Chris McDonough <chrism <at> plope.com> writes:
Although the change makes it possible, and it is very useful for very low level WSGI apps, the issue this change addresses really isn't really 100% about "needing to define native strings". It's also just preservation of a resource in pretty short supply: developer energy.
Apparently developer energy is a limitless resource when it comes to arguing over PEPs ;-)
This, IMO, can only be a good thing, because the nominal impact of some
It can also have some downsides, at least according to some points of view. For example, I regard elevating "native strings" to undue prominence, and the continued use of u'xxx' in Python 3 code, as unfortunate consequences. For example, with PEP 414, it will be possible to mix Unicode with and without prefix - how would that not be at least a little confusing for users new to Python? Remember, "native strings" are a Python-only concept.
future user who must now understand u'' syntax is (again IMO) not as consequential as that user having less software to choose from because porting to Python 3 was just that much harder for existing Python 2 developers.
I don't believe it's because porting to Python 3 is especially hard. I'm not saying it's trivial, but it isn't rocket surgery ;-) Even if porting were trivially easy to do technically at the level the PEP addresses, there would still be additional tests, and perhaps documentation, and perhaps release-related work to be done. Since Python 2.x is a very good platform for software development, where's the incentive to move over to 3.x? It's the chicken and egg effect. Many people are waiting for other people to move over (perhaps projects they depend upon), and while the transition is happening, it's not as quick as it could be. I think a lot of it is down to inertia. Possibly another factor was the "just use 2to3 message", which we now know doesn't work well in all scenarios. However, I don't believe that the "use a single codebase, use six or six-like techniques, use unicode_literals, use the 2to3 fixer to remove unicode prefixes, and use native string markers where you need to" message has received anything like the same level of airplay. If you talk to people who have *actually tried* this approach (say Barry, or me) you'll hear that it's not been all that rough a ride. Regards, Vinay Sajip