This is just a short note to explain the plan for future release
numbering. Basically, we are just going to be much more rigorous
about following everyone's expectations about release numbers.
1.1.0 will be released ASAP and will include some minor API breakage
and new features
1.1.x will be a purely bug-fix release
1.2.0 will be released by the end of the summer and will include some
minor API breakage and new features
1.2.x will be a purely bug-fix release
At this point we will probably have a longer period of code
stabilization in the 1.2.x release series before starting to work on a
1.3 release. I also hope to get a SciPy 0.7 release out soon based on
NumPy 1.1 and then we will release a SciPy 0.8 based on NumPy 1.2.
When/if we release NumPy 2.0, we will allow much more radically API
breakage. The idea being that a major release might require people
'rewriting' code where a minor release would require a small amount of
refactoring. Bugfixes to a release shouldn't require any changes to
code depending on our software. Just for reference:
I.e., we are in the 1.x major release cycle and we are making two new
minor releases this summer (1.1 ASAP and 1.2 by the end of summer).
We will release bug-fixes as necessary. For example, we could release
1.0.5 if anyone feels there are some important bug-fixes they wish to
backport from the trunk to the 1.0 branch (which I will be happy to
make if anyone so wishes).