[Python-Dev] Complexity documentation request

Terry Reedy tjreedy at udel.edu
Tue Mar 11 04:11:16 CET 2008

"Greg Ewing" <greg.ewing at canterbury.ac.nz> wrote in message 
news:47D5C2C4.6090701 at canterbury.ac.nz...
|| Yeah, there's no substitute for having at least a
| rough idea of how the object you're using is implemented,
| e.g. array vs. linked list.

As I understand it, the new 2.6 docs include a new one on CPython 
specifically.  A page there  might be appropriate.  But someone has to 
write and submit a patch for review.

| This kind of very basic information is something that
| I think ought to be documented, and some guarantees
| made in the language definition. For example, I think
| a Python implementation that implemented lists as
| linked lists would make many people unhappy, as their
| algorithms suddenly went from O(n**m) to O(n**(m+1))
| without anyone telling them.

Such an implementation should document such a design decision, but I don't 
see that an interpreter that runs the test suite should be prohibited from 
calling itself a 'Python interpreter'


More information about the Python-Dev mailing list