Message230579
FWIW, I think doc changes like this need to be made by the people most familiar with the implementation details and with the original design intent (i.e. the authors of PEP 207).
The knowledge here is fragile and it would be easy to accidentally make-up new normative rules that aren't actually true or necessary, or to subtly misrepresent what was trying to be accomplished originally. The risk is greatest when the OP is just now learning the ins and outs of NotImplemented (i.e. whether in-place operations can or should return NotImplemeted or whether those operations can be duck typed).
Changing "can be returned" to "should be returned" is normative (making-up a new rule without sufficient discussion). This doesn't add explanation; instead, instead, it make a rule change about how a feature should be used.
https://hg.python.org/cpython/rev/26d0a17affb5 |
|
Date |
User |
Action |
Args |
2014-11-04 05:41:27 | rhettinger | unlink | issue22780 messages |
2014-11-04 05:38:17 | rhettinger | set | recipients:
+ rhettinger, gvanrossum, tim.peters, georg.brandl, ezio.melotti, eric.araujo, r.david.murray, ethan.furman, python-dev |
2014-11-04 05:38:17 | rhettinger | set | messageid: <1415079497.68.0.279190620958.issue22780@psf.upfronthosting.co.za> |
2014-11-04 05:38:17 | rhettinger | link | issue22780 messages |
2014-11-04 05:38:16 | rhettinger | create | |
|