On Tuesday, June 28, 2016 at 11:12:59 AM UTC+5:30, Steven D'Aprano wrote: > On Tuesday 28 June 2016 14:31, Rustom Mody wrote: > > > On Tuesday, June 28, 2016 at 6:36:06 AM UTC+5:30, Steven D'Aprano wrote: > >> On Tue, 28 Jun 2016 12:23 am, Rustom Mody wrote: > >> > Also how is GG deliberately downgrading clear unicode content to be kind > >> > to obsolete clients at recipient end different from python 2 → 3 making > >> > breaking changes but not going beyond ASCII lexemes? > >> > >> Oh yes, I completely agree, obviously GvR is literally worse than Hitler > >> because he hasn't added a bunch of Unicode characters with poor support for > >> input and worse support for output as essential syntactic elements to > >> Python. > >> > >> /s > > > > Gratuitous Godwin acceleration produceth poor sarcasm -- try again > > And while you are at it try and answer the parallel: > > Unicode has a major pro and con > > Pro: Its a superset and enormously richer than ASCII > > Correct. > > > Con: It is costly and implementations are spotty > > That's a matter of opinion. What do you mean by "spotty"?
We've had this conversation before. Ive listed these spottinesses See http://blog.languager.org/2015/03/whimsical-unicode.html Specifically the section on ½-assed unicode support > > It seems to me that implementations are mostly pretty good, at least as good > as > Python 2 narrow builds. Support for astral characters is not as good, but > (apart from some Han users, and a few specialist niches) not as import either. > > The big problem is poor tooling: fonts still have many missing characters, > and > editors don't make it easy to enter anything not visible on the keyboard. > > > > GG downgrades posts containing unicode if it can, thereby increasing reach > > to > > recipients with unicode-broken clients > > And how does that encourage clients to support Unicode? It just enables > developers to tell themselves "It's just a few weirdos and foreigners who use > Unicode, ASCII [by which they mean Latin 1] is good enough for everyone." > > Its 2016, and it is *way* past time that application developers stop > pandering > to legacy encodings by making them the default. If developers saw that 99% of > emails were UTF-8, they would be less likely to think they could avoid > learning > about Unicode. > > > > Likewise this: > > > >> a bunch of Unicode characters with poor support for > >> input and worse support for output as essential syntactic elements to > >> Python. > > > > sounds like the same logic applied to python > > > > JFTR I am not quarrelling with Guido's choices; just pointing out your > > inconsistencies > > Oh, it's inconsistencies plural is it? So I have more than one? :-) Here's one (below) > > In Python 3, source files are treated as UTF-8 by default. That means, if you > want to use Unicode characters in your source code (for variable names, > comments, or in strings) you can, and you don't have to declare a special > encoding. Just save the file in an editor that defaults to UTF-8, and Python > is > satisfied. If, for some reason, you need some legacy encoding, you can still > explicitly set it with a coding cookie at the top of the file. > > That behaviour is exactly analogous to my position that mail and news clients > should default to UTF-8. But in neither case would people be *required* to > include Unicode characters in their text. Python2 had strings and unicode strings u"..." Python3 has char-strings and byte-strings b"..." with the char-strings uniformly spanning all of unicode. Not just a significant change in implementation but in mindset Yet the way you use Unicode in the sentence above implies that while you *say* 'Unicode' you mean the set Unicode - ASCII which is exactly Python2 mindset. So which mindset do you subscribe to? -- https://mail.python.org/mailman/listinfo/python-list