Is the new GIL in Python 3.2 sufficient to make the switch?

Question

I was reading this page on the new GIL found/to be found in Python 3.2 and I was wondering if it is the "killer feature" that will trigger a transition from Python 2.x to 3.x.

What do you guys think?

Answer 1

3.2 is roughly a year away. By that time hopefully there will be some progress on switching to py3k, it also coincides with the two year "grace" period that was supposedly promised by Guido to some big projects. I don't think that there will be any single "killer" feature, and even if anyone would need it, I'd imagine unicode support would be much higher on priority ladder.

Answer 2

The New GIL could be considered a killer feature of Python 3.2 if you use threading heavily. Integration of unladen swallow could also be considered a killer feature from a performance perspective.

In general, I think py3k already has enough features to warrant switching; what it doesn't have is broad library support. Most libraries have already started looking at the transition, but it is a good ways away. By the time 3.2 comes out a year from now, I expect most of the libraries I depend on to have made strong progress with compatibility, and that is what would make me switch.

Answer 3

It's still a GIL. Python implementations without any GIL at all have been available for over a decade now. Python implementations that are much faster than CPython have been available for years.

(Almost) noone uses them, which clearly shows that nobody cares about performance or the GIL, so I hardly see them being a "killer feature". Killer features are something that people actually want.