Its such a statement about the free rider problem that python can have only 2 people working fulltime on it. Python is one of the most popular languages in the world - there's got to be trillions of dollars worth of companies depending on it. Small improvements to the language would result in millions of dollars of productivity gains in aggregate globally.

And yet ... collectively we can only find enough money to pay two full time salaries?! Seriously?

And large improvements to the language could result in millions of dollars of productivity losses! :-)

I'm not a big fan of "Now this language has gotten popular, let's add all the features to it that made our other language so unpopular."

Everything about language maintenance isn't "just" adding features. There's actual bug fixes and a bunch of other stuff that take a lot of time. Not to mention that Python's standard library is relatively large.

All the evidence is 2 Full Time Developers are equal to the workload. People aren't using Python because it is going to be good enough in the future, it is good enough now.

If more people are working on it full time, we'll just get more Walrus Operators. A very hypothetical gain at cost of an increasingly complex core language.

I always have been and always will be against the walrus operator. I don't care how much anyone tries me to sell it on its elegance. There's just no _need_. It's already the #1 example of wtfpython https://github.com/satwikkansal/wtfpython#-first-things-firs...

/rant. Those kind of features are going to keep coming anyway, I'm afraid.

It's the #1 example of wtfpython for no good reason.

It's like someone used to PHP or JS style coercion, doing: 1 + "2", getting a type error and asking "wtf python?".

The walrus operator is obvious, and very handy in many cases.

The capture part of the pattern matching PEP, on the other hand, has been designed shodilly.

I don’t like the walrus operator personally, but wtfpython has it wrong. It’s simply syntax for an expression that includes a binding. It parses like any other expression, and is completely unsurprising.

It's not about going million miles an hour. The safe speed can be used even if there's 200 people. But the commenter above has a good point: tons of businesses and billions are linked to a product with almost no workforce behind it. The same is with many other OSS: tons of value and no money to those who maintain/create.

If we want to solve this problem we need to start with a way to assign credit and identify neglected parts of the stack. After we have awareness we can find the funds.

Not necessarily. One can be done without another,but it would obviously help if there's some roadmap that can then be shown to the potential sponsors. Getting funding is ultimately a sales function and I'm sure any meaningful project would have enough dough to run on if such function would exist.

It taking 10 years to transition from 2 to 3 wasn't necessarily a bad thing.

Definitely concerns that more hands could make another Angular, which had 2 incompatible branches, with one going from v2 to v11 in 5 years.

> It taking 10 years to transition from 2 to 3 wasn't necessarily a bad thing.

The python3 interpreter could have been largely (or completely) source-compatible with python2 code. They could have done it rust2018 style - where source code files can opt in to a newer parser, and in return get access to the new features. Then the whole python3 drama would have been avoided. MacOS still ships with a python 2.7 interpreter because of all that nonsense.

But if they only had 2 full-time people working on it, no wonder they didn't have the man power to do the transition any other way.

Disappointing.

Which puts a:

> WARNING: Python 2.7 is not recommended.

> This version is included in macOS for compatibility with legacy software.

> Future versions of macOS will not include Python 2.7.

> Instead, it is recommended that you transition to using 'python3' from within Terminal.

banner on launch nowadays...

Right; none of this should be necessary. I don’t need an old version of my C compiler in order to compile old C files. Rust made breaking changes to their language and yet a modern rust compiler will happily compile the older 2015 rust syntax. Python could have gone down the same road; but I suppose all of python’s corporate users chose to fight python compatibility changes locally rather than donate time / money fixing python itself.

Improvements ≠ new features.

because software that is provided for free cannot capture the positive externalities it creates and thus is under-supplied, it's just very straight-forward economics.

Yet it's become such a widespread sentiment to expect software for free that charging for something like Python would probably be seen as obscene and kill the language through loss of reputation alone.

>Its such a statement about the free rider problem that python can have only 2 people working fulltime on it

It's also a big statement about how non realistic were the ideas that FOSS will thrive on volunteers working together.

Back in the day where the "Cathredral and the Bazaar" was published, VALinux got its IPO, and FOSS was the latest hype, people really believed that this will work without corporate sponsorship, just by volunteers doing the work.

Then again, they also believed in many eyes available through the magic of FOSS making bugs shallow (Torvalds), and you get 20+ years of some obvious bug in something as critical as OpenSSL.

Is that actually a bad thing?

Much of the popularity of Python is based upon third party libraries that have been contributed by both individuals and organizations. These are still contributions, and there are many cases where the contributors are paid.

Should Python receive more funding and pay more people to work on the language? Probably, yet it is doubtful that having a large team working on the core language will improve the health of the language.

> And yet ... collectively we can only find enough money to pay two full time salaries?! Seriously?

But what’s the problem here? Is python struggling as a language? If there were clear calls for help companies could find it a lot easier to justify donating money or dev time.

>Is python struggling as a language?

As a language, interpreter, runtime, yes. In terms of adoption, no.

With what? Why haven’t they said so? There are a bunch of tech companies that would love the exposure of employing core developers.

I am seriously wondering: is it actually possible for stock listed companies to donate? I mean aren't they legally obliged not to spend more than necessary (towards their stockholders obliged). Also what does the IRS say on that?

No. This is a misinterpretation of the responsibility. Companies have great latitude in what they do, as long as they're forthright.

They have a duty to try to operate efficiently, but that doesn't make corporate social responsibility, a bad hire, or wasting something nobody liked in the canteen illegal.

Well, bad hire, bad food can always be justified (avoiding hindsight fallacy). But giving away money for nothing received (meaning of donation) is something different (at least to me)

companies donate money to stuff all the time what are you on about?

So I read up a little. To tax deduct the nonprofit must have registered as 501(c)(3). Is this the case here?

I can pay you to put my company name on your business website or a physical banner at a trade show and account for this as a business expense (deduct it) without you being a non-profit. I have broad latitude to make business expenditures in a manner that I believe is profit-seeking or optimizing for my business.

People do this with Google and Facebook all the time. We call them ads. Not every ad campaign has a positive RoI.

Yes, its possible, and, yes, they actually do. And the IRS is fine with it.

To me that is more like evidence of a well-engineered codebase.

Python is as popular as it is because it's free, not because it's good. If companies had to contribute, many would pick something else.

So... why don't you pay?

I say this as someone who's spent decades of my life donating time to opensource projects. - But the model where big companies have their technology stacks underpinned by charity is ridiculous. Obviously ridiculous, if python only has 2 full-time-equivalent staff.

This problem isn't solved by little old me pulling out my check book. My bank and my insurance company don't need charity from me in order to fund the technology they've built their businesses on. They need to learn to pony up and start funding the "free" software they use. Paying myself would only help kick the can down the road - and encourage and enable this sort of corporate mooching to keep happening.

In 60 years most of the volunteer maintainers of the software we depend on every day will be dead. Some time between now and then we need solution to this problem of opensource sustainability. Delaying buys us nothing.

If python continues to be this popular, there will be more maintainers who join the project.

Surely you’ve seen volunteers join existing projects out of their own interest and need now. The decades that you contributed (thank you, BTW): were those all on projects you started, or did you also join some existing efforts?

I don’t see any reason to think that will stop in 10, 20, or 60 years.

> If python continues to be this popular, there will be more maintainers who join the project.

Will there? I feel like most of the opensource attention flows gradually up stack / toward trendy things on github and away from infrastructure older than github. When heartbleed happened a few years ago it was revealed that OpenSSL was maintained by basically one guy. The same is true of the linux timezone database. Apparently python is mostly 2 people.

I'd put money on this being is a theme for lots of tools we take for granted. How many people work on bash? Or busybox, grub, libpthread or libicu? I bet the number is tiny.

Cpython had 5 pull requests opened yesterday. 6 the day before that (from 10 total distinct contributors). That doesn’t seem like a project starving for contributors. If everyone with commit access up and quit, there’d be turmoil for several weeks and something reasonable would emerge.

People want to work on popular, useful things. Python is both of those.

> the model where big companies have their technology stacks underpinned by charity is ridiculous

Worse: they are exploiting people's good intentions.

They co-opted the Free Software movement.

Copyleft has been painted as "business-unfriendly" and replaced with permissive licenses - aka unpaid labor without "strings attached".

> They need to learn to pony up and start funding the "free" software they use

...instead they even dodge taxes.

Why do you assume he isn't paying? The payment should obviously be proportional, though. You can't expect a single individual to finance what a gigantic multibillion corporation could.

Just because some random individual came up with some arbitratry agenda of whom should finance what, doesn't mean that it makes sense.

There are so many better (and worse) things to finance than Python Foundation.

It's not very arbitrary. The reasoning (in one of the upstream posts) is that Python is one of the world's most used programming languages. It definitely makes sense for large multi-billion corporations which are the users of this language to help finance it.

Your point would make more sense if the claim was that a significant portion of the corporation's profits should go towards financing Python. Then I would agree that "there are so many better things to finance than Python". But if we're discussing a salary or two in the context of corporations that employ hundreds of thousands of people, then your argument makes much less sense in the current context. There are very few things that would have better marginal utility then paying for the (effectively) third full-time Python developer.

Someone else should!

That's a pretty distorted view IMHO. There are a lot of people that are working in the Python ecosystem and improve the language, and while they are not directly paid by the PSF their employers give them time to devote to working on Python. I think e.g. GvR had plenty of freedom to work on his Python projects while at Dropbox. I know a lot of companies in Germany that give people ample time to go to Python conferences and engage with the community, as it's also a way to meet and recruit good Python developers.

So it might be true that there are only a few FTEs working on Python, but the larger community of people contributing to it is much larger.

> GvR had plenty of freedom to work on his Python projects while at Dropbox.

That post was made while Guido was at Dropbox, and it claims he had one day a week?

They talk about maintaining Python, which includes things like preparing and managing releases. Why would you need more than a few people to do that? The core language itself also evolves at a slow pace (which is good), it's not that there are tons of new features every year that get merged into the codebase, so a few FTEs is probably adequate to handle that kind of work.

Considering the larger Python ecosystem the community does a lot of the work, so saying there are only two people working full-time on Python is quite misleading IMHO, as Python has probably one of the largest communities in the world for a programming language, the Github repo alone lists almost 1.500 contributors.

That's insane. A thought occured recently - some lines of code in the python code base have been executed an innumerable number of times (trillion-trillion? Can't even start to calculate the order of magnitude). Even the subtlest of optimizations to these lines could collectively save probably hundreds or thousands of tons of CO2 from being released. That alone might deserve a couple of People looking into such massively common codebases fulltime.

And yet… Did you know about this (serious) proposal to make CPython 5x faster?

https://mail.python.org/archives/list/[email protected]/...

Needed funding though.

Why does this not happen or anything like this?

Ruby got MJT in a big effort to make Ruby a lot faster. PHP is constantly increasing it‘s performance and tries to change a lot to get faster (basic building blocks for JIT in PHP8).

Why is python not iterating so much on it‘s performance? Or is it doing this but it‘s not as public in promoting it like my former two examples.

There's already PyPy. It's faster for, I estimate, 90% of workloads that don't use NumPy or Tensorflow, which offload work to faster implementations in different ways. It can certainly be 5 times faster.

It doesn't have that much adoption, which suggests that performance isn't critical to Python's user base. If PyPy got significant market share, perhaps the CPython developers would prioritise an approach like this. (Or perhaps they'd have even more justification to say look, there's already a performance-focused implementation, we'll do our thing).

Pypy isnt used that much because you may run into problems. And nobody wants to research strange runtime errors in production.

But the speed improvements would be welcome if they were in cpython because you can be quite sure there will be no problems.

Agreed (and I like pypy)

One example is that Pypy switched away from reference-counting to fully garbage collected. They are probably right that's faster, but if I'm running a webserver, it's actually really comforting to know that any garbage produced by the processing of a request is collected when that request ends (barring cycles of course) so I don't end up with weird hangups.

Using a GC doesn't preclude doing a correct use of ResourceManager.

The author claims to have _existing code_ for the first stage, but provides no code and no benchmarks in the post or linked github repo. Additionally, the repository does not seem to make a particularly strong case that each improvement will result in the massive 50% speedup compared to the previous as claimed.

I'm not in the python community, so it's possible this person is well known and respected enough that part of this can be taken on faith. However, because the author does not provide benchmarks or any other proof of the effectiveness of the supposedly almost complete stage 1, I cannot personally call this proposal serious.

If you actually read the thread, it is evident that the author is in fact well-known and has worked on this for many years.

I decided to read the entire thread, and my final conclusion remains roughly the same. Although "not serious" is perhaps too harsh, the author does not provide the level of detail I'd expect of someone asking for 250 thousand dollars for their existing work (as well as 750k for future work).

The most information we got was "on a few simple benchmarks I'm seeing about a 70% speedup vs master for both default and LTO configures." However, this information is almost completely useless. What benchmarks? What architectures? What hardware?

It would take the author 5 minutes of their time to say "here's the benchmark code I ran, it executed in [time] on CPython master commit [hash], and [lower time] on my patch, running on a machine with [hardware]." To be completely honest, I fail to see why they are not providing this information that would make their proposal significantly stronger.

> The author claims to have _existing code_ for the first stage, but provides no code and no benchmarks in the post or linked github repo.

Since he was only asking for payment once he hit his benchmarks for each stage, why does it matter?

Very interesting thread!

difficult to get this through without going through the established channels and dealing with the politics straight on

> Two years ago it was estimated that it only had (the equivalent of) two full-time employees working on it.

The discussion thread on that reveals that it was a bit low even by its explicit premise (core devs that had allocated time for OSS that was used for Python), and that methodology itself undercounted reality because it ignored people who had and made use of employer permission to use paid time but did not have a set allocation.

And these caused a lot of disagreement in the core team. Many thought that they were introducing too much churn.

Coupled with the fact that they get paid for the alleged churn and others have to catch up on unpaid time, this can ruin a project.

The Python dev community has certainly gotten far less vibrant (and productive) in the last decade due to stifling corporate influence.

Only two people working full-time on python is not the same as “(the equivalent of) two full-time employees”.

One is two people. The other is a few thousand person-hours of effort per year.

I’m not sure what you mean. How is “two people working full-time“ not equivalent to “two full-time employees”?

Two people working full time on Python have two names.

The equivalent of two full time employees could be 10 people with 20% of their time. I suspect there are a lot of patches submitted to python that were developed by someone being paid while they did the work that isn’t accounted for in the 1 actual full-time plus 5 20% time people.

IOW, paying for 1 additional named person to work ~2000 hours in 2021 will almost surely not be a net increase of 50% of capacity.