On Thursday, June 28, 2012 7:47:24 AM UTC+2, Stefan Behnel wrote: > Serhiy Storchaka, 28.06.2012 07:36: > > On 28.06.12 00:14, Terry Reedy wrote: > >> Another prediction: people who code Python without reading the manual, > >> at least not for new features, will learn about 'u' somehow (such as by > >> reading this list) and may do either of the following, both of which are > >> bad. > >> > >> 1. They will confuse themselves by thinking that 'u' actually means > >> somethings. They may then confuse others by writing about its supposed > >> meaning. This might get amusing. > >> > >> 2. They will use 'u' in Python 3 only code, thereby making it > >> incompatible with 3.2-, even if it otherwise would not be. > >> > >> These two actions will reinforce each other. > > > > Yes, this is what I mean. I can even make a prediction: in just 5 years, as > > this feature would be banned in a decent society. The authors of the books > > will be strongly advise not to use it, and in software companies 'u' will > > be prohibited in coding style. But get rid of this will be difficult. > > Once Py2.7 is out of maintenance, we can deprecate that feature in one > release and start warning about it in the next one. You're then free to use > the corresponding 2to3 fixer to get it back out of your code with a single > patch. > > Stefan
On the other side, one can argue this (elegancy): b'a serie of bytes' u'a unicode, a serie of code points' 'python2? str? python3? encoded _unicode?' jmf -- http://mail.python.org/mailman/listinfo/python-list
