If you continue the current trends, this process could involve making WebAssembly more powerful, making CSS and Web Components into an actual widget framework and adding more capabilities to service workers to enable OS hooks, background activities and offline use.
However, all of those changes will push web apps into the direction of becoming more "app-like", not more "web-like".
So after the improvements are done, I think we might end up with just another framework for building apps, not having moved away from apps. (With the one downside that HTML will still have to drag around reams of cruft to maintain backwards compatibility which other app framework could easily do away with)
I’d add that that’s the good scenario. The bad scenario is that everything goes the way of desktop apps, where everything just becomes a lowest-common-denominator web page with crappy UI design that doesn’t really use the platform effectively, just because it’s cheaper that way and cheap always wins.
For the longest time it was. Because $newsCompany was subscription it had to leave the apple store, or give apple ~25% cut. To get round this we had a fully offline HTML5 webapp.
it was very good, but despite mega optimisations, it was slower than native, and had limited storage.
Users were constantly asking for a native app, In every customer council/survey. Bear in mind that these people were not techhead, they were babyboomers who struggled with tech, they knew enough however, to know that native apps were faster. (for example the times, grauniad and telegraph. )
I agree it’s still suboptimal for now, incremental improvements are making it less so year over year. More parts of native apps are also getting written with web technologies. Eventually, maybe, most of the app icons in our phones might just be a shortcut to some website