Speak for yourself. After using Python 3, I can't stand touching Python 2 codebases. Is a given str object an actual text string that's been decoded into a valid charset, or is it an array of bytes fresh from the network / database / file? Who knows, unless you trace back to its point of origin. I personally love that Python 3 says "this is text, and this is some binary data I got somewhere, and they are not the same thing".

Please no! Text is text. Bytes is bytes. Convert to the correct form on input boundaries, and that's it. Don't switch back and forth internally, it's overly complicated, error prone, and slower.

unix pipes (stdin, stdout) are bytes, files are bytes, filenames are bytes. yet, for some reason python3 thinks al of those are text. its not the coders that are wrong, it is the language.

No, it's Python 2 that thinks those are all text. Python 3 makes you explicitly say "convert this stream of bytes into a text encoded string", and from then on it's a str object.

Python 2 was happy to (try to) let you call text methods on a JPEG. Python 3 draws an appropriate distinction between the two.

Python 2's "str" is a bytestring. All of those are bytestrings in Python 2 as they should be. Python 3 makes them text and suddenly i can not output data over stdout anymore without extra steps of switching the mode of stdout from some arbitrary text default.

kstrauer is right, but I'll elaborate and say that places where you are interfacing with Unix threads are input boundaries, so having the programmer make a choice makes sense. Arbitrary switching within a program does not.

the issue is that the language arbitrarily decides those things should be treated as text with some arbitrary encoding when all of them are decidedly not text.