Asking for prophecy

[Terry Reedy]

Stefan Spoettl wrote:
> In the pass it was always a good idea to use the newest Python verison 
> for starting the development of a new application. First one could 
> benefit from the additional features and second one could be sure that 
> the community would have been passing during development.
>  
> Nowadays we have at least three Python versions (2.5, 2.6, 3.0) on our 
> machines and - damned! - I really don’t know which version I should use 
> for my next development. The Unix-like systems as much as the major part 
> of well maintained third party libraries are remaining "penetrantly" on 
> 2.5. Why the vangard of the community don’t like to use at least 2.6 for 
> bridging to the future Python? Is this the mutiny against the empery of 
> the BDFL or is the vangard just asking for some more time? If I want to 
> attest my personal attachment to the king by using 3.0, what will 
> happen? Will I be deserted someday?

It typically takes about a year before *most* 3-rd party libs have been 
converted.  Windows binaries for C extentions tend to be slowest to 
arrive.  Many people routinely skip 2.x.0 and wait to 2.x.1 both to skip 
initial bugs and wait for libraries.  I have no idea if 2.6 conversions 
are slower than usual or not.

3.0 conversion was expected to be a bit slower.  On the other hand, it 
is an opportunity to increase mindshare for libs which do make the 
conversion.

I expect usage of Py3 will increase noticeably when 3.1 comes out in a 
few months with some notable fixes and improvements.

I think there is at least half a chance that 2.7, which should arrive 
with 3.2, will be the last 2.x version.

In the meanwhile, use whichever one meets your needs.  I am currently 
using 3.0, but have 2.5 loaded in case I want to do something that needs 
3rd-party libs before they are available for 3.x.

Terry Jan Reedy