Python becomes the most popular language

[Picuino]

It is a good moment to teach Computer Science, with Python as first programming language, in Secondary School.

[bd139]

They do that here 

[Cerebus]

Oh great, a whole generation who think that indentation = semantics. 

[bd139]

I think it’s more a generation who can solve problems without getting pissy 

[pcprogrammer]

And also with a lot of patience.

I'm indifferent on Python, but compared to C it is slow. For the FPGA reversing I'm doing the code for the open source tools is written in Python and I used it to do some preliminary work, but boy it takes it time when you need to process millions of bits. So I switched back to C in a heartbeat.

[Cerebus]

Not getting pissy, taking the piss possibly, but not getting pissy. 

I really don't think that Python is a good starting place as an instructional language, too many idiosyncrasies, way too many ways of doing things and it's quite hard to reveal how to do things in "layers" as students grasp concepts one by one.

The older 'instructional' languages are probably badly suited to educating people for the future - they were all designed with building up to using full second or third generation programming languages. That I think would poorly serve current students. Sure, they need the concepts that second and third generation languages use, but the older 'instructional' languages completely omit features that are de rigueur for modern languages.  Perhaps it's time for a new 'instructional' language.

The "indentation as semantics" is a serious point however. It's a terrible feature for an instructional language because (1) only 1 popular language works that way resulting in (2) crashing and burning when you hit a language that doesn't work that way. All of us who have habitually indented languages with classical block semantics 'get' the Python indentation thing immediately. Some of us may not like it, but we get it. If all you knew was the Python "indentation as semantics" approach where would you be when faced with a language that doesn't work that way? Very confused for quite a while I suspect.

[bd139]

It’s a bit of a straw man that. It’s not an instructional language. You can use it as one if you choose.

I mean it could have semantic braces instead like the C-like languages. That left us in the domain of non memory safety C, bonkers JavaScript, general fucking shit show PHP and industrial scale misadventure and the main reason we’ve never been visited by an alien race C++ . Those are, well someone more mired with idiosyncrasies as I think I alluded to.

l ask you a question; have you ever done a 3-way merge on 7,000 Java files? Then you’ll know why braces aren’t necessarily the best idea out there. The test suite checks semantic errors in python. In Java the compiler twangs your ball sack for an hour asking you to inject squiggles not because it needs to be told what is semantic and what isn’t.

The real problems in the world are far greater than semantic whitespace which might not be a bad idea after all.

[SiliconWizard]

It is a good moment to teach Computer Science, with Python as first programming language, in Secondary School.

In which case, that should be definitely not be called Computer Science, but rather Programming courses. And I'll be fine with that.

[jfiresto]

... If all you knew was the Python "indentation as semantics" approach where would you be when faced with a language that doesn't work that way? Very confused for quite a while I suspect.

I might believe that except Python's syntax is pragmatic not dogmatic. Real Python code often relaxes the "off-side rule".

[Cerebus]

It’s a bit of a straw man that. It’s not an instructional language. You can use it as one if you choose.

Erm, no it's not a strawman because I never said that it was an instructional language. I merely opined that to use it in place of an instructional language was probably not a great idea.

There's quite a body of experience from the world of education that suggests that specifically designed beginner instruction languages are a good thing^TM. It's instructive (pun unintentional, but I'll take credit for it) that Smalltalk came of the the Learning Research Group at Xerox' Palo Alto Research Labs from an effort to make a programming language that could be taught at infant school level.

They actually succeeded in teaching 6 year olds to program using it. Remember that's about when you start to learn to read and write the language that you've grown up with, language that will be parsed with a whole human brain for syntax and semantics, not the pathetic fussy brain of a literal computer. It's quite an achievement to have got six year old to write any code that actually works.

I mean it could have semantic braces instead like the C-like languages. That left us in the domain of non memory safety C, bonkers JavaScript, general fucking shit show PHP and industrial scale misadventure and the main reason we’ve never been visited by an alien race C++ . Those are, well someone more mired with idiosyncrasies as I think I alluded to.

l ask you a question; have you ever done a 3-way merge on 7,000 Java files? Then you’ll know why braces aren’t necessarily the best idea out there. The test suite checks semantic errors in python. In Java the compiler twangs your ball sack for an hour asking you to inject squiggles not because it needs to be told what is semantic and what isn’t.

The real problems in the world are far greater than semantic whitespace which might not be a bad idea after all.

You seem hung up on braces. Bad experience with orthodontistry?  A litany of the failures of assorted programming languages is really neither here nor there, nor is getting yourself into the situation where you even need to contemplate the idea of doing a three-way merge of 7,000 source files**.  I'd opine that none of them are particularly good, and especially they're none of them particularly good as an introduction to programming. Far, far, too many introductory courses to programming are introductory courses to programming in language x rather than courses in programming per se.

Of relevance is my own introduction to programming. It was at a one week residential summer school in 1974 as a test subject for the "Computer Education in Schools" syllabus that ICL cooked up at the behest of the government of the day. They produced their own programming language 'CESIL*' which was 'assembler made easy' which they taught in the first half of the week, and they taught BASIC in the second half of the week. I spent the next year helping one of the math's teachers at school teach that course, taught myself FORTRAN and COBOL, and took a commercial job as a junior programmer in 1976 where I was bashing out COBOL for a software house that serviced the Merchant Banks of the day. So that approach clearly had some merit as far as the computer industry of the time was concerned.

The question is what would make a good introductory language now? It certainly isn't assembler-lite like CESIL was, but neither is it a language like Python that throws in everything including the kitchen sink and has idiosyncrasies that make it non-representative of programming languages as a whole.

* Computer Education in Schools Introductory Language

** To paraphrase Lady Bracknell "To have to do a three-way merge of one source file is unfortunate; to have to merge 7,000 of them smacks of carelessness."

[nctnico]

IMHO Pascal is still pretty good at forcing people to think about the structure of a program before starting to code. Python is just scribbles on a napkin that you code in a real programming language later on.

[olkipukki]

I'm indifferent on Python, but compared to C it is slow. For the FPGA reversing I'm doing the code for the open source tools is written in Python and I used it to do some preliminary work, but boy it takes it time when you need to process millions of bits. So I switched back to C in a heartbeat.

It's getting better 

https://docs.python.org/3.11/whatsnew/3.11.html#faster-cpython

[emece67]

.

[Karel]

CPython 3.11 is on average 25% faster than CPython 3.10 when measured with the pyperformance benchmark suite,
and compiled with GCC on Ubuntu Linux. Depending on your workload, the speedup could be up to 10-60% faster.

Which means still 20 times slower than C/C++...

[voltsandjolts]

CPython 3.11 is on average 25% faster than CPython 3.10 when measured with the pyperformance benchmark suite,
and compiled with GCC on Ubuntu Linux. Depending on your workload, the speedup could be up to 10-60% faster.

Which means still 20 times slower than C/C++...

...but with 100x faster development cycle than C/C++.

That means Python is great for initial dev/experiments and perhaps production if it's fast "enough" or the critical bits are written in C/C++. But, yawn, this is age old argument repeated a thousand times here already, I must be bored at work or I wouldn't even have posted this. I guess Python helped me finish my work early

[madires]

We have to wait a few years to see how this Python hype evolves. Kids start learning Python in school now. Will Windows 13 be written in Python? But please don't complain if the greybeards will remind you later 'we told you so'.

[pcprogrammer]

...but with 100x faster development cycle than C/C++.

But only when you are up to speed with Python

And when you are experienced enough with both it might not make a difference in which language you write.

[coppice]

IMHO Pascal is still pretty good at forcing people to think about the structure of a program before starting to code.

Pascal, or derivatives like Turbo Pascal? Turbo Pascal makes you focus on structure. Real Pascal makes you focus on working around all the basic things it doesn't have, like word length control and logical operations.

[brucehoult]

...but with 100x faster development cycle than C/C++.

But only when you are up to speed with Python

... and not up to speed with the C++ standard library, which contains all that useful stuff such as expandable arrays and trees and hash tables and algorithms such as sorting and so forth.

It's possible Java and separately JavaScript might lie somewhere between the development time of C++ and Python -- Java closer to C++, and JavaScript closer to Python, and bot run maybe 20% slower than C++, not 20x slower.

And when you are experienced enough with both it might not make a difference in which language you write.

I would think if you know both well, the difference would be no more than 2x, and maybe quite a bit less than that.

C++ compilation time is not significant on reasonable sized programs on modern computers. Any difference is purely in convenience of libraries and the amount of typing needed.

[Karel]

It's possible Java and separately JavaScript might lie somewhere between the development time of C++ and Python -- Java closer to C++, and JavaScript closer to Python, and bot run maybe 20% slower than C++, not 20x slower.

I don't agree. I ported a library from pure C to pure Python and for sure it's 20 times slower.
The lib provides an api to write ADC samples to a file in a certain format and it also applies some scaling to all samples.
Other people wrote a Python wrapper around the C library because in pure Python it was terribly slow.
Both those people and me did the speed comparison and the speed difference is huge (at least 20x).

There's a reason why Python (and Numpy as well) is written in pure C...

[olkipukki]

I ported a library from pure C to pure Python and for sure it's 20 times slower.

Not sure why you did this  , but overall - a terrible idea, at least cover in cython with a minimalistic porting and let it's available to Python world

[SiliconWizard]

We have to wait a few years to see how this Python hype evolves. Kids start learning Python in school now. Will Windows 13 be written in Python? But please don't complain if the greybeards will remind you later 'we told you so'.

I for one think that Python has too many "flaws" and is knowing too much "hype" not to eventually get phased out.
When and how, I do not know. It will have to be replaced by something else of similar nature - a language easy to use for a very wide range of people including a lot of non-software developers.

I ported a library from pure C to pure Python and for sure it's 20 times slower.

Not sure why you did this  ,

Me neither, but the only "valid" reason I see is that Karel would have been missioned to do this by his boss, or a customer. Doing that from your own will doesn't seem very plausible, so that must come from some order you can't decline.

[Picuino]

This is an electronics forum, which is one of the areas where Python is not as widely used due to its speed. This is very noticeable in the comments, always obsessed with speed, as if speed were the only parameter of importance.

[nctnico]

We have to wait a few years to see how this Python hype evolves. Kids start learning Python in school now. Will Windows 13 be written in Python? But please don't complain if the greybeards will remind you later 'we told you so'.

I for one think that Python has too many "flaws" and is knowing too much "hype" not to eventually get phased out.
When and how, I do not know. It will have to be replaced by something else of similar nature - a language easy to use for a very wide range of people including a lot of non-software developers.

I agree. For starters such a language will need to be governed by some form of a standard committee to ensure prolonged compatibility where it comes to the language syntax and libraries. Python is too much of a moving target to be useful for commercial software development.

[voltsandjolts]

Python is too much of a moving target to be useful for commercial software development.

I don't think that is true at all.

There are many big tech firms successfully using Python, I won't list them here as they are so easy to find. Indeed that is a big part of the momentum Python has, no doubt many contributions coming from there. However, I think it tends to be more server side, web tech stuff rather than desktop apps. Dropbox is the only hugely popular desktop app I can think of.