Notices by Alex Russell (slightlyoff@toot.cafe)

Apple's trying to scuttle PWAs out of pique, blame regulators for their tantrum, and gaslight users, businesses, and developers about who is breaking their apps and why.

It's bullshit, and @owa is fighting back. Join them by signing the open letter:

https://letter.open-web-advocacy.org/

What Apple is going to do to PWAs in less than two weeks in the EU is dastardly.

Their goal is to remove the only versions of Push, Fullscreen, Homescreen Icons, and Badging that didn't require paying them to be in the App Store *and make sure no other browser can have them either*.

They tried to sneak it under the noise of the alt-store drama, in a clearly premeditated fashion.

It's no exaggeration to say they've gone to war with the web & we don't owe the benefit of the doubt ever again.

A decade ago, a tribe of JS partisans took the web by the reigns, forked HTML and JS syntax, and yeeted useland abstractions into the critical path because "a better user experience".

This was premised on the idea that everyone's CPUs/networks would get faster the way their top-end phones did.

They could not have been more wrong.

JS-first web development has been a planetary-scale exercise in the rich making life harder for the less well-off.

https://httparchive.org/reports/state-of-javascript#bytesJs

https://infrequently.org/2022/12/performance-baseline-2023/

Lordy:

https://www.techdirt.com/2023/05/22/apparently-elon-doesnt-think-he-needs-to-pay-rent-because-sf-is-a-shithole-so-why-should-we-pay-for-twitter/

@aral cool; you're blocked.

If you make websites for a living, know that you are *extremely likely* to be wealthier than the people on whose devices your code will run. This matters to *how* that code runs.

Let's say you're a webdev in the US, building for the US market exclusively. There's huge variability in cost of living across geos, but even so, Indeed shows a an average base salary for a webdev of ~$80K/yr:

https://www.indeed.com/career/web-developer/salaries

Median* household* income, meanwhile, is closer to $70K/yr:

https://www.census.gov/quickfacts/fact/table/US/SEX255221

Now, let's say you decided to rebrand in the Lost Decade as "JavaScript developer"; you feel pretty comfortable packing those webs -- what does it do to the equation? A LOT:

https://www.indeed.com/career/javascript-developer/salaries

When I go on about how JS cultists have cost companies more for worse outcomes, this is what I'm talking about. A 35% cost premium, which pushes the income of these folks *as individuals* to be ~50% higher than most households, and more than 2.5x the median dual-income earner.

This might not sound like a big gap, but recall that *most* US households are dual-earner:

https://www.pewresearch.org/ft_dual-income-households-1960-2012-2/

This means that you *very likely* make 2x or more than *most* of your adult users.

But then there are kids and the elderly. If they're potential users, they'll skew income data further.

This is, in part, due to differences in replacement rate. If you, as a developer, have a laptop that's more than 3 years old and cost less than $700 new, odds are you're pretty anxious about it. Can't pack many webs on that.

But as a user? That's just *the average*. Real-world gets slower from there. Edge telemetry shows that something close to half of our users are on HDDs, not SSDs.

When was the last time *you* booted from spinning rust?

A 2.5x difference in earnings is a huge difference in life experience. Developers may not realise how different those experiences are because they don't go looking for them, but it's in their data if they cared to investigate.

*Most* users do not have fast laptops or phones, even in the US.

Lower income also correlates with increased network latency.

In the US, this is maps onto persistent income and wealth disparities between white and (specifically) black communities. Digital redlining is now A Thing:

https://themarkup.org/still-loading/2022/10/19/dollars-to-megabits-you-may-be-paying-400-times-as-much-as-your-neighbor-for-internet-service

When data is expensive and slow, you use less of it.

HDDs correlate heavily with lower core counts, and generally signal slower (older) cores, all of which have a big impact on desktop performance related to your web apps.

Browsers are heavily threaded today, and fewer cores means less performance on JS-heavy sites *in particular*.

Less expensive phones, meanwhile, tend to feature MANY cores, but the also tend to suck. A lot:

https://infrequently.org/2022/12/performance-baseline-2023/#devices-1

All of the data in this thread is from the US where inequality is a problem, but is not at its worst worldwide:

https://en.wikipedia.org/wiki/Gini_coefficient

Technology has broken down borders in many ways, which now means that the merely 2.5x wealthier JS bros in the US may make 20-50x more than users in other geographies:

https://worldpopulationreview.com/country-rankings/median-income-by-country

PPP is a thing:

https://data.oecd.org/conversion/purchasing-power-parities-ppp.htm

But even with that caveat, $108K/yr US income is still 6-7x the effective median salary in India.

The digital divide has a colour, and it is being excavated and widened by the actions of a well-paid upper crust that bear none of the consequences of their actions.

Wealth acts as a blanket, keeping those wrapped in it from feeling the chill as they open the windows and turn down the heating for everyone else.

...and that's before we consider the lower price differences of electronics vs. services and other sorts of local goods.

Phones and laptops are *expensive* in much of the world, which is why Android accounts for 80+% of device shipments and the average selling price, worldwide, continues to hover near $300 (new, unlocked) even as high-end devices accelerate in price. There are just that many more sub-$300 devices being sold for every trade-up of a wealthy user from a $400 to a $1K phone.

Which brings me to the argument that *really* makes me wince; the one @seldo et. al. seem to be falling back into: that everyone knows what they're doing, so it's fine.

This bears no resemblance to the conversations I've had with teams.

Managers and TLs are universally shocked and upset to learn their JS-bro-special stack is losing them tons of money that the "legacy" site didn't. Further, they have none of the collateral one would expect if it was explicit (a financial model).

So when you export technology that was barely working in the US to other markets, it's a *disaster*. And the JS bros take no responsibility for the predictable (and predicted) consequences of their actions.

Instead, they have continued to spread the (always false) narrative that CPUs are getting faster and that people can afford the JS they're sending.

The "it's fine, we know we're losing users with these technology choices" crowd *really* needs to start showing their work.

What is the expected revenue gain in wealthy user segments per KB of JS?

What is expected revenue loss per KB of JS in other segments?

What does the distribution look like?

Just read @seldo's (well written and argued) response to my post from last week, and while I owe it a larger airing, a few things jump out:

First, "SPA" vs. "not-SPA" is about data model locality, and I'm not sure that sites that could easily be built with progressive enhancement and Ajax partials count.

See the diagram from the end of my longer post:

@seldo OTOH, I think it's fair to say that this has given me more direct and high-bandwidth access to the decision-makers and the narratives they have used to justify their choices.

Laurie's model is so far divorced from this argumentative and marketing morass that I do not recognise it...except in the effect of shrinking the TAM.