Feature request: sorting a list slice

Raymond Hettinger python at rcn.com
Sun May 21 18:55:51 CEST 2006


Getting a patch ready for checkin (corrected, documented, reviewed, and
tested) is only part of the battle.  The real challenge of language
design is figuring out whether something should be done.

Adding optional parameters to a method makes its invocation slower and
makes the API harder to learn and remember.

There has not yet been a thorough discussion of use cases.  In
particular, people should scan their existing, deployed codebase for
examples of where a particular code fragment would have been improved.
We should then compare the fragment with and without the feature -- how
much more concise is it, how much is performance improved/harmed, is
the code more or less readable, etc.

If the perf gain is small and the use cases are infrequent, the
addition is likely unwarranted.   There is an entire class of feature
requests that are more appropriate as recipes than for inclusion in the
language.  If python had accepted most requests like this, it would
have become somewhat bloated and unattractive.




More information about the Python-list mailing list