = ['95668', '95669']
nosy_count = 3.0
nosy_names = ['gvanrossum', 'mark.dickinson', 'Retro']
pr_nums = []
priority = 'normal'
resolution = 'later'
stage = None
status = 'closed'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue7389'
versions = ['Python 3.2']

[Retro]

The Python interpreter should have the decimal type built into its core.
The mechanism for dealing with decimal numbers should be handled by the
decimal type and not by the processor which spits binary floats. The
time is now. But ask yourself these questions:

Is this doable? If yes:
Would it break anything? If yes, note that:
Python 3.1 opposed to Python 3.0 also had major changes that break the
two appart.

I say go for revolutional change in Python 3.2. But the call is yours.

[mdickinson]

I think the bug tracker is the wrong place to discuss such a wide-
ranging and (currently) ill-specified change. The python-list or
python-ideas mailing lists might be better places.

Decimal in the core is out of the question for Python 3.2, thanks to PEP-3003. What *is* feasible at this stage is to replace the current slow
Python implementation of the decimal module with a C version; there's
some ongoing work on that.

Removing binary floats from the core entirely sounds like a bad idea to
me, for performance reasons. And it probably couldn't be done before
Python 4.x without breaking backwards compatibility.

I'm closing this for now; if you're interested in this, please take the
discussion to python-list or python-ideas.