Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> How large is the Flutter team, today? Google doesn't publish this information, but my guess is that the team is about 50 people strong.

> That's 50 people serving the needs of 1,000,000. Doing a little bit of division, that means that every single member of the Flutter team is responsible for the needs of 20,000 Flutter developers! That ratio is clearly unworkable for any semblance of customer support.

This is a weird exercise. Python, for example, is a #1/#2 language in the world and there's only 50 active core devs, 90% of which don't even work on Python full time. Somehow we make that work.



I have statistics about real team size behind open-source repositories, and it is exceedingly rare to be larger than 50: https://play.clickhouse.com/play?user=play#U0VMRUNUIHJlcG9fb...

For comparison, my open-source product has around 20..30 according to the monthly stats (I want these numbers to be higher, though): https://play.clickhouse.com/play?user=play#U0VMRUNUIHRvU3Rhc...


This is significantly undercounting in at least several cases I know of. e.g. I was employed by Elastic and Kong and in both cases, the number of developers employed by the company to work on these projects was several multiples the size reported here


Ah, I now see why: I hadn't really looked at the query, but now that I have, here are some of the problems:

1. Using median is a problem for many projects because many often start off (sometimes for years) with only a few developers before they really take off. The median() number for Elasticsearch was 28, but the max() is 49, which I suspect is closer to the current number.

2. Not everybody uses a pull request flow, and not everyone working on a project is a developer. You could have product managers, technical writers, engineering managers, etc, all working on a project and not opening PRs and you can have users just directly adding commits, especially when a project is young. The original query here for the kong/kong project shows a team size of 6, but if you remove event_type = 'PullRequestEvent' and switch to max(), you get 33.

3. Sometimes, parts of the project are kind of "elsewhere." e.g. Elastic employed a number of Lucene committers to help move Lucene along, Kong employed a number of folks that would e.g. commit/maintain OpenResty plugins which got incorporated in the build system, the docs live in separate repos in both cases, etc.


Python's standard library rarely changes except deprecating features or new additions.

Underneath - POSIX + handful of Windows APIs for porting it to Windows. Cross platform aspect is mostly for the compilers to produce the binaries.

Flutter is not Python, PHP or Whatsapp. It has to emulate UI of six platforms (Web, Mac, Windows, Linux, Android, iOS) with easily 50+ widgets.

Flutter unfortunately is not comparable to a language interpreter. The surface area is altogether different.


Every project is different. Claiming it needs more developers because "it needs to emulate UI of six platforms" as if that justifies a larger team is also fallacious.

I wanted to see the evidence that the community is lacking support and/or flutter is lagging behind competition. That would justify Flock's goal... Otherwise, it sounds a lot like both naivette and a unpublished agenda.


“Flutter team members don't actually use Flutter.” (as per the website).

This could explain a lot.


Very rarely do the people working full time on a framework/tool also use that framework/tool in a non-toy setting. They aren't working two full time jobs after all.


Always, always, dogfood what you produce. The number one way devs cease making products people care to use is by not using what they make.


I think there should be a distinction here. E.g. if you work on a browser, possibly implementing parts of image loading, or javascript parser, etc.

Are you consider a dogfooder if you use the browser? or do you need to lots of write Javascript yourself, etc. to be considered "a user of your product"?

Typically, these are two different sets of people.

So, I don't buy the "always, always" part


I suppose this problem is timeless, back when I was active in the PHP community it was a long-running joke that people who "graduated" to committing to the actual php source (in C) were not doing web development work anymore. And I suppose it was actually true for the majority. On the other hand, designing language features wasn't really related to using it for web work.


Dogfooding doesn't mean you know how to build it. Take for example the graphics engineers on Unreal Engine. They almost certainly know how to fly around test zones or make stress scenarios, but they aren't going to be building large, detailed, open world maps like will be seen in the next Cyberpunk or whatever. And it's not reasonable to expect that of them, either.

Nor are the people doing UI work in Blender going to be able to make Big Bunny.

Nor are the people working on fusion360's parametric system going to build complex mechanical-fluid simulation scenarios.

Flutter is simpler than those, sure, but even in that world you have people that do nothing but fonts & text for their entire careers. They'll know how to do text stuff in flutter, but they'd be lost if you demanded they make a tiktok clone.


I'm a man who once made medical software for pregnant women. There's no point being dogmatic.


I'm sure all those devs working on fighter jet's software would love to but alas...


> This could explain a lot.

Only if you're oblivious to the day-to-day activities of a software project. They work on building a framework, not on using said said framework to build something entirely different that has no bearing in how to build a framework.


I think the experience of building something atop a framework should absolutely have bearing on how to build the underlying framework.


> I think the experience of building something atop a framework should absolutely have bearing on how to build the underlying framework.

You'd be wrong. If your job is maintaining a framework then your focus is on internal details, and how the framework is used would be limited to your concerns in putting up test sets.

Believing that working on a framework gives you equivalent or even similar experience to using said framework in professional settings is a kin to believing that all mechanics are excellent drivers just because they work on cars.


Django comes to my mind here. The people behind the framwork had real experience and needs in web publication.


Commenters point is that this is normal


But not necessarily optimal.


It does, however, have bearing. Despite how common the practice may be in the corporate world, developing a framework without any regard to the user experience thereof is pretty suboptimal


Rails has 12 core developers + 5 committers.

https://rubyonrails.org/community


And 17 users.


I know this isn't a Apples-to-Apples comparison but antirez was the sole maintainer (after creating it) of Redis for more than decade[1]

[1] https://redis.io/blog/thank-you-salvatore-sanfilippo/


PHP Foundation has 10 developers and runs a substantial part of the web. 50 developers are definitely the wrong reason to fork flutter.


Not sure this is a good comparison because the PHP Foundation is "relatively" new and afaik there's no strong correlation to "the php development team" besides these people being paid and also work on it. PHP had a 20y history before this foundation, and the main difference is that it was never steered directly by one big company. Zend, yeah, but they had influence, they had neither started the project, nor ever had a long-stretching majority of people in the important positions. (Disclaimer: have not been an active part of the project for a couple years)


Flutter has much bigger scope than PHP Foundation. It's more fair to compare PHP Foundation scope to Dart Lang and Dart SDK. Flutter scope on top of that is Flutter Engine, Flutter Widgets, Flutter Framework, Flutter Packages, Dev Tools (IntelliJ + VSCode). PHP also support desktops/server environments instead of Flutter 6 platform (iOS, Android, Windows, Linux, MacOS, Web). Making crossplatform UI toolkit is the most labor intensive - PHP Foundation is not developing that. PHP had also 2 decades ahead of development.


50 developers would be definitely a good reason to fork. It's way too much. 1, max 2 would be the perfect size.


Python is an engine, flutter is the whole car.


If Flutter is a car, Python is at least four different models of car, plus three models of light truck and one airplane.

The scope of Python's standard library is just enormous.


true, but still - these libraries can be updated independently, one from each other - with a framework, things have to be orchestrated.


In general in PLs this is not true—the standard library is usually very carefully designed and all of its interactions are well-considered—but Python's libraries do seem to be more independently-updated than most.

Even so, isn't that an argument for a smaller, integrated team for the framework, rather than a large one made of ~1000 volunteers? A tightly-knit framework is much more likely to suffer from too many chefs than a disparate collection of library functions.

Between the parallelization problems inherent in a tight framework and the fact that its total surface area is also smaller than a PL that's maintained by about the same number of people, it sure doesn't sound like Flutter's problem lies in developer headcount.


And Python is written in C, unlike Flutter which is written in Dart.


Flutter engine is mostly written in C++.

Other parts are written in Dart like you said.


I doubt that complexity of a language is comparable to that of a widget library. After the text input cell and scrolling window it’s basically dumb geometry/drawing code all along.


It's much harder to to implement a language in a proper way than writing a widget library.

But that hardness comes from the level of knowledge one must have in different fields such as compiler theory, cathegory theory, hardware, beside classical computer science and coding.

To write a widget library you just have to know how to code and have some experience.

That being said, the amount of work for implementing a widget library might be higher, but is less qualified work being done.

I dabbled in computer language theory and compilers and it's hard. I have no issues reading source code for widget and GUI libraries.


The code behind "intelligent", performant and scalable Tree/Greed widget can be quite big and complex and include concurrent/parallel code.


Flutter has its own language (Dart) which I suspect is included in these stats.


The Dart team is separate from the Flutter team, though they do interact as Flutter is now the primary use of Dart and as such, influences what is worked on in the Dart team.


Then Django or Pytorch are cars, too.


Like a Fisker Ocean?

Launched, abandoned. Scammed again.

When are you all gonna learn?


Heh, as long as the ocean does not cause them to burn - https://jalopnik.com/more-than-a-dozen-fisker-karma-hybrids-...


Even whatsapp had about 50 members when they were handling billion users


> Even whatsapp had about 50 members when they were handling billion users

I would add that running a service deployed to N regions is considerably more demanding and work-intensive than working on a releasable software package.

No one working on Flutter is woken up in the middle of the night by a pager because a user was not able to install a package, to then create a call to pull in half a dozen of Flutter engineers.


Seems to me that as projects mature they require less effort too. It's hard to account for that since

Python started in 1989, it's older than Linux and it has a more narrow scope.


Python doesn't have a narrow scope. At least compared to Flutter. It's in so many places, from Linux installers, to medical, aerospace, math,... and finally the web (which is probably what we hear about the most).


But those are modules added on by developers other than the 50 core devs.


The standard library alone has a huge scope which includes a (mediocre) GUI toolkit.


I did not meant to start a pissing contest, I thought the biggest idea was that Flutter isn't 35yo and because of it, it's more likely to drastically change, and in a fast way too.


It includes another scripting language with a GUI toolkit. The Python part is just a shallow wrapper over third party code.


If only some homegrown GUI toolkits were half as good as Tcl/Tk.


Have you seen the python standard library? It ain't JavaScript kiddo


But does python use protobufs and monorepos and <insert google induced fetish here>?

Maybe they have more time to work on the software because of less overhead...


> This is a weird exercise. Python, for example, is a #1/#2 language in the world and there's only 50 active core devs, 90% of which don't even work on Python full time. Somehow we make that work.

But that leaves Python slow and single threaded.



Besides, it's pure speculation; where's the hard data behind 1.000.000 developers? That's a very large number to be a guess.


Python is really not the same as Flutter. Compare Python to Dart instead.


Yeah... no. The team was recently transferred from the US to Germany, the numbers were estimated around 10 engineers.


I think author put incorrect words into the statement you are quoting. Flutter has much less mature environment, UI packages look good, but most are hot garbage without even unit tests. You can't trust even a single one if it doesn't have at least 5 active developers. They are drying like flies, no one maintains them, updates on time and Dart package manager doesn't help at all, makes things worse by having that system of requirements, versions. I used to work on a side project in Dart and Flutter and quickly fall in love, because i was able to write UI and backend code super quickly, but about 18 months later, I was spending more time maintaining packages that were dropped by original authors, updating, fixing conflicts between versions, and this was the reason I simply rewrote everything back to Java, even frontend to PrimeFaces (: Nothing broke by itself since early 2021.

Even Google dropped 3 packages I used in my projects. Including Dart Angular, they archived it 7 months after promising in AMA they will maintain it for a decade ahead (: They also abandoned their flutter charts library, it stopped working in newer versions of Dart and community had to fork it, they made like 80 forks in a week.

I think this is the problem author is trying to say, not Flutter as the framework, but as the general ecosystem that's consistent and provides at least somewhat positive experience using it. Google is neglecting it.




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: