All up, I say “two major versions” is obviously a completely unsuitable definition to use. It needs to treat Safari differently, e.g. “two major versions of Safari and twelve major versions of all other browsers” (which would catch Firefox ESR) or “thirteen months of browser support” (which would be less for Safari most of the time, but really, consider the deployment cadence and it’s probably not so different, though maybe you’d bump it to fourteen or fifteen months, I dunno).
So if Safari is adding a slew of major features once per year, and Firefox 6x per year, it doesn't matter. We have a baseline of what features as developers we can now implement without worrying about which browser version supports this or that. We see the baseline, we add the feature.
Last two major versions has been utilized (unofficially) by most people through Browserlist. It's been both suitable and productive.
In my experience, people haven’t tended to use just “last two versions” with Browserlist: IE has normally been special-cased historically, and Safari has regularly been special-cased in some way, and Safari has actually been the salvation of the “last two versions” strategy because it dragged the baseline down to a more acceptable place, where “last two versions” of only Firefox and Chromium would regularly be disastrous. It has long been a really poorly-thought-out mechanism.
The other point you’re missing is user adoption. “At least one year” for Safari is one thing, but “five weeks” for everyone else is pretty crazy when you look at actual update statistics.
I'm hoping that we'll be able to identify better usage data and a better understanding of developer expectations to refine this definition.
What's worse is people copycat this "baseline", and if you try to argue sanity, they just say "sure it's what Chrome/Firefox recommend" without knowing or caring that they are only supporting months old browsers now. It's madness.
Chrome 109 is the last version to support Windows 8.
There are a lot of people on systems even older than that.
But internet-connected devices are a special case which deserves consideration.
Unfortunately Operating Systems, and the software that runs on them is imperfect. People are spending a lot of energy finding, and creating exploits based on, those imperfections.
An old car can be used almost infinitum, if there's any danger it's likely to be to itself, or at worst very local to others.
However an exploited old machine on a LAN, is dangerous to the entire LAN. And beyond.
There are hardware iterations that create a specific cut-off, but I don't think Windows 10 was one. (As i recall, if it could run Windows 8 it could run Windows 10 AND the upgrade was/is free.)
Yes there is -really- old hardware out there running XP, even DOS, but I'd argue its not something you should expose to the Internet, even as a client.
Running old browers, and related software, to browse the modern Web, is a very risky game with a very large penalty for failure.
In that context I'm happy to build sites targeting "current software".
If it is e-waste you are worried about you can always use Linux to keep your older machines useful as that both allows them to run better than Windows with a lightweight DE while also using the latest versions of the available software with all the security fixes.
You also get to have more control of how your computer will behave too.
This is also why XPs complete ESR support ended in 2019. What I just said about people also applies to certain organizations like governments, who'd rather pay through the nose for that ESR support than upgrade.
That's just the reality of it all. There's a reason why Android is so extremely pushy with installing updates to Chrome et al. in the background (and why Androids old OS depreciation process is so slow) - users just flat out don't update their software or their hardware if they don't feel the need to.
EOL from a company means nothing. People will absolutely stick on years old hardware if they can help it as long as it works for them, and you're going to still be the one on the hook unless it's truly ancient tech. (We're currently at the point where if someone uses Vista, you could probably tell them to update? Keep in mind that Vista got EOLed in 2016. Windows 7 is kinda heading in that direction, but I'm not seeing that happen until another 2 to 3 years have passed at least).
Can't be too careful.
The argument of “this doesn’t cater for the super-minority of nerdy power user ideologues”, that’s honestly not going to roll commercially, sorry. If we are talking about the regular adoption curve, enterprises holding things back, or literally just…anything more than you and RMS, I just don’t see that as remotely relevant.
I don’t even know what “two major versions” means. Or why it would be reasonable to consider as a baseline.
For Safari: we’re currently at 16.4 (March 2023); does “two major versions” mean from Safari 15 (September 2021), or from Safari 16.3 (January 2023)? I’m going to guess 16.3.
For Firefox: so this just ignores Firefox ESR? It’s at 102.11.0, while regular is at 113.0. (That is, it’s at nine versions/nine months ago.)
For everyone else: you know that’s permanently as little as about five weeks ago, right? Put bluntly: that’s stupid, nothing like what a baseline should be.
By the sound of it, this may be genuinely ignoring 25–35% of users, and declaring features “baseline” not supported by the browsers used by fully 15–25% of users. (Numbers taken from quick estimation based on https://caniuse.com/usage-table which mostly uses StatCounter’s woefully misleading and wrong numbers because I’m lazy, but I think my error bars are large enough that they’d hold up to more rigorous calculation using other data sets too.)
My only hope here is that “two major versions” means Safari 15. Then all up it’d mean roughly “since September 1–2 years ago” rather than “mostly 5–10 weeks ago”, and that 15–25% would drop to under 5% for most features in the delta, half by simple coincidence.
(Aside: I really wish browser makers would switch to some form of calendar versioning, e.g. YY.MM. There was a great opportunity about now when Firefox caught up to Chromium due to a slightly faster cadence.)