> Could we get there? Absolutely. We just haven't yet.

What else is needed then?

Take a look at the known issues section regarding iPhones. It’s good evidence of apples non competitive behaviour regarding browser support. PWA/websites are not allowed to be good on iPhones.

Their libraries sometimes don’t even work for low scale though.

The protocol buffer compiler for Swift actually at one point crashed on unexpected fields. Defeating the entire point of protos. The issue happens when only it tries to deserialize from JSON, which I guess none of them actually use due to large scale.

To clarify, I'm not thinking of code/libraries written by a huge company, more about open source code that has been scaled far beyond your deployment size by someone/anyone else.

Also, if you're using some feature that isn't regularly exercised (like your Swift protobuf example), it's probably doesn't have the variety of use to be covered by Hyrum's Law (see https://www.hyrumslaw.com ), which is definitely a different aspect of the Ubiquity criteria.

It’s death by a million papercuts with safari.

I made a reader app for learning languages. Wiktionary has audio for a word. Playing the file over web URL works fine, but when I add caching to play from cached audio blob, safari sometimes delays the audio by 0.5-15 seconds. Works fine on every other browser.

It’s infuriating and it can’t be unintentional.

My workplace is a bird conservation non-profit with java code from 2006 and the website is many jsp files.

I've added a react SSR system. It has node subprocess code for rendering HTML from java via stdin/stdout. There's a Node/Vite proxy server that adds the fancy HMR you expect from SPA apps.

It supports multiple roots on a page, every SSR component has data-props and data-componentname, and the entry script just queries those attributes and hydrates everything.

The node renderer script is packaged as an EXE which is deployed in WEB-INF on the server.

It's fun to add the amazing React tooling to an old codebase. It also shows how you really, really, really do not need NextJS.

It’s not easy on iOS either.

For me to get to it from App Store, I have to click a tiny profile pic (smallest allowed button on iOS?)

The next screen displays Purchase History and other items. However, no way to cancel from this page (insane?)

To get to subscriptions, I have to click my name at the top of that page (which doesn’t even look like a button) which loads for 3 seconds then pops up.

On this hidden account page, it shows Purchase History along with subscriptions list to cancel.

If any other site hid subscription cancelling behind a flat contact header secret account page, it’d be an issue, so yeah it’s an issue for Apple too.

Getting there via the App Store is just one option. You can also get there from Settings under Apple Account > Subscriptions.

And if that’s too tricky, you can just type “Subscriptions” (shows up after just typing “sub” for me) into Spotlight Search, and it’s the top option.

What version are you using? Click 'Settings' -> 'Profile' (huge button on top) -> 'Subscriptions'. I don't know how it could be easier than that. Ah wait, pull down on the home screen for search, type 'subscriptions' and tap on the result for direct access to the setting. From there you can see and cancel any subscription made in the App Store.

At a certain point, C++ compile time computation becomes something you really can’t do in C. https://codegolf.stackexchange.com/a/269772

Title is incorrect. 178 days is the walking time of the optimal tour, not how long it took to solve for the best route

3 months, using 44 CPU-years, is that time.

(Submitted title was "Shortest walking tour to 81,998 bars in Korea — TSP solved in 178 days".)

Haven’t got through the whole article yet. The beginning was slow for me, but it did start to get interesting quickly.

A tool for learning German via reading. The main feature is click a word to see its dictionary entry (not always easy with so many declensions). Flashcard creation from the dictionary entry. Real human spoken audio from Wiktionary.

Will be working on some LLM features soon (explain "X" in the sentence "Y" prompt). It'd be nice to get a word frequency ranking dataset integrated, so I don't waste time learning rare words.

I’d check this out