The web is not poor man’s native (2015)

PWAs also fit in a niche that is difficult for most organizations to digest. It's fundamentally front end tech, but service workers are non-trivial and also are decidedly outside of the typical front end dev's experience. They resemble backend, but the life cycle is somewhat complex and difficult.

Out of the box frameworks can provide basic service workers, but that doesn't make a PWA really shine.

It's a powerful but unusual tool that's slotted into a strange place. They are rarely used even though they can significantly improve user experience without going full PWA.

Been working in back end and front end for years. The number of front end devs I've met in the wild who understand service workers, their use cases, and how to build one, I can count on one hand.

> Users rarely, if ever, install apps

My local car wash now has an app and my wife promptly installed it, because of some vague rewards tracking.

The vast majority of mobile usage is through apps. That's a fact. This is primarily because of the low bar to adoption (click a link from a QR code/click an icon) paired with the expectation that the experience will be better than a website. If the previous website experience was bad, it's almost an instant conversion (hence the prompts to "install the app" before the user might find the web UI too problematic).

Mobile users prefer apps (and probably trust them more) than the browser, on a mobile platform. You can say it's baseless supposition, but that's ignoring the existing evidence that companies have done (and continue to do) over the last decade. Find any company or data that contradicts that and a lot of people would be interested to see it, because nobody has for almost a decade. I can't be sure why you think that someone would consider the web to have a better experience/UI, but it doesn't matter.

What are those reasons? It seems to me that browsers are perfect for shopping as shopping is essentially browsing a product catalog. It's all basically document oriented content, static text and images, some short videos.

Sharing links is absolutely critical for shopping. Leaving product pages open for later is how many people research what they may want to buy. Product pages need to be indexed by search engines and be clickable by everyone, including on desktop.

All data lives on the server. There's little interactivity and no content creation. Offline use is rather pointless because the product database is on the server. There's no useful background activity. Notifications from shopping apps are mostly spam. Access to low level device functionality is neither necessary nor desirable.

There are far more shops than anyone wants apps. Desktop remains an important platform for shopping. So there needs to be a fully featured website anyway.

>- Applications

I think for applications it depends on where the data lives and whether desktop is an important platform. That's why essentially all business apps are web first unless their purpose is content creation and complex editing.

The real limitation of web apps is local data. There's no way for a web browser to reliably store data locally. Even though it sometimes looks as if you could use various forms of local data in a browser, on closer look it always turns out to be fragile, unreliable and only really useful as a cache.

1. Reading is not transactional or remunerative in most cases. There are exceptions (subscription-based content). Those might still be supported without going full-blown shopping, though in general I'd prefer alternate monetisation mechanisms.

2. Shopping involves payment methods and creates tremendous privacy concerns. Both of those are problematic in a generalised browser, to the extent that commerce sites are increasingly limiting which browsers they function with, which is to say, commerce is not a generalised Web capability. Payments functions could be removed from my generalised browser. The commerce-oriented app would have specific attention paid to security and privacy measures.

2a. My uparmoured Web browser using auto-deleting cookies, uMatrix, uBlock Origin, Ghostry, and other features to fight against surveillance and general Web annoyances plays poorly with most shopping sites as is. This means I've got to specifically remove armour to use such sites. This becomes more complicated when having to support others, at work or home, similarly.

2b. Information leakage between non-commercial browsing and commerce-based activity is a major concern and increasing threat. App-separation would provide a further firewall between the two.

3. A hypertext commerce platform could be URI based much as the present Web is. There are non-HTML URIs (ftp:// mail:// news:// doi:// ...). A "shop://" or "httpc://" transport (HTTP Commerce) would clearly distinguish shopping from other Web traffic and invoke the shopping application itself.

4. A possible risk would be that a large and well-resourced commerce provider might decide on delivering its own commerce app, or capture development of that app in much the same way as, hypothetically, a major Internet advertising entity might conceivably optimse a dominant Web browser as an advertising-delivery mechanism. These cases would have to be regulated by a conscientious, empowered, and principled competitions / anti-trust agency.

5. A standardised and capabilities-limited commerce application should reduce the use of dark patterns, or at least make them more difficult and apparent when employed.

Specification of a commerce-oriented application should reflect interests and concerns of vendors, shoppers, competition regulators, finance and payment processors, and advocates for groups of concern (the elderly, disabled, minorities, unbanked, etc.). The limited focus would make this far more tractable than for generalised Web browsers.