Python 3000 Status Update. Doesn’t look like we’ll get multiline lambdas, but the other stuff looks great. I’m not looking forward to years of Python 2 and Python 3 co-existing and splitting the community though (ala PHP 4 and 5).
Sign in with 
OpenID
I don't think you need to worry too much. Python 2 is not nearly as broken as PHP4, and the delta in goodness between Python 2 -> Python 3 will not be nearly as huge as PHP4 -> PHP5. It's weird, but the improvement in PHP5 is holding people back in PHP4.
One of the reasons that PHP4 continues to flourish is that none of the major projects that use PHP will take the bold leap to commit to PHP5 only. If e.g. WordPress in a future version converts to PHP5 that will drag all the ISPs and end users along with it. PHP5 would dominate a year later.
But why would WordPress do that? PHP4 is very different than PHP5, and the WordPress code is pretty much architected for PHP4. You could say PHP4 is in its DNA. A big part of that is because PHP4 object capabilities are so woefully awful. Rewriting WordPress to really be PHP5-savvy would be a massive undertaking, for (what appears to be) very little (if any) gain.
The one small ray of hope for PHP5 right now is the Zend Framework. It's PHP5-only, and if developers really start to adopt it, then PHP5 will be dragged along into the light. A lot depends on how the ZF turns out, documentation, sample code, and all the other things you need to do right to get developers to adopt your platform.
Python today doesn't suffer from the same level of limitation that PHP4 does. It will probably be much simpler for major existing projects to adopt Python 3 features, and in such a way that it's a reasonable trade of features (or code clarity, or ...) for compatibility. This in turn will allow developers to adopt it, without requiring the flying leap of faith that PHP4 -> PHP5 seems to take.