I'd say 50 or so people would be fine.
>How big do you think the Chromium team is?
Around 60-80 people judging from the names listed under the various Blink teams (Rendering, DOM, Memory, Style, etc).
With 50 or so devs (let's forget for this example about managers, UX researchers and designers, HR, etc.) you'll get maybe a JavaScript VM and a small UX.
Not nearly a browser :(
> Around 60-80 people judging from the names listed under the various Blink teams (Rendering, DOM, Memory, Style, etc).
That sounds like a really, really vast underestimation. To the best of my recollection Chromium embedding teams inside Google that are 30+ developers (again, let's forget managers, UX researchers, etc.). I know that there are at least 4 such teams at Google.
I would be very surprised if Google didn't have at least 1000 developers working on Chromium.
This list has 400 pages of contributors (~8000 people). Even with a very conservative assumption in that only 10% of them are full time developers, it's still 800. This doesn't even include other derivative projects and non engineers.
Safari does those two specific things with a quarter of the number you mentioned. The entire team is nowhere near a thousand people.
git log --since=2019-01-15 | grep "^Author:" | fgrep chromium.org | sort -u | wc -l
If I git clone https://github.com/WebKit/webkit.git as of today (rev ba925cdbc8c3f2dff44cdcb92d9a374816b0215b) and run:
git log --since=2019-01-15 | grep "^Author:" | fgrep webkit.org | sort -u | wc -l
If I hg clone https://hg.mozilla.org/mozilla-central/ (the repository for Firefox) and run:
hg log -d 2019 | grep "^user:" | sort -u | fgrep mozilla.com | wc -l
Note that we may be undercounting QA here, since not all of them might commit to the main repository.
On the other hand, we might be counting one-off contributors who are not really on the "browser team" per se, especially for "apple.com" and "google.com". We're also overcounting somewhat depending on how much churn there was over the course of the year (people leaving, new ones joining).
But pallpark, I suspect an unrealistic lower bound is 100 and a reasonable lower bound is 150-500, depending on how much of the tech stack you want to delegate to other entities. The Chrome team is a lot bigger than any of those numbers, of course.
Opera's long gone, now it's a chromium fork. Internet Explorer is gone, now MS uses a chromium fork. The hype new browsers like Brave and Vivaldi are just chromium under the hood.
It's like the difference between making a Linux distro and maintaining a full OS.
FYI, WebKit itself takes hundred of engineers from Apple (which would be roughly similar to Blink). And this is only for the rendering engine, which is pretty small compared to the entire browser codebase. Thus Apple is investing a comparable amount of engineering resource into Safari. Where are "Other browsers"?
Just writing a javascript runtime alone isn't just a "handful of engineers." WebGL stack? WebRTC? Layout engine/compositor? Notifications? You're kidding me, there is no small team in the world that could ship a browser stack end to end.
But it is. You can verify it yourself: https://github.com/WebKit/webkit/commits/master/Source/JavaS...
> You're kidding me, there is no small team in the world that could ship a browser stack end to end.
https://webkit.org/team/, ⌘F "Apple". Balance the people on that list who have left or are assigned to work on something else with those who aren't listed there.
WebKit does not have hundreds of Apple assigned to it.
> And this is only for the rendering engine, which is pretty small compared to the entire browser codebase.
Not really.
WebKit runs on macOS, iOS, iPadOS and watchOS across Intel and ARM architectures.
WebKit provides the web views for countless 3rd party apps, including Mail, Calendar, iTunes, etc.
Apple certainly has fewer people who get paid to write code for Safari/WebKit than Google has on Chrome/Blink. I wouldn’t be surprised if Mozilla has more people too, especially since they’re rewriting pieces of the browser engine at the same time.
Well, it's close to 100, according to https://news.ycombinator.com/item?id=22059393 ? You can bring up with other numbers to refute this.
> Not really.
Please don't assert this so lightly unless you have any evidence to support it. A rendering engine is just a tiny fraction and you gotta take care of literally thousands of other components to build a modern browser. This applies to Chrome, Firefox and even old good IE. I don't expect any valid reasons why the same logic cannot apply to Safari.
https://www.chromium.org/getting-involved/become-a-committer
They've managed to actually get GMail to render - a not insignificant task: https://www.ekioh.com/devblog/full-google-mail-in-a-clean-ro...
The timeline on this page alone is in years: https://www.ekioh.com/company/#team
From what I understand, that team is relatively small in comparison, but actually does have this widely-ish deployed. So it theoretically could be done with less... but it's still insane to even consider. This isn't simple, and anyone who's trying to imply otherwise is wrong.
The other commentator you're responding with also never discloses they worked with Apple previously, while pretty much endlessly pumping up their work here. shrug
Not so sure. I remember the Webkit guys being a very small team (and they basically did the whole of Safari). There was some such mention on Dave Hyatt's blog at some point.
And, as far as the "chrome" part (UI, settings, etc) goes, wasn't Firefox at first the work of a couple of people, who forked their own UI version of Mozilla? And still it got to be the most popular browser at the time.
Not to mention how whole OSes and other challenging things have been done by smaller teams...