Spoiler: unlike so many blogs, this actually contains original content, cf the title question.
Go to the salient point of this post here: [0]
<Intro: NOT Original Content>
Hi fellow human,
You might have seen ad-nauseam this recurring theme on (e.g.) Hacker News that once Chromium will truly dominate the web marketshare, it will become a Google-only monoculture project unlike every other single de-facto implementation projects (Linux, QT, GTK, rustc, blender, libreoffice, etc).
Something people fail to understand is that beside some spec work, actors such as Mozilla and Apple have zero utilitarist impact of chromium's users. Mozilla generally does not improve at all the privacy or performance or feature richness of the engine that people increasingly use, and the day they will become irrelevant (following marketshare trends of a mere 3.26%) their lasting impact on the codebase of the lasting browser engine will be of a striking 0 lines of code.
The actor that will care about the web and improve the experience for actual end users (not a 3.26% niche) and protect important rights such as maintaining a manifest V3 ublock-compatible variant, could be mozilla and they would become relevant again, a place that could be taken by Brave browser eventually. There exist also an alternative future about Mozilla where it would remain Gecko but hybridify more and integrate some useful chromium's modules/components (e.g. Spidermonkey is a net cost and has lost its main value propositions against V8 a decade ago, in return Mozilla could concentrate on supporting newer web capabilities (which is is absolutely remarkably lacking (in case you were out of the loop)) or even better, contribute to V8 and backport some good spidermonkeys ideas to bring a best of both worlds.
A converse hybridisation is possible, such as contributing and adapting the amazing Stylo engine to chromium.
I already presciently outlined a detailed plan years before the fall of Servo
But this intro is not the topic at hand and one should not go past the short attention span of HN readers, so let's jump to the Original Content:
</Intro>
<Original Content>
Since Mozilla nor Apple nor any relevant actor decided to actually care and invest into diversifying the contributions ratio of chromium, that burden has been left to the very Rich (and #1 worldwide open source contributor) Microsoft.
The salient question is: What share of Microsoft devs have on Chromium commits ratio vs Googlers/Chromiumers?
So I ran this experiment:
Clone Chromium's ~40GB github repository
and run
git shortlog -e -s -n --no-merges > .\Results.txt
According to stackoverflow this give us a list of ALL git email addresses, ranked by total commit count. Please leave a HN comment if there are flaws with this method.
A control find of @Microsoft give us 270 distinct email addresses. For comparison @google give 4000 result, although contrary to microsoft some are duplicated so a minimum of >2000 addresses.
and there are 8000 @chromium addresses, > 4000 with duplications.
there are 4 mozilla addresses, 7 from @collabora (would have guessed a bit more) and ~74 @webkit addresses (most probably being from the pre-blink fork?)
It's major to note there are 20000 @ symbol in the file (with some duplication) and that there are some important actors (Intel at 384) and 900 unattributed @gmail addresses.
I have no clue why there are some @google addresses, are most of them old/dead accounts long replaced by @chromium addresses? And are there @chromium addresses that do not comes from Google? Does Microsoft has any?
Iff we assume by the negative, that's ~ between 4000 to 6000 Google developers vs 270 Microsoft developers.
A tragic flaw of my methodology is that we have no information about which accounts are active and which accounts stopped working on chromium 10 years ago.. Heuristically, Microsoft addresses are very recent and therefore active while Google and Chromium addresses do not necessarily reflect the present, IIRC at least that was true a few years ago, Google was continuously hiring more chromium devs so one could believe Google has between 1500-5000 active developers vs 270 for Microsoft.
270 fulltime paid Edge Blink developers is underwhelming, Microsoft should hire more engineers to make the web a better place, as an extremely potent meaningful goal.
However 270 fulltime developers should be enough to have some major positive impact, including the resources for maintaining a Manifest V-3 ublock compatible variant..
If we look at the facts, Microsoft has made many important contributions to Chromium both regarding performance (e.g. moving input handling off-the main thread) and features (CSS-subgrids, etc)
However the most salient point of this post [0] is: Where are all the Microsoft Edge devs?!
OF THOSE 270, only 86 have more than 15 commit and only 44 devs have more than 40 commits.
I bet those "commits" must be squashed and therefore account for the count of merged Pull Requests and some devs might have made many commits in a long running branch but still, that means that only 86 devs worked on more than 15 features/sub-feature/bugfix.
</Original Content>
<Conclusion>
: I think it is likely that my method for counting an account total commits/PRs is flawed ? Please leave a comment.
- If it is not then 86-44 full time devs is really low Microsoft and then if so, you could and should do better.
- What are the 184 remaining developers doing? It just doesn't add up.
- either way, 270 is a nice start but I hope Microsoft will invest more, and publicly address the Manifest V3 ublock limiting issue
- Reality is that if Mozilla does not hibridize they will fall into irelevancy. Not realizing the current hybridization (Angle, Skia, etc) is ignorance at best, hypocrisy at worst.
- The web is the most user impacting platform on earth, google should not just be scrutinized, they should also be praised for being the ones that cares the most about improving the platform, something Amazon, Facebook, Netflix and so many others could largely, trivially afford but it seems they don't care about the end user.
note however that I did the same for the webkit repo:
there are 97 @webkit addresses, with 77 of them above 15 "commits", and 65 above 40.
344 @apple addresses with 183 above 15 "commits" and 148 above 40.
340 chromium addresses pre-blink fork and 105 google addresses.
It means there was more google devs in webkit before Chrome version 28 than there are todays (or all time) Apple devs on webkit. And keep in mind that I'm cloning latest git in 2022 and that Apple has recently (doubled?) it's developers team (still far too small though).
I can't say about active accounts, but in all time addresses, Apple has tragically more invested in the web than Microsoft currently is, unless of course there is a flaw in my methodology.
I believe those are some great points Alex Russel, Edge's product manager should answer, browser Human's resources investment matters, as it is an investment for humans. And checking the epistemic veracity of blogs too.
Bonus:
Chromium is not as open as it seems, many design documents are locked and you must ask for permission. Usually that permission is granted in the next 24 hours but this adds friction for what should be open source documentation. Secondly, I had found some specific URLs that redirected to google internal websites that are not available to us, this should never happen. This is a concrete and easy issue that chromium should fix.
Fun fact:
Apple is so miserable/pathetic Google had to contribute to them WebGL 2.0 support, and the fact Apple deprecated OpenGL (and OpenCL because hey science/AI is not trendy enough /s) did not help.
Note: The --no-merges flag does not seem to yield different results.
Source of Chromium email addresses:
https://justpaste.it/edit/48034590/463shvy3b02g1htm
Source of Webkit email addresses:
https://justpaste.it/edit/48035787/i9qktz2fk6o2gzjb
> and OpenCL because hey science/AI is not trendy enough /s
This is an euphemism for not saying they are degenerates obscurantists and retrograde human beings. There I said it. Oh and BTW they probably forgot they are the ones that invented OpenCL in the first place, turnover must be high?
</Conclusion>
Comments here: https://news.ycombinator.com/item?id=31865870
If you are not from HN, please tell us from which circle you come from :)