Clear Linux Is The Latest Distribution Figuring Out What To Do With Python 2

[Bitter]

Oh please stop being so politically correct, not to mention idiotic. "It's 2018" is not a sane argument. It's just a reflex for people who unquestioningly accept everything and don't actually have the capacity to form their own opinion. Therefore "oh it's 2018, everything in 2018 must be correct compared with everything that came before". Completely stupid reasoning.

What I meant, and what you clearly don't want to acknowledge, is that a this point you have to handle the global environment, and Unicode must be a top priority.

Nobody is saying Unicode is not good, only that for 99% of "crossing the boundary" applications, for which Python is supposed to excel ("glue language, remember") we're still using bytes. So having strings suddenly become unicode every time you're talking bytes to msgpack, zeromq, or any of dozens of other packages, you're having to tear your hear out encoding/decoding and wondering which you'll get back. Go's UTF8 handily solves the problem. Unicode where you want it, bytes otherwise. Predictable, easy, no f*cking around with encode decode.

Sounds to me like it's just your limitation to handle such concepts.

[vegabook]

No... I'm saying Python 3, like a new car, has a lot of desirable (but optional) features and overall low-level improvements than previous generations.

So despite there basically being no major differences that will negatively affect you, you'd rather add fragmentation to the industry and stick with Python 2. Remind me to avoid any applications you have written, because good programming requires good logic skills, and your logic is severely flawed.

Stop lying to yourself. Python 3 does not do anything that Python 2 doesn't already do fine. YOUR logic must be severely flawed since you cannot understand that that's why it took 10 years! Nobody needs Python 3! 10 years later and transition is still an issue. How do you explain that, Einstein?

Oh an if you think that the transition is done, let me remind you of the title of the story:

"Clear Linux is the Latest Distributioin Figuring Out What To Do With Python 2"

[schmidtbag]

Stop lying to yourself. Python 3 does not do anything that Python 2 doesn't already do fine. YOUR logic must be severely flawed since you cannot understand that that's why it took 10 years! Nobody needs Python 3! 10 years later and transition is still an issue. How do you explain that, Einstein?

Seriously how are you so blind to not realize the inverse of your statements apply: Python 2 doesn't do anything better than Python 3.
Python 3 has been plenty usable for far fewer than 10 years. I've been using it since the 3.2 days, when I realized "wait a minute, why am I still using Python 2, since it has no inherent advantages?"
Meanwhile, had it not occurred to you that the transition is still a problem because of people like you? This is why Windows XP (and in turn, older versions of Internet Explorer) were crippling the industry - it's because people like you think "well what I use is working fine for me and I don't care about holding back anyone else because it's not my problem".

Oh an if you think that the transition is done, let me remind you of the title of the story:

"Clear Linux is the Latest Distributioin Figuring Out What To Do With Python 2"

Right, which in case you haven't noticed, is not exactly speaking fondly of Python 2; it's implied to be a burden.

[vegabook]

What I meant, and what you clearly don't want to acknowledge, is that a this point you have to handle the global environment, and Unicode must be a top priority.

Why exactly "must" Unicode be a top priority? The world didn't globalize suddenly in 2009. The internet has been a big deal since 1990. Python 2 does Unicode just fine where you need it. And UTF8 does it even better if you need it everywhere. Python 3's Unicode-by-default is a poorly thought-through, politically motivated dogma that has a lot more downsides than upsides.

[vegabook]

Seriously how are you so blind to not realize the inverse of your statements apply: Python 2 doesn't do anything better than Python 3.
Python 3 has been plenty usable for far fewer than 10 years. I've been using it since the 3.2 days, when I realized "wait a minute, why am I still using Python 2, since it has no inherent advantages?"
Meanwhile, had it not occurred to you that the transition is still a problem because of people like you? This is why Windows XP (and in turn, older versions of Internet Explorer) were crippling the industry - it's because people like you think "well what I use is working fine for me and I don't care about holding back anyone else because it's not my problem".


Right, which in case you haven't noticed, is not exactly speaking fondly of Python 2; it's implied to be a burden.

Unfortunately, Einstein, "people like me" are everywhere. That's why you are still using the x86 instruction set and that MS-DOS can still run on an Core i7. It is almost a law of the computer industry universe that "you don't make breaking changes" unless you are providing MASSIVELY more value. You see, that's what you don't get. Python 3 is not good enough to get people to move. People don't move when what they're using is good enough. That's how it is, and you Python 3 guys have ignored that at your peril.

The Python 3 transition, as I am sure you are aware, is used as a case study by ecosystem designers on how NOT to do things, so it is disengenuous and indeed foolish to blame things on "people like me".

This thing has been a complete fuckup, you know it, I know it, and the angst in these forums about it is the rock solid proof of that fact.

[schmidtbag]

Unfortunately, Einstein, "people like me" are everywhere. That's why you are still using the x86 instruction set and that MS-DOS can still run on an Core i7. It is almost a law of the computer industry universe that "you don't make breaking changes" unless you are providing MASSIVELY more value. You see, that's what you don't get. Python 3 is not good enough to get people to move. People don't move when what they're using is good enough. That's how it is, and you Python 3 guys have ignored that at your peril.

Yes, it is quite unfortunate - thanks for holding back the world. Unlike x86 (which was inherently designed to evolve over time, hence being CISC), your methods aren't helping anyone. It doesn't matter if Python 3 "massively" provides more value - the fact of the matter is, you willingly are preventing progress. And for what, your petty principles? Because you're too lazy or fussy?
Ironic how you're basically admitting that your stubbornness is actively slowing down progress, simply because "python 2 is good enough". Throughout this debate, you have consistently shown that you ignore the needs of anyone but yourself.
How exactly are you contributing anything of value?

is used as a case study by ecosystem designers on how NOT to do things, so it is disengenuous and indeed foolish to blame things on "people like me".

Facts and sources, please. Let's not waste time on conspiracies and opinions.

This thing has been a complete fuckup, you know it, I know it, and the angst in these forums about it is the rock solid proof of that fact.

Except, despite how many times I ask you to prove your point, you have yet to show any evidence of an actual fuckup beyond your personal preferences.

So as far as I'm concerned, the only thing fucking up Python 3 are people like you.

[TheBlackCat]

Nobody is saying Unicode is not good, only that for 99% of "crossing the boundary" applications, for which Python is supposed to excel ("glue language, remember") we're still using bytes. So having strings suddenly become unicode every time you're talking bytes to msgpack, zeromq, or any of dozens of other packages, you're having to tear your hear out encoding/decoding and wondering which you'll get back. Go's UTF8 handily solves the problem. Unicode where you want it, bytes otherwise. Predictable, easy, no f*cking around with encode decode.

It is a really simpler concept: text and bytes are not the same thing. Text is text, bytes are arbitrary binary data that can represent anything. "In the face of ambiguity, refuse the temptation to guess." Python 2 ignored this by guessing that arbitrary binary data was supposed to represent text. Python3 fixes this issue by making text and arbitrary binary data two different things.

In Python 2, if you want Python to interpret bytes as integers, you need to explicitly convert it. If you want it to interpret bytes as floats, you need to explicitly convert it. If you want it to interpret bytes as some packed data, you need to explicitly convert it. If you want it to interpret bytes as text, well it just sort of guesses you want one of the dozens of available kind of text and does it on its own. Python 3 fixes this inconsistency by requiring you to be explicit about what the binary data is supposed to represent in every situation, rather than special-casing one. "Explicit is better than implicit." "Special cases aren't special enough to break the rules."

This led to tons of hard-to-find bugs that can and did occur in pretty much any situation where you actually needed to explicitly treat bytes as text, bugs that were only found and fixed once people had to be explicit about what they wanted. The Zen of Python are not just words, they are good policies for designing a programming language. Python 3 fixed a violation of this rule (several, actually) that was causing a lot of trouble for people.

[TheBlackCat]

This thing has been a complete fuckup, you know it, I know it, and the angst in these forums about it is the rock solid proof of that fact.

What angst? We have a maybe two or three people complaining about it with vague conspiracy theories and three or four times that many happy about it and hundreds more that don't care one way or the other.

[vegabook]

It is a really simpler concept: text and bytes are not the same thing. Text is text, bytes are arbitrary binary data that can represent anything. "In the face of ambiguity, refuse the temptation to guess." Python 2 ignored this by guessing that arbitrary binary data was supposed to represent text. Python3 fixes this issue by making text and arbitrary binary data two different things.

In Python 2, if you want Python to interpret bytes as integers, you need to explicitly convert it. If you want it to interpret bytes as floats, you need to explicitly convert it. If you want it to interpret bytes as some packed data, you need to explicitly convert it. If you want it to interpret bytes as text, well it just sort of guesses you want one of the dozens of available kind of text and does it on its own. Python 3 fixes this inconsistency by requiring you to be explicit about what the binary data is supposed to represent in every situation, rather than special-casing one. "Explicit is better than implicit." "Special cases aren't special enough to break the rules."

This led to tons of hard-to-find bugs that can and did occur in pretty much any situation where you actually needed to explicitly treat bytes as text, bugs that were only found and fixed once people had to be explicit about what they wanted. The Zen of Python are not just words, they are good policies for designing a programming language. Python 3 fixed a violation of this rule (several, actually) that was causing a lot of trouble for people.

Oh please. Thanks for pedantically mansplaining that. The issue is that 99% of the time bytes by default would be better.

[TheBlackCat]

This entire post is basically telling me that Python 3 is just Python 2 with a few small syntax changes. Yes I agree!! That's why it was a completely unnecessary endeavour for which the only recourse of the zealots is bullying. Python 3, as you are basically saying, is a complete waste of time, 'cos it's Python 2 + a few extraneous things (that nobody uses).

What a a huge waste of everybody's time this Python 3 thing has been.

Just because you personally don't use something doesn't mean nobody does. I use a lot of Python 3 features, to such an extent that I simply gave up on trying to make things backwards-compatible. Projects aren't abandoning support for python 2, or not supporting it to begin with, for lols, there are useful features that only Python 3 supports and having to keep Python 2 support is holding them back.