Update, 2023-12-27: The project described herein has now concluded. Please see “Testing, testing: ending a fool’s errand,” for an explanation.
My recent spate of browser-hopping while trying to base each such choice on The Data™ has, of course, required gathering data. And, because the world of browsers involves frequent changes, that’s become an ongoing task. So I decided to version-control it and put the results out there for all to see.
The public form of my new browser-tests repository, always very much a work in progress, will show you my latest findings. There’s one page for macOS and another for Arch Linux. Here’s an alphabetical list of the browsers I currently am including in the tests:
- Arc (macOS)
- Brave (macOS and Arch)
- Google Chrome (macOS and Arch)
- Microsoft Edge (macOS and Arch)
- Mozilla Firefox (macOS and Arch)
- Orion (macOS)
- Safari (macOS)
- Ungoogled Chromium (Arch)
. . . a list which can and will change over time, depending on several factors which, themselves, will remain somewhat in flux.
The repo also has a pre-git area where you can see what I did before going Full Git with this thing. In short, it was a growing set of text files (including those on which I based the earlier “Testing, testing”) and involved a lot of needless repetition. I realized that, in the end, the best bet is to keep these two Markdown files and update them as I do new tests. That way, there will always be one source of truth with the latest data for each config (i.e., macOS-on-Mac-Studio and Arch-on-iMac), yet with a clear Git trail to earlier versions — at least, that’ll be true once I actually add to it.
I made the repo public mainly because I suspected that some of my readers would be similarly interested in such data but didn’t want to spend their own time running the tests, especially not every time there’s a need to re-do them. So, there ya go. Enjoy.
Update, 2023-12-16: I later decided to do my Linux testing in Fedora rather than Arch, due to the easier availability of truly official browser versions for that distro. The resulting page will be here.
A follow-up to “Firefox on the brink?”
I would be remiss if I didn’t acknowledge that my previous post got a lot more attention that I’d ever dreamed it would.1
First, it got mentioned in the 2023-12-04 edition of the Changelog News podcast. For a few hours, that produced a decent spike in my site’s numbers, and I figured that would be the end of it. Oh, my, was I ever wrong.
The next day, somebody — perhaps someone who’d found the post through that Changelog News podcast? — submitted it to Hacker News. It ended up on the front page and stayed there for hours, long enough to gather hundreds of comments and, not incidentally, generate tens of thousands of hits to the post.2 (As I’ve learned ruefully in the past, certain topics really trigger the HN crowd. I’d already been aware that Tailwind CSS is one of them; and, it now appears, another is the decline of Firefox.)
Note, 2023-12-17: I later learned that the article also got picked up in multiple subreddits (e.g., /r/webdev/) and on Slashdot.
Thanks to anyone who started following my feed as a result of all the attention that “Firefox on the brink?” amazingly received. I hope you consider sticking around. And, just for the record: I actually like Firefox very much; I just fear for its future due to the reasons I cited, all of which have to do with real-world facts and policies, not anyone’s fleeting feelings.
Although I no longer use a separate analytics service (because the small size of my usual traffic doesn’t merit it), I still can see relatively accurate numbers from the anonymized server-side analytics that come with using Cloudflare’s free tier. At least, trends are pretty clear. ↩︎
Indeed, my post’s venture into semi-viral status actually cost me a little money. The resulting “invocations” of the Cloudflare Pages Function in front of my site blew ’way past the free tier’s 100K-per-day limit, forcing me to drop a few bucks to go to the next tier and keep the little sucker running. ↩︎
Latest commit (12aed64a2) for page file:
2024-05-04 at 10:49:37 AM CDT.
Page history