Add an easy way to provide total ordering now that __cmp__ is deprecated/gone

[mrts]

BPO	5479
Nosy	@loewis, @birkenfeld, @rhettinger, @mdickinson, @merwok, @florentx

^{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 = 'https://github.com/rhettinger'
closed_at = <Date 2010-08-08.01:19:51.952>
created_at = <Date 2009-03-12.08:18:35.996>
labels = ['type-feature', 'library', 'docs']
title = 'Add an easy way to provide total ordering now that __cmp__ is deprecated/gone'
updated_at = <Date 2010-08-08.01:19:51.951>
user = 'https://bugs.python.org/mrts'

bugs.python.org fields:

activity = <Date 2010-08-08.01:19:51.951>
actor = 'rhettinger'
assignee = 'rhettinger'
closed = True
closed_date = <Date 2010-08-08.01:19:51.952>
closer = 'rhettinger'
components = ['Documentation', 'Library (Lib)']
creation = <Date 2009-03-12.08:18:35.996>
creator = 'mrts'
dependencies = []
files = []
hgrepos = []
issue_num = 5479
keywords = []
message_count = 15.0
messages = ['83490', '83491', '83499', '83501', '83502', '83510', '83528', '83574', '102315', '102400', '102401', '102402', '102403', '102412', '113228']
nosy_count = 8.0
nosy_names = ['loewis', 'georg.brandl', 'rhettinger', 'mark.dickinson', 'LambertDW', 'eric.araujo', 'mrts', 'flox']
pr_nums = []
priority = 'normal'
resolution = 'accepted'
stage = None
status = 'closed'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue5479'
versions = ['Python 2.7', 'Python 3.2']

[mrts]

See http://mail.python.org/pipermail/python-dev/2009-March/087000.html
and http://code.activestate.com/recipes/576685/ .

[lambertdw]

That's the best version I recall seeing at activestate.

Still, I'd deprecate and remove > and >= from mathematics.

[loewis]

What is the purpose of this submission? What do you want to happen about
Python?

[lambertdw]

(As I recall) in python-dev mailing list Ray claimed he could clean up a
cited active state recipe to address this issue. He succeeded to the
extent I'm aware---he's the author of
http://code.activestate.com/recipes/576685/.

I haven't used the redundant >, >= comparisons operators in code since
1981. The chances of simplifying python by removing them from the
language are None, but the mistake predates python by 10**n years, where
(3 <= n <= 7) or (7 >= n >= 3).

[loewis]

Then why was this classified as a documentation issue? And why did Mart
Sömmermaa submit it, and not Raymond? AFAICT, Raymond said he would
propose something when it's ready (which I assume it currently isn't).

[mrts]

Then why was this classified as a documentation issue?

As the documentation section of
http://docs.python.org/reference/datamodel.html#object.\_\_lt__ needs to
be updated as well to mark the eventual solution as the recommended easy
way to provide total ordering.

And why did Mart Sömmermaa submit it, and not Raymond? AFAICT, Raymond
said he would propose something when it's ready (which I assume it
currently isn't).

Raymond's recipe at http://code.activestate.com/recipes/576685/ looks
more or less complete, do you feel that his posting on the mailing list
does not count as proposal? I submitted the feature request instead of
him because I was the one who noticed the problem (as discussed on the
mailing list) and felt "responsible" to report it here.

[loewis]

As the documentation section of
http://docs.python.org/reference/datamodel.html#object.\_\_lt__ needs to
be updated as well to mark the eventual solution as the recommended easy
way to provide total ordering.

This is the 2.6 version. What about the 3.0 version in

http://docs.python.org/3.0/reference/datamodel.html#object.\_\_lt__

needs to be updated?

Raymond's recipe at http://code.activestate.com/recipes/576685/ looks
more or less complete, do you feel that his posting on the mailing list
does not count as proposal? I submitted the feature request instead of
him because I was the one who noticed the problem (as discussed on the
mailing list) and felt "responsible" to report it here.

I see. So it's a feature request.

[mrts]

This is the 2.6 version. What about the 3.0 version in
http://docs.python.org/3.0/reference/datamodel.html#object.\_\_lt__
needs to be updated?

When functools.total_ordering (whether it lands in functools is open)
lands that section should be amended in the lines of the following:

"There are no implied relationships among the comparison operators. The
truth of x==y does not imply that x!=y is false. Accordingly, when
defining __eq__(), one should also define __ne__() so that the operators
will behave as expected.

However, given a class defining one or more ordering methods,
functools.total_ordering_ class decorator can be used to fill in the
rest. Please see the documentation of functools.total_ordering_ for
further details."

[rhettinger]

See r79730.

Leaving open until tests are written and all is forward ported to 3.2.

[merwok]

Hello

Small documentation question: Does the expression “total ordering” have established usage in maths or computer science? Its meaning is not obvious to the non-maths person that I am.

Regards

[mdickinson]

Yes, it's a standard mathematics term.

http://en.wikipedia.org/wiki/Total_order

[rhettinger]

http://en.wikipedia.org/wiki/Total_order

[merwok]

Thanks for the link. Please include it in the future doc if you judge it useful for a large number of users.

I’m still wondering if “total_ordering” is the best name for a decorator that fills the blanks to provide total ordering.

Regards

[lambertdw]

http://en.wikipedia.org/wiki/Total_order

For pair of items from a set, (that's the total)
if a <= b and b <= c then a <= c (part of the order)
if a <= b and b <= a then a compares the same as b, a == b, (the
other part of the order)

On Mon, 2010-04-05 at 19:46 +0000, Éric Araujo wrote:

Éric Araujo <merwok@netwok.org> added the comment:

Hello

Small documentation question: Does the expression “total ordering” have established usage in maths or computer science? Its meaning is not obvious to the non-maths person that I am.

Regards

----------

---

Python tracker <report@bugs.python.org>
<http://bugs.python.org/issue5479\>

---

[rhettinger]

Applied in r79810
Thanks for the suggestion.