PWAs, especially those built with an offline-first approach, are so much closer to the native experience than those from half a decade ago. However what concerns me is Google's and Apple's incentives to help them get better to a point where they can rival native apps which could hit their bottom line.
The nice thing about the web is that while it’s no company’s favoured platform, it’s valuable to all of them as a way into each other’s platforms. That’s why whoever is “winning” the platform war a particular point tends to neglect the web. See Microsoft in the 2000s, and Apple now. But all their competitors are incentivised to push the web forward, because it’s a useful fallback as a way to get their apps into the walled garden. Then, eventually things shift and the company that was dominating falls behind a bit, or is facing antitrust problems, and suddenly they care about the web again. In that way, the web is a perpetual underdog, but never goes away.
> The nice thing about the web is that while it’s no company’s favoured platform
It is effectively Google's favored platform. They effectively get to be the gatekeepers for what standards become accepted. They control the most widely used engine, and fund the only opposition so they have an incentive to avoid going against them. Just look at how everyone dropped the ball in the recent JPEG-XL vs WebP news.
This. Apple's Safari (on iOS and macOS) is the Microsoft IE of 2020s. It's unbelievably behind the more sophisticated browsers (Firefox, Chrome) and yet, Apple couldn't care less.
It's not IE by any imaginable standard. And it's only "unbelievably behind" if you count all the non-standard APIs that Chrome ships by default and calls them standard.
A browser that doesn't support extensions unless they're installed as separate apps on the platform is not worthy of the name "browser"; it's a lock-in mechanism.
> what concerns me is Google's and Apple's incentives to help them get better
I think that despite that, they did a decently good job. A few years ago, I created a voice-only social app: people can easily record and send voice memos.
I built it using fairly new tech at the time: the web audio api and native web components. (I refuse to touch react-style frameworks with a 20-foot pole.) It was 99% compatible with all major web browsers from the get-go, even with its fairly advanced "tech."
The only thing that wasn't cross-platform was the audio files themselves. I had to make a deep-dive into the world of encodings and codecs so that I can offer users audio files in the formats that work on their machines.
Obviously, there are areas where cross-platform-ity (??) is abysmal, but things will probably continue to evolve for the best ― hopefully.
And yes, PWAs with their offline capabilities are just awesome!
> PWAs . . . are so much closer to the native experience than those from half a decade ago.
On the one hand I love this. I'm glad PWAs are helping drive people who wants and expect a certain thing towards the web.
I am also however super sad that PWAs removed so much greatness of the web. Most don't launch in the browser. None have visible url bars. They just added a new mode so PWAs can have some kind of tabs. They don't integrate with bookmarks, or extensions. They don't have forward/back. I like the web a lot, and I feel like PWAs are a major step backwards for me.
What's really scary to me js that new permissions are being added which are PWA only. I no longer can have a web like experience; I am forced to have an app like experience, without any of the affordances I get with a user agent/browser at my back. I feel so undermined here.
My understanding is that there is a lot of political power inside Google (from the Chrome side of things) working to ensure that PWAs are the primary way that users interact with their phones. It's definitely a point of contention, with pressure on both sides, but there is a balance there.
Apple on the other hand is only incentivized to funnel everyone through their App Store, and for some ridiculous reason, app developers use this to justify going all in on native apps, which I think is what this article is addressing.
One really good thing about native apps versus PWAs is that once the user installs a native app and learns the way it works they don't need to learn it again until they upgrade. With anything sourced from the web, the user is at the mercy of the wevdevs, who are at the mercy of the marketing and merchandising wizards who want to shake up the UI to try and goad users into buying more stuff. It's kind of like those round baskets that supermarkets like to clutter the aisles with; somehow they think it prompts you to buy things, but for me it adds one more product to my "no, not buying this ever again" list.
not my experience. Apps update automatically for most people so they are no different than web pages. You can turn off the updating and then in 2-3 months you'll get a message the app no longer works with whatever service it's tied to and you're required to upgrade.
On my desktop there might be a few apps that don't need to upgrade but on my phone, 90% of all apps are tied to a service. Uber/Lift, bank apps, apartment apps, hotel apps, bike rental apps, social media apps, dating apps, streaming music apps, video apps, etc.
Yes because for example I'm sticking to K9 mail 5.6 because of the UI change in the next versions. I picked K9 exactly because of the original UI so I'm not using the new one, which is like every other mail client. I saved the APK and I'm installing on any device of mine, also when my fingers slip on the update button. Uninstall and install 5.6 from the file manager.
And no, because if the company behind the app wants to force people to upgrade they'll make the backend incompatible with the old app. Also no, because I don't think that many normal people invest time not upgrading an app because of the UI.
Right. Google is incentivized to push PWAs and other forms of cross-platform mobile app development (hence Flutter).
How often will the hot new app from some SV startup launch for iOS first with no Android app? Resource strapped companies that have to prioritize one platform will tend to pick iOS and that hurts Android (but is obviously good for Apple)
Apple has a giant advertising business, but yes, it's not on the web like Google's is. It's on their own app properties like Apple News and the App Store.
Google and Apple should be regulated to offer the same functionality and performance to the web. Their grip on the most important technology of the century is astounding and harmful.
Honestly first-class web mobile installs without hoops and scare walls would be the thing to strive for.
Google and Apple, both of whom make web browsers as well as OSes and could therefore cheat loads with anything like that. Of all the places to find someone conflating the companies with the operating systems Android and iOS.
But even then…
> same functionality
> Honestly first-class web mobile installs without hoops and scare walls would be the thing to strive for.
Coming to a future near you: npm install rm_minus_rf_slash
What's the old quote?
"""
Tech Enthusiasts: Everything in my house is wired to the Internet of Things! I control it all from my smartphone! My smart-house is bluetooth enabled and I can give it voice commands via alexa! I love the future!
Programmers / Engineers: The most recent piece of technology I own is a printer from 2004 and I keep a loaded gun ready to shoot it if it makes a strange noise.
"""
Except all the banks and railways require me to have a smartphone these days whether I want it or not.
I'm not aware of any scare walls on iOS. On Android is a site is installable a little banner is supposed to pop up and ask if you want to install it. On iOS one has to open the site's share sheet and pick an install option nestled in there somewhere. I suppose that counts as a hoop, but how else would you surface the functionality without it being invasive and annoying for most users?
I think a more pertinent problem is that most devs don't want to distribute their apps as PWAs because Apple deletes saved data at unpredictable intervals and PWAs can't push engagement notifications as effectively as native apps.
When I switched from the Uber PWA to their native app I suddenly started getting bombarded with ads from them in my notifications.
> devs don't want to distribute their apps as PWAs because Apple deletes saved data at unpredictable intervals
Does anybody know what the current situation in iOS looks like? I researched this recently and the information is conflicting. The OPFS data (Origin Private File System, which is the new storage API for PWAs and is supported in Safari) is supposed to not be deleted without explicit user confirmation, but what this confirmation looks like is unknown (is it just the usual delete Safari history button?) and I didn't run any experiments. However I've seen someone claiming Apple separates a website and it's installed PWA's storages. This is reinforced by the fact that I can see the storage space each PWA takes in Settings (which was oddly not updated when a PWA using OPFS used some more storage). However I just deleted that PWA but there were no special prompts, just the usual native app deletion prompt, absent of the usual phrase "Deleting this app will delete its data.", only the phrase "Do you want to delete this bookmark." was present.
Those are downloads though the app stores. The links merely funnel you though the stores.
Native web downloads of apps are impossible on iPhone/iOS. You have to go through the app store.
Native downloads of apks are possible on Android, but not until you navigate to the hidden settings and enable them. And even then, Google scares users from installing apps this way. Nobody in practice does this. It's effectively not permitted.
> but not until you navigate to the hidden settings and enable them
This used to be the case for older Android versions, but isn't anymore. Currently any attempt to install an APK from an app without permission to do so will pop up a missing permission warning with a deep link into the relevant settings page with a switch you can flip. Flipping the switch will initiate the installation of the APK. I can't imagine sideloading being any easier without removing the permission barrier entirely, which seems ill-advised.
