Python is becoming the world’s most popular programming language

https://www.economist.com/graphic-detail/2018/07/26/python-is-becoming-the-worlds-most-popular-coding-language

by miohtama • 8 years ago

192 243 8 years ago

economist.com

miohtama• 8 years ago

Non-paywalled summary http://www.blog.pythonlibrary.org/2018/10/09/python-worlds-m...

Dowwie• 8 years ago

Hey Mikko!

accrual• 8 years ago

Non-canvas fingerprinting summary: http://www.blog.pythonlibrary.org/wp-content/uploads/2018/10...

I found the data points coarse (x = 5 years) but useful to see similar trends of popularity between C, Javascript, and Python. Note: these metrics are defined by global search engine popularity.

noobermin• 8 years ago

What happened to lisp in the last five years?

usrusr• 8 years ago

That drop looks like what I would expect if you do not include Clojure in the lisp numbers, despite not having a separate category for it.

tralarpa2• 8 years ago

The online documentation of Racket became too good :)

empath75• 8 years ago

Doesn’t seem plausible that both c and c++ are more popular than python.

yellowapple2• 8 years ago

Why not? C++ is still the de facto standard for desktop software (including games), last I checked, and C is still the de facto standard for systems and embedded programming. Python hasn't really displaced anything in either realm as far as I can tell (some C/C++ codebases use Python for user-customizable scripting or plugins, sure, but it probably ain't doing the heavy lifting).

denisw• 8 years ago

That entirely depends on how “popular” is defined. If you mean “preferred choice for greenfield consumer and B2B apps”, then yes, C/C++ would be ranked way lower. But as far as I understand this statistic is based in search results, so it takes into account the huge body of existing software that needs maintenance (after all, you need to stay up-to-date and also look things up again from time to time), as well as domains in which C and C++ are a good still dominating, such as programming embedded devices.

yellowapple2• 8 years ago

I wonder if those statistics are accounting for people searching for snakes.

sumedh• 8 years ago

Took me a while to figure out what you mean by snakes, I thought what do snakes have to do with programming languages then it bit me.

tralarpa2• 8 years ago

Or for coffee.

bakery2k3• 8 years ago

Or, of course, for British surreal comedy groups.

Swizec• 8 years ago

Wait wasn’t JavaScript supposed to be most popular? Every device in the world shipping with at least one environment for JS?

Or does it not count because you don’t distribute binaries?

martinsbalodis• 8 years ago

I was thinking the same thing. Their google search language comparison is probably incorrect. For example no one includes the keyword "javascript" when searching for nodejs or react. Instead they should have used Stackoverflow's developer survey results.

From an article that analyzes Stackoverflow's developer survey results[1]:

  1. JavaScript
  2. HTML
  3. CSS
  4. SQL
  5. Java
  6. Bash/Shell
  7. Python
  8. C#
  9. PHP
  10. C++

[1] https://www.businessinsider.com/14-most-popular-programming-...

realusername• 8 years ago

Yes that list sounds much more realistic to me.

cm21872• 8 years ago

But nodejs and react aren’t programming languages. No one is including “c#” when searching for spreadsheetgear, asp.net or wpf even though they are .net libraries and frameworks.

cimmanom• 8 years ago

Yeah, JavaScript also has too many names for the language itself. I mean, I search for “Django” or “Python 3” instead of “Python” half the time. But with JS, if I’m looking for info about the language itself it’s usually with the keyword “ES6” instead of “JavaScript”.

cyborgx7• 8 years ago

Well, this ranks them by search engine popularity. That metric certainly has its problems, but I have to assume it does correlate with people coding in those languages more than it correlates with people using things coded in that language.

jorisw• 8 years ago

I don't think so.

When I have a React or Vue or Angular problem, I don't search for "JavaScript".

Only when I need docs on JavaScript's native features, for instance Array.sort(), I include "JavaScript". And even then I may use "JS" instead, but I assume Google knows that's a synonym for "JavaScript".

While each of those searches would be part of, say, a React project, I would definitely consider that part of the JavaScript space.

To be fair, all of the above would also go for Python vs Django, etc.

goatlover3• 8 years ago

Every device in the world shipping with at least one environment for JS

I doubt that's true, while it would be more true to say that most devices in the world is running something compiled from C or C++. Also, The JS environment (web or Node) is running on top of C++ code.

qualsiasi• 8 years ago

Python is really widespread, but at least in my company it's used for analysis and prototyping only. Then the prototype is ported either to Java, C# or C++.

I don't know if it's common practice, but we work like that and totally makes sense to us.

metildaa2• 8 years ago

Depends on the organizations needs, for what we work on performance does not matter terribly much, thus a MVP in Flask is often cleaned up and put up on Gitlab (with a Github mirror) once its vaugely usable.

qalmakka8• 8 years ago

Languages have always come and gone; Python has always mattered, it's simply finally showing itself as a much more suited tool for certain tasks (such as CS education, lightweight scripting by data scientists, glueing together large frameworks, ...). While popular, it can't really hit every single spot required in software development; try to maintain > 10k LOC codebases in Python to understand what I mean.

The point that I'm trying to say is that, while we're living in a moment where python is at the center of everything, it won't last forever; it never does. Something better always shows up, given enough time.

pmiller22• 8 years ago

I’ve worked on Python codebases >100k LOC with no issues. What actually is your point?

toolslive3• 8 years ago

having been on the receiving end of a large python project, I can say that refactoring is (trying to stay polite here) a challenge.

qalmakka8• 8 years ago

Like someone else said, I personally find that very large codebases written in dynamically typed languages tend to discourage refactoring. You may try to mitigate the issue by writing the whole application as standalone, independent modules, but we all know that a bit of interdependence always creeps inside even the best designed application.

I love Python; that's the first language I've learnt and I can't really live without its REPL. I only think that the industry has a tendency to excessively obsess towards certain languages, even when they are not the best tool for the job. For instance, I strongly believe that the academia's morbid obsession with Java has crippled a whole generation of new programmers, which are now unable to think and code in no other way but the Java™ Way. I won't say that python is as bad as Java in this regard (it isn't even close), but it annoys me to death when it looks like we're heading once again towards a language monoculture; at least this time the language isn't utter crap.

mherrmann• 8 years ago

I'm very happily maintaining a 15 kLoC code base (my file manager [1]). No issues at all. Though it must be said I'm alone, so know the code base inside out.

My point: Likely, there is a project size where Python's dynamic nature becomes a challenge. But 10 kLoC isn't it.

[1] https://fman.io

qalmakka8• 8 years ago

Writing code ≠ maintaining code someone else has written. Duck typed code relies in general in implicit behavioural assumption on how the values you're receiving should behave, and it's much better to be the one making choices than the once trying to understand them.

I'm not arguing that it's not possible to do it; I just think that it requires a bit of extra steps and care to document which assumptions your code takes and how it behaves that could be more meaningfully described by a statically typed language. But these are just my 2 cents.

scarface748• 8 years ago

Of course it’s possible if you write every line of code and you know it inside an out, but that’s very much not the norm.

pistolpeteDK• 8 years ago

Not trying to split hairs, but we have no problem maintaining our 50K LOC python/django codebase.

What I like about Python is the ease of which you can layout a project to fit your needs. Maintaining a large codebase is more about layout/architecture than the language itself.

Maybe other languages are simply better suited for programmers who lack the need of keeping their codebase clean and tidy :)

sumofzeros3• 8 years ago

50K LOC python/django codebase.

How much of it is generated?

WaxProlix2• 8 years ago

Similarly, we have a ~125k loc java service and a 47k loc one and they're both real, honest burdens to maintain.

ramijames• 8 years ago

This is totally true. We are trying to maintain a codebase that is hundreds of thousands of lines of code, based on Drupal 5. It's a fucking nightmare.

scarface748• 8 years ago

When it comes to refactoring, there are so many automated, guaranteed safe refactors you can do with statically typed language that you can’t do with dynamic language.

I can’t understand why developers would want to take a tool out of their tool belt like static checking.

I’m a fan of Python as a low overhead scripting language but for large systems with multiple developers, static type checking is a godsend.

devilmoon• 8 years ago

I'd prefer maintaining a 10k LoC python project rather than a 10k LoC Java project (which would likely be 8k lines of boilerplate anyway)

senorjazz5• 8 years ago

horses for courses. I would prefer a 100k java LOC program over a 10k loc python program, any day.

The ease and safety you can make huge refactors is a huge win. Sure you can do it in python, but you will be writing x10 the amount of tests and final LOC counts including tests will probably have python come in closer to Java than the 10k > 100k loc indicates

Flip-per• 8 years ago

I'm contributing to a >20k LOC open-source python/django project (counting .py files only). It's quite clean with an architecture that makes sense, and thus way easier to understand and refactor than an actually smaller Java project at work. Admittedly, I have more of a Python background than a Java background.

Jian-Yang• 8 years ago

Java is just as bad as python. For maintainability, header files make life a lot easier.

gaius2• 8 years ago

try to maintain > 10k LOC codebases in Python to understand what I mean.

My previous job we had about 3 million lines of Python and it was fine.

devxpy12• 8 years ago

I think your real problem is with duck typing and not specifically with python.

There are efforts to have gradual typing suppprt in python though, and you might be interested in having at least partial typing suppprt.

qalmakka8• 8 years ago

Yup, it's a complaint that can be applied to every duck typed language out there, not just Python.

devxpy12• 8 years ago

it won't last forever; it never does

I would love for a real contender to arrive, that challenges python's level of abstraction in a real way...

alehander422• 8 years ago

wow, what's python's level of abstraction? there are languages with much better abstraction capabilities

firdak4• 8 years ago

Do you think a statically-typed language could ever be a "contender" - i.e. compete with Python in terms of ease-of-use etc?

Nowadays, I don't think any new, dynamically-typed language has a chance to become mainstream.

stevesimmons4• 8 years ago

try to maintain > 10k LOC codebases in Python to understand what I mean

Lots of companies successfully manage far larger Python codebases.

We have a 35M LOC Python codebase with 2500 active developers making 25000 commits per week and continuous deployment globally of new code.

(I gave a talk on it at PyData London in April: "Python at Massive Scale")

mcintyre1994• 8 years ago

If anyone else is looking: https://www.youtube.com/watch?v=ZYD9yyMh9Hk :) I'm definitely interested in this - I've seen a struggle to manage a far smaller codebase than that in Python and really interested to see what you're doing differently to make it work.

collyw4• 8 years ago

Is there a quick way to find out your LOC in a python project? Does it include whitespace and comments? What about HTML and JS code?

simonh3• 8 years ago

I used to work on the subsequent brain child of Kirat Singh et al at another big financial institution, and ditto. Admittedly we used a custom built source control system and release tooling, but also developed in Python.

jernfrost4• 8 years ago

Is this made as a cluster of independent programs or is this actually a monolith, where you can access all functionality from some python method or class?

patagonia• 8 years ago

Perhaps paradoxically it is easier to maintain a gigantic Python code base than it is to maintain one that is small or medium sized.

The reasoning, a gigantic code base means infrastructure and tooling investment will obviously pay off. It probably even means you get one or more dedicated resources of this sort. Whereas a small to medium sized codebase might bump into some of the same issues as a larger code base but not justify such investments.

gigatexal2• 8 years ago

Dropbox has 1m lines of python — it’s doable.

qalmakka8• 8 years ago

Dropbox has had several headaches from its reliance on python, first among all from the fact that Python is still bound on running once thread at once on a single core after all these years (thanks GIL), so it does not parallelize at all without abusing subprocess + IPC.

That's the reason why they attempted writing a new runtime first (Pyston), and then settled down on rewriting their performance critical infrastructure in Go (https://blogs.dropbox.com/tech/2014/07/open-sourcing-our-go-...)

Google iself wrote grumpy for the same reason, which is that CPython scales pretty bad.

senorjazz5• 8 years ago

doable sure. But enjoyable? ;-)

Rotareti4• 8 years ago

try to maintain > 10k LOC codebases in Python to understand what I mean.

Since Python 3.5 you can use type annotations:

    def my_func(foo: str, bar: User) -> None:

        baz: int = User.qux

This is a useful feature for larger code bases.

In the long run, I'm sure some org will come up with a decent Python compiler, that:

* compiles Python to C

* makes use of (optional) type annotations

* improves performance

* creates single file binaries

I guess this would be less of an effort than porting the whole Python ecosystem to e.g. Julia.

demonshreder• 8 years ago

Nuitka (http://nuitka.net/pages/overview.html) seems to be attempting to do those.

TurboHaskal• 8 years ago

Everyone keeps saying that, but my experience is rather the opposite. Static typing is great when the program is small and you can keep the model in your head, but the larger the program, the more you benefit from dynamic typing. Sure, static typing helps with tiny refactorings, but making non trivial changes that don't play well with the existing model proved to be too much effort.

I can't speak for Python but since the codebases were Perl and Common Lisp based, I believe it would be similar, should there be anything similar to the MOP for Python.

kccqzy• 8 years ago

In my experience, it's the opposite.

Static typing is great but not essential when the program is small and you can keep the model in your head, but the larger the program, the more you benefit from static typing. Static typing helps with tiny refactorings, but it truly shines when making non-trivial changes that don't play well with the existing model; it's as effortless as making the desired changes to the model, then mindlessly fixing all compile errors, and it will likely just work afterwards.

amrx4312• 8 years ago

Java. C#, C++ will never be replaced. Want job security learn and code in these languages. Enterprises love them and are not going to incur cost to replace their code-bases with anything else.

rusk4• 8 years ago

Java code is a bitch to write from scratch but it never dies.

TuringTest3• 8 years ago

People said the same thing about COBOL. It was replaced anyway. You're using a really short version of "never".

oliwarner• 8 years ago

I have personal projects that are bigger than 10KLOC and professional projects that are 10× that.

I'm not sure what you're getting at but code management in Python is really no harder than anything else. It can be clean and well delineated, or you can completely screw it up... But that's not Python's fault.

simonh3• 8 years ago

it won't last forever; it never does.

The rest of my life will do fine.

I'm sure you're right, but just for fun I'll make some counter-arguments.

All technologies follow an S-curve in their 'bang for buck' life cycle. Initial improvement and return on investment ramps up slowly, then it hits a middle era where all the low hanging fruit get picked for massive returns, then it enters a third stage where most of the easy benefits have been reaped and further improvement is incremental.

We see this with the internal combustion engine, jet engines, nuclear power, all sorts of thing. Eventually another technology might come along that supplants the old technology for many or most use cases.

The question I'm pondering is, are languages technologies in this respect, or are programming languages in general a technology? If the former then Python will top out in the benefits it provides and will almost certainly be supplanted by a successor. If the latter, maybe Python is near the top of the development curve (along with many other modern languages) and further improvements will be too incremental to lead to anything supplanting it?

But if that is the case, what technology might supplant programming languages?

mch82• 8 years ago

Interesting observation! Throwing a couple ideas out there...

One possibility is visual languages that emphasize mistake proofing, interchangeable parts, and logic visualization over expressiveness. I’ve noticed an increasing number and use of visual languages (Blockly, Unreal Blueprints, NoFlo, WebGME). These are typically criticized as too limited for real programming, but seem to be climbing the S-curve.

A more likely near-term possibility is continued adoption of systems that emphasize package management, which values reuse of proven code over ability to solve problems from scratch. It seems like the language package managers are a big draw in Python & NodeJS. Maybe this trend moves even farther toward stuff like IFTT, where most people working with code are really chaining together applets.

We’re also seemingly headed toward the Star Trek TNG computer where humans query a computer with voice commands and work with it to solve problems.

jernfrost4• 8 years ago

While I may personally prefer statically typed language for large code bases, I think people greatly exaggerate the problems of maintaining large code bases in dynamic languages. Python projects such has NumPy, SciPy and Pandas are several hundred thousand lines of Python. Pandas is 250k LOC in Python.

GNU Emacs is 1.5 million lines of LISP code, another dynamic language.

It is ultimately a question of sound software architecture. You don't have to care about how many lines of code is in the libraries you use daily or in the Operating System you use.

In similar fashion any software can be modularized so you only need to concern yourself with your part.

tobmlt• 8 years ago

Chiming in anecdotally here: 50k ``real'' lines of Python and climbing. No problems maintaining it but I wrote it all for my dissertation. It's scientific/CAGD programming, -Not sure but it feels like that makes it easier. Pretty sure I need to re-write it to entice others in on the fun. Anecdotes aren't valid far from the experience, I am afraid.

I, along with one other guy, maintain a >500k fortran 77 code base (bit of C/C++ sprinkled in) at work. It's way cleaner than my home-brew python. Maintainable is about wise practices, not language choice.

nuttymoon• 8 years ago

How do you compare the popularity of full featured languages like Python to languages like JavaScript that have numerous frameworks built upon. Were React, Angular, Vue, Typescript search results taken into account to draw this graph ?

peter-m80• 8 years ago

Don't forget node.js

easytiger3• 8 years ago

In your reductive analysis x86 ASM is the most popular language or every Python application is a c application.

posamari3• 8 years ago

That’s not the point. Many JS devs google “Node xxxxx” but still code with JS. Not many people search for “Python xxxxx” and code in C.

Offtopic: I disagree with grandparent, for different reasons.

Rjevski• 8 years ago

that have numerous frameworks built upon

Python has Django though, which is light-years ahead of anything I've seen in the Node ecosystem for backend development.

bausshf• 8 years ago

And I'm light-years ahead of Django. (D Programming though.)

https://github.com/DiamondMVC/Diamond

cm21872• 8 years ago

It’s interesting that by this metric, vb.net beats javascript. It’s remarkable for a language which maintainers have placed in maintenance only mode.

I think it is a language that people should consider for beginners. Rich and helpful IDE, easy and permissible syntax (english keywords rather than symbols, case insensitive, optional parenthesis). You can do pretty much anything with it (console, desktop, website). And it provides an upgrade path to business users which only programming skills are usually toying with VBA macros.

jschulenklopper• 8 years ago

Here's a nice combination of _Python_ and an _economist_: Paul Romer, this year's Nobel prize laureate in economy, moved from publishing recent research using Mathematica to Python and Jupyter notebooks.

https://qz.com/1417145/economics-nobel-laureate-paul-romer-i...

amirmasoudabdol• 8 years ago

I found it funny/interesting that the title called it "coding" instead of "programming" language!

gaius2• 8 years ago

Yes, you can imagine the author of the article saying Nikon is the best for clicking, or that Gordon Ramsey is the best at chopping.

a_imho• 8 years ago

With such a rapidly growing user base and wide array of capabilities, Python might seem destined to become the lingua franca of coding, rendering all other competitors obsolete.

I realize it is weasel worded not to mean what it implies but statements like this is a surefire way to lose any credibility. Not that there was much underpinning besides some search engine popularity.

Nevertheless I'm baffled by python's popularity. With my admittedly limited exposure I would not reach for it again any time soon. When it is the right tool for the job?

FabHK• 8 years ago

How does that make The Economist "lose any credibility", when the very next sentence is

That is unlikely, according to Grady Booch, IBM’s chief software scientist, who compares programming languages to empires.

pmart123• 8 years ago

Coming from a data science background, I wasn't surprised by this.

* From doing Engineering problem sets in Matlab and also using Java and C/C++ in CS classes, Java's inability to overload operators made it unbearable for anything math related.

* C/C++ wasn't bad to write small algorithms but as someone wanting to look at data, I really wasn't up for solving memory leaks.

* Then the big problem with languages like Matlab is they aren't general purpose programming languages. This meant things tooling around things like sending an email notification, creating a light web framework, etc. were limited, meanwhile, numpy's eco-system began offering many of the same benefits to the commercial data science libraries.

jarfil• 8 years ago

Python is in part a "quick and dirty" language that nevertheless has many advanced features, which means it's a great language to easily hack something together.

Whether to use it afterwards in production... it doesn't shine in the maintainable aspect (it allows for some killer one-liners close to perl level), or in the low level performance aspect, so there might be some better alternative for a particular case.

Still, Python and in particular Python 3.x is a decent compromise between hackiness and something more serious, so it's no surprise that many choose it as a starting point.

saintPirelli• 8 years ago

lingua franca of coding

In my very bubble, it kind of already is though, not because Python is so good in and of itself, but because you can express general coding/programming ideas in a readable and concise manner. The syntax provides a general way to get your ideas across to programmers with all kinds of backgrounds and to proove they also actually work, anyone can open a shell and see for themselves and tweak and play with your idea.

TarpitCarnivore• 8 years ago

On the testing side of things it’s a really good option. It’s what I’ve mainly used and the Pytest test framework is incredibly flexible and powerful. I personally find it a much cleaner and more readable option when compared against Java. I’m migrating over to Websriver.io now only because I feel there is finally a good JS option that offers the flexibility needed for UI tests.

mywittyname2• 8 years ago

When it is the right tool for the job?

Anytime you need a simple web app.

Writing powerful CLI tools. Take a look at AWS's Python API.

Data wrangling -- it has libraries for handling so many data formats.

Quick and dirty scripts. I've used it for many one-time scripts that need a bit more than regex.

Data science / analysis -- it has so many great libraries for math and data wrangling.

The biggest pain-point I have with it is around module handling. I'll often run into issues the first time I'm deploying a new pycharm project. The second biggest pain-point I have is around concurrency/multi-threading, but I think the concurrent.futures module might provide a decent solution.

jokoon2• 8 years ago

Well now we have to wait for a v8 equivalent for python. I guess pypy doesn't really make python that much faster, can it be faster?

rkeene2• 8 years ago

You could look to TclQuadCode for inspiration as it can make Tcl code 66x faster in the best cases.

jacinda2• 8 years ago

Have you looked at Cython?

jacinda2• 8 years ago

You can also profile your code to find the hot spots and write C extensions just for those areas.

bakery2k3• 8 years ago

Python is, despite its simple appearance, a much more complex language than JavaScript. A V8 equivalent for Python is certainly possible, but would require even more work than went into creating V8 itself - who would be motivated to spend that much effort on Python?

Dropbox? They seem to be the company driving Python nowadays, but gave up trying to speed up Python in favour of writing "performance-sensitive code in other languages" [1].

PyPy uses a meta-tracing approach in an attempt to reduce the amount of work required. Its speed is impressive, but I'm sure meta-tracing leaves some performance on the table.

Another way to reduce the amount of work required would be to support only a subset of the language - for example, a V8 equivalent for MicroPython rather than for full Python. I'm sure this could be made very fast, but it would have very poor compatibility with existing Python code, so what would be the point?

[1] https://blog.pyston.org/

jokoon2• 8 years ago

What makes python so complex? What are the features that if removed could really improve performance?

I love python's syntax, but I could do without its lazy interpretation, without duck typing if possible, and without its reference system.

In short I just like the syntax of python, I don't really care about its modules.

diminoten2• 8 years ago

Having worked in Python as a way to pay for food nearing 7 years now, the biggest complaint I have at this point is that, at an enterprise level, it's severely under-equipped to enforce its touted "There's probably one right way to do things" mantra.

People are often introduced first to Python by way of scripting, but the tricks and tools learned while scripting are exactly the sorts of things I've found cause dirty code. Applying clean code principles to Python is hard, because it's so easy to write dirty code, and that's a byproduct of how easy everything is to do in Python. That ever-nagging "but it's working" continues to get in the way.

It's a weird problem. I dunno what to do about it, and currently end many of my debates with colleagues on topics related to the problem without agreement.

houzi• 8 years ago

Most programming languages suffers from lack of enforcing idiomatic coding style, don't you think? Perhaps python more than others? Scientists having learned that OOP is the one true methodology make horrendous code if a language has classes. I've also seen FP coders being a bit too extreme in Python, inlining everything and not using comprehensions.

ubernostrum3• 8 years ago

People are often introduced first to Python by way of scripting

I don't believe this is true, and I don't believe it's been true for quite some time.

First encounters with Python now tend to be via frameworks, whether for web development, machine learning, data science or automation.

diminoten2• 8 years ago

It was true recently enough ago that the people writing Python code professionally were introduced to Python from a scripting entrypoint.

For non-programmers, I think it's definitely still true that "scripting" is their primary use case for Python, and when those non-programmers shift into software development jobs because they have "3 years experience" with Python, they bring all of their bad habits with them.

moomin• 8 years ago

StackOverflow has the data on this: Python for most purposes is entirely static. Python for data science is utterly dominant. That’s what’s driving the stats.

sathishvj2• 8 years ago

I can see the sudden jump in popularity for python alongside data science - it really is great for that.

My view is: python is great for offline tasks, like functional testing and training in machine learning - tasks that do not have huge, direct customer interaction. But to put something in production, please give me a strongly typed language with compiled binaries and no additional installations/VMs (and for me that's Go).

devxpy12• 8 years ago

Yeah distribution is a _huge_ problem for python right now.

timClicks• 8 years ago

Is it though? I mean, yes there is fragmentation and lots of old documentation but that's because Python is 25 years old and several approaches have been tried.

What's especially difficult about the following two steps?

  pip install pyinstaller
  pyinstaller main.py

ubernostrum3• 8 years ago

Python is strongly typed, insofar as the term has a meaningful definition at all.

You meant "give me a statically typed language". Which may not actually be what you want, since statically typed languages can be, and some historically have been considered, weakly typed (C, for example).

sathishvj2• 8 years ago

doesn't strongly typed mean that the type cannot change once assigned? and you can do that in python, can you not?

archi42• 8 years ago

I usually shudder at duck typing and/or GC for anything bigger. Both have the ugly tendency to explode in your face at runtime. The first type of error doesn't occur in C++ (at runtime), and in the other instance I feel like having more control. But maybe my academic experience with the two is to blame (multi-person projects in python (multiplayer game using lecturers lib) and Java (SQL DBMS, GC killed performance while in C++ I would have written a specialized allocator)).

hyperion20102• 8 years ago

I started programming in python in 2010. It has been a good tool for solving a variety of problems.

I do have to wonder though. Is python popular right now because so many more people who have never programmed before are now starting to program?

I came from a biology background and my hatred for matlab following one quarter of exposure was sufficient to push me to python because that was what was offered as an alternative (taught by a gradstudent rather than the guy who funded the lead dev for the Octave project). There was no other choice. Julia didn't exist etc.

In the 8 years since I have written a lot of bad code have come to very much dislike python (though I still write it daily). The reason is not because of any of the visible failings of the language. It is due to the deep hidden inconsistencies, and unexpected and unpredictable behavior of the runtime (try using list comprehension in class scope (What is class scope you ask? Well. Python is a language for newbies... so by the time you can ask this question you are too invested to realize that you might have made a bad decision)).

Python is kludge on top of kludge without the diversity of implementations to take another path when someone calls bullshit. Python can get away with it because they have runtime hacks that let you sort of do what you want, but there will be a corner case that doesn't work because there is some unspoken and undocumented assumption that you are violating and will be duly punished for.

None of this is the language's fault. It's the runtime's. The python runtime (don't even bother with discussion about the GIL, which is a red herring and a bad case of Stockholm syndrome at the same time) is a mess. The language is fine, even great. But the fact that different scopes have radically different meaning and that expressions only work in certain scopes [0] is a complete nightmare.

My main emotional interaction with python is that if I follow the BDFL approved path, everything will work. If I stray from the fascistic predetermined happy path, good luck to me, I was on the wrong side of history TM. 99% of the code that I write that doesn't start as a class I end up rewriting as a class (probably because python doesn't have closures ...). Python lures you in with promises of easy to write easy to read, but when you get to that point where your actions are no longer approved then then pain and sadness and rage are all you have left. I won't name the language that has almost the opposite interaction that I come to vastly prefer, but the emotional experience is radically different, I get extremely frustrated because I don't understand something in that language, and eventually come to realize it is because I did not understand something simple. In python I initially think that I am the one to blame, but after months or years, I finally realize that it is actually the language that is fundamentally broken and inconsistent and I have been mislead the entire time. I get frustrated because there no possible way for me to solve my problem by increasing my knowledge, I would literally have to fix a 4+ year old problem in the language (which I'm sure I couldn't do and that I'm sure has not been done for very good technical (debt) reasons), or find a library. There are a lot of libraries, but don't be fooled. Every library is a new language. Python is just the gatekeeping runtime.

Python will thrive for a large set of use cases, but when you hit a wall in python, the problem is not because you don't understand, it is because python is broken. That is not a happy realization. It is like growing up and realizing that people are actually evil, not that you were just misunderstanding them. Or more generously, that expecting people to be on their game 100% of the time is unrealistic. The difference is that python is that attractive person that is great at first impressions, but a disaster in the long term, and if I hadn't had the good fortune to meet a language that was super awkward for the first 6 months or so, but turned out to be a keeper, I'd believe that python was the only thing there is.

Python is and abusive language, and probably exists and is successful because things like perl also exist (and because it solves a whole bunch of practical problems). I'll keep using python for a long time, and the coming of age of pypy brings me great hope for the future, but there are classes of problems where I now go elsewhere. There are problems everywhere, but in some places they seem to be rational rather than contingent.

tl;dr Don't pick a language, pick a runtime.

0. https://github.com/ipython/ipython/issues/62

lizmat• 8 years ago

Perhaps it's time to look at Perl 6 ?

tootahe45• 8 years ago

I do have to wonder though. Is python popular right now because so many more people who have never programmed before are now starting to program?

In recent years coding has become hip/cool. I'm pretty sure none of the 10,000 Bootcamps/online courses that popped up are teaching c++ at the moment, it's all the easy stuff like Python and web. This gets people to their 'aha' moments quicker and keeps them feeling like they're progressing, even though they're doing barely any thinking. I will admit though if there is one language for anybody to learn for general purpose, it's Python.

restalis• 8 years ago

"a language that was super awkward for the first 6 months or so, but turned out to be a keeper"

I wonder, do you consider the language's awkwardness somehow related to its quality? Do you see it as a prerequisite to avoiding language shortcomings?

hyperion20102• 8 years ago

No. In this case that part of the analogy is actually not apt at all. To use another bad analogy I think my point is that the feeling awkwardness actually comes from the fact that I was doing the equivalent of holding the hammer by the head and trying to pound nails, only to later realize that I should have been holding the handle, and that there actually was a handle. In python it is like having a sledge hammer that you can't possibly hold by the head, but that no matter how hard you try you keep putting holes in the wall when you try to hang a picture.

Someone will say 'but why not use the tack hammer?' and, at that point I argue, I am no longer using python the language, but a whole new language that happens to look like python but that has a different runtime and a different set of assumptions. This even happens inside of the core language where `[d(a) for a in b if a == c]` can be valid in some scopes but not in others. The other language does the opposite, it says "you can't do that" and so I try to hold the hammer by its head and think that it is the language that is awkward because I don't know how to use it and that there is an fact a nail gun that is configurable along every dimension from nail diameter, to nail length, to insertion speed and pressure profile over the whole insertion timecourse, it is just not there in plain sight and I the first few times I used it I accidentally nailed my feet to the ground with railroad spikes.

I guess it is more that every time I learn more about python it is about what I have to NOT do in order to be safe. In the other case I learn what TO do in order to solve the problem the way I want to. Maybe the clearest example I can think of is lambda, where you have something that looks like a lambda, but that doesn't doesn't even work like the def (): that it copies due to the biases of the language designer. Honestly it would be better if Python just removed all the syntactic features that don't have generalized runtime support because that way you don't have to know which features you can use where. Sure, there is a use case for lambda in certain contexts, but if Python refuses to support lambda as a general abstraction that I can use everywhere, then it should be removed from the language, because including it leads to those hidden inconsistencies which require an enormous memory burden on the developer just to know what is not allowed. That is dead knowledge that in no way empowers and enriches the programming experience.

In essence the push for pythonic code is not just about maintainable code, it is about keeping people on the happy path without removing delimited abstractions, which is fine, except that the explicit solution to this problem would be to have `from __i_know_what_im_doing__ import lambda, comprehensions` so that the core language didn't misleadingly offer these delimited abstractions. Of course this means that many of my favourite features would be thrown out until some core technical challenges were resolved.

TimJYoung2• 8 years ago

I only know about Python from what I've read and tinkered with, but IMO, Python = FoxPro (look it up): it is fairly easy to learn, immensely useful and you can do lots of cool things with it, but it a) relies very heavily on C or external libraries for any type of raw code performance, b) it's inconsistent and the inconsistencies are somewhat hidden and not obvious, and c) distribution is a nightmare, so developing anything professional that needs to be dropped on a target machine is a problem.

cmcginty• 8 years ago

Unfortunately you have not provided a lot of examples to go along with your major complaints. I feel like at some level, every language will have its idiosyncrasies that create annoyances and issues. From what I can tell, you are describing concerns that would not void the use of Python entirely. These issues tend to increase with higher level languages as the complexity of the feature set expands and it become unreasonable to support every combination of every idea.

I don't know if you'll have better luck switching after making such a large investment in Python, but if you do find something to supplant it please document your experiences for other to learn from.

bsimpson• 8 years ago

At the bottom of the page, there's a link to "the full article":

https://www.economist.com/science-and-technology/2018/07/19/...

It's fascinating to see the differences between the two. The OP is a bit shorter and a bit more awkwardly worded. It also inlines some names (for instance, referring to the package manager as "an online repository" instead of "the Cheese Shop").

I wonder if the OP was written for the website (with a lower quality threshold) and the "full article" was edited/expanded-upon for the more prestigious print edition.

pinoytambayan• 8 years ago

What about Asp.net?

romanovcode• 8 years ago

It's not a language

legostormtroopr• 8 years ago

As a huge fan of Python, my go-to statement is that Python is the second or third best language for whatever you want to do.

Want to do web server development? Flask, Django, Tornado, etc... can do that. Probably not as concurrently as Go, but it'll do.

Scientific computing? Its no Matlab, but the syntax is easy and the time you lose in processing you make up for when developing.

Desktop computing? PyQT is pretty good. Electron may be faster to build with, C will be faster at runtime, but Python, even with Tkinter, is ok.

Functional programming? There are list comprehensions, map, reduce and filter. Not as good as Haskell or Lisp, but it'll do.

Python isn't the killer language for any of these really, but its good enough in almost all cases that you'll be very productive and can iterate to a good MVP quicker than with most other languages.

zimablue2• 8 years ago

It's the killer language in scientific computing I'd say, having used both Python and Matlab. The open source nature/libraries and being a more fully-powered programming language (expressive object system) make it better.

dagw2• 8 years ago

As someone who does a lot of scientific computing in Python, I'd probably recommend learning Julia as the 'killer language' if scientific computing is the main reason you want to learn programming.

Pythons strength comes from being pretty good at scientific computing, and equally good at just about everything else.

dazzawazza• 8 years ago

For me python's killer feature is readability years later.

I've gone back to websites and tools YEARS later and the code as quick to pickup and easy to modify.

35 years of programming and I've never met a language so well integrated with my mind.

deaps• 8 years ago

35 years of programming and I've never met a language so well integrated with my mind.

That's a pretty good way to put it - and mirrors my feelings. I've been programming for 20+ years (with breaks here and there), and it took a while for me to find Python. But when I found it, it's like I finally found the language that my brain thinks in.

mazesc• 8 years ago

Any language with static types should be easier to make sense of, though. I get it's faster to write dynamically typed code, but your point is specifically about reading code, which I can't follow.

Of course Python nowadays has support for static types too.

tpaschalis• 8 years ago

Personal anecdote as well, but I think you nailed one of the reasons I love Python.. It's the way the language was designed (PEP 20)

I had written some POC scientific code for my undergrad thesis, but the thesis was sidetracked for about a year or so due to real-life issues. When I went back, I could try and test things again in just a day or two, in comparison to much more detailed documented Matlab and C code.

- There should be one-- and preferably only one --obvious way to do it.

- Simple is better than complex.

mmargerum• 8 years ago

This is also what i love most about go. The code is so easy to read and thanks to `go fmt` everyone's code is structured the same way.

skohan• 8 years ago

I would seriously recommend Swift for this reason. It's the most self-documenting language I have worked with.

devxpy12• 8 years ago

I was just talking to my friend about how in python I can just write "and" "or" "not" instead of the symbol orgy in almost every other language xD

3rdAccount4• 8 years ago

I was going to say "no" to embedded, but micro python and other projects work on some controllers I think.

jononor• 8 years ago

In the microcontroller it is good enough for hobby/proof-of-concept level things. Same as the various Javascript engines. Lua might actually be doable for some simple products. But a lot of embedded these days is on Linux, and then Python is pretty great.

sibinpthomas• 8 years ago

Seems like this applies here - https://www.ribbonfarm.com/2018/04/24/survival-of-the-medioc...

ts330• 8 years ago

Just because people are searching for it doesn't mean it's popular. It means people are trying to answer a question about it. I'd posit there is more of a correlation between searches and newusers/issues/problems with the tool/language than how much it's being used. Stack Overflow is a better source for these stats anyway.

iliaznk2• 8 years ago

Python is great, I just wish it had JS-like object destructuring applied to dicts – so convenient!

rataata_jr• 8 years ago

If you can think it, it can be done. There is probably a library for it. It is just not standard.

iliaznk2• 8 years ago

That I have no doubts about, but having it in the standard lib would be awesome!

Rotareti4• 8 years ago

Me too. There is a library that solves this: https://github.com/Infinidat/munch

    from munch import Munch

    foo = {'a':1, 'b':2}

    foo = Munch(foo)

    print(foo.a) # 1
    print(foo['a']) # 1

jasongi• 8 years ago

You mean like

  a = {"foo": "bar", "bar": "foo"}
  b = {**a}

You can do this since 3.5 https://www.python.org/dev/peps/pep-0448/

You can also do dict comprehensions which is much more expressive

  b = {k: v for k,v in a.items() if v == "bar"}

This way you can copy only the parts of a dict you want in a simple one-liner.

https://www.python.org/dev/peps/pep-0274/

devit• 8 years ago

Is it really?

Seems unlikely Python can be more popular than JavaScript/TypeScript since they are equivalent to Python and have the critical advantages of working on the web (and thus being known by all web developers) and having optimized engines in web browsers.

goatlover3• 8 years ago

The computing world is a lot larger than the web, believe it or not. Python is huge in scientific computing, and is used for lots of server-side scripting. JS/TS don't have much of a presence in scientific computing beyond the web presentation layer, which can be generated from Python libraries. Python is used everywhere in science, and the ecosystem is incredibly rich.

Jenz• 8 years ago

While pythons popularity is undeniable, Im sceptical to the article as it’s based on the TIOBE index.

zmmmmm2• 8 years ago

I understand it (and code a hell of a lot of Python) - but it's a shame. While it's a lot better than the many low-barrier-to-entry languages that came before it, it is such a heavily compromised, suboptimal language in so many ways. I feel like a whole generation of programmers is becoming stuck in a local optimum that will cost us very dearly later on.

djaychela• 8 years ago

How so? Not trying to pick an argument, just wondering (as someone who's learning Python) what you mean, specifically, that will lead to this?

zmmmmm2• 8 years ago

Well, this is how lengthy arguments start about pros and cons of languages, as there are many many people who will argue these are not problems or can be worked around or are actually preferable for whatever reasons.

But my biggest problems with Python are:

    - Lack of strong enough / useful / culture around typing
    - Effective limitation to single core through GIL
    - Lack of proper solution to closures

Really the first two are the ones that make it for me, a fundamentally crippled language. And they are exactly the things you don't notice when starting out, and the things that limit you eventually.

batty_alex• 8 years ago

While you're likely correct that FORTRAN, COBOL, C, Visual Basic, C++, Perl, ColdFusion, Java, Javascript, and Python, were sub-optimal choices, they're just tools whose popularity boils down to current programming pop culture.

That being said, it could be something much worse than Python.

senorjazz5• 8 years ago

When ever you start feeling like this, just think to yourself how bad things would be if it was javascript that won and was used everywhere...

oh wait :(

asien2• 8 years ago

Never been a fan of Python , but all the comments about Java are legit.

I’m a big Java User mostly for Enterprise Apps & Services. A third of my code is boilerplate, Python really shine when it comes to making stuff concise but in the same way not having type is something that has always bothered me.

Dynamic typing is something I have accepted with JavaScript but for Python for some reason it has always bothered me...

a3n2• 8 years ago

Python is a strongly typed language.

https://duckduckgo.com/?q=python+strongly+typed&t=fpas&ia=qa

meepmorp2• 8 years ago

Strong typing and dynamic typing are not mutually exclusive.

RickJWagner• 8 years ago

I'm not surprised.

Python is a very easy language to learn and use. In the end, ease of use always wins.

puranjay• 8 years ago

I picked up Learn Python the Hard Way. Really enjoyed it. Was super easy even for someone new to programming.

But got lost when I reached the OOP part. This speedbreaker, along with a few other distractions effectively eroded my enthusiasm for coding

village-idiot2• 8 years ago

A few jobs ago was in Python. Joy to write, absolute terror to deploy.

Seared into my mind was a failure caused by delivering .pyc files (oops) to a server whose shared libs were compiled with different flags than our CI box.

To be fair, this was before docked (or at least before I’d heard of it). Maybe that’s improved things a bit.

mywittyname2• 8 years ago

Deployment is still mostly awful. The community is still tackling this issue as of 2018.

https://realpython.com/pipenv-guide/

village-idiot2• 8 years ago

It's a shame. I rather like the language, but failing to fix the deployment story for this long is frankly unacceptable.

casper345• 8 years ago

Question: How is python on more mechanical systems, like robotics in such? Will we see a migration there besides web env?

kjullien• 8 years ago

I learned development with Python (after trying to grasp the concepts of C for a while) and I have to say that alongside .NET it is one of the easiest languages I had the pleasure to work with. So much so that after 10 years of mainly Javascript and Node.js work I have decided to go back to Python simply to make my life easier and be more productive.

anonu• 8 years ago

This is a shorter version of this July 19, 2018 article: https://www.economist.com/science-and-technology/2018/07/19/...

Python is becoming the world’s most popular programming language

Related Stories