12 years of Python and only at v2.2

[Greg Ewing]

Tim Peters wrote:

> [Greg Ewing]
> 
>>Hmmm. I wonder if there's some generalisation of the
>>notion of monotonicity when you're dealing with a
>>tree rather than a linear sequence?
>
> you don't need reals, you just need lexicographic ordering on
> tuples.


I was going to reply with something to that effect, when
I realised that there is a subtler point. While it's
certainly possible to arrange the version numbers in a
well-defined order, the versions in question were *not*
released in that order time-wise. In other words, the
function v = f(t) is not monotonic.

However, the version numbers have clearly been chosen
to try to make them monotonic with respect to some other
independent variable, which we might call "features
implemented" or some such.

But the values of *that* variable don't necessarily
have a total ordering, in which case the usual notion
of monotonicity doesn't apply. Hence my comment about
trees!

-- 
Greg Ewing, Computer Science Dept,
University of Canterbury,	
Christchurch, New Zealand
http://www.cosc.canterbury.ac.nz/~greg