Hacker News Reader

Comment by msoad | Hacker News Reader

Let's face it, web on mobile has failed.

Web abstraction layer for UI is too heavy and very limited. Even today we can't have multithreaded scrolling or manage painting with the latest web standards.

Native platforms don't have to worry about "other browsers" so they can optimize their UI abstraction layer specific to their target devices. They also don't have to wait for ever to get consensus for new APIs and architectures firm multiple parties that don't specially care about your platform.

No wonder native apps are so much better for the user and with recent developments in tooling they are as easy to work with as th eweb technologies

View on HN

That's probably why Google is using the Progressive Web App version of Google Maps on Android Go instead of a native app...

The issue with web UIs in general is that it's still too hard to get optimal performance (eg. make sure you leverage the GPU as much as possible, don't force reflows, etc.). That could be solved by better tooling but we're not there yet.

We have seen how WebOS and FirefoxOS turned out. Wishing Google good luck with PWAs on slow devices. Having seen Inbox web app animations and memory usage I'm not holding my breath

Did you actually use WebOS? I've had a great experience with the Palm Pre phones, way ahead of its time.

I was a big WebOS advocate, and loved it very much. Owned 3 Pre series phones.

But you do have to admit that UI latency on it was quite high compared to iOS. I mean, so was Android back then, but they improved it significantly.

Exactly. Android had a native stack and it also had terrible performance. I don’t think it was any worse than WebOS’s.

There’s good reason to blelieve that given enough time and even a fraction of resources that the Android team had WebOS would be able to significantly up their performance game (I say a fraction of resources because there is a lot of well funded open source development being done to improve web tech on mobile, through Chromium, WebKit and Firefox that would be leveraged).

Open secret: FirefoxOS "failed" because it could not get a few super popular services in the targeted markets to write apps for it. It's very hard to break the chicken and egg issue without that support when bootstrapping a new mobile OS.

This is what I truly hope PWAs can solve. If they are 'good enough', they will bridge the gap for new platforms until they're popular enough for companies to develop for directly.

The problem is that a web app can never be good enough, because a web app can never be secure. I simply can't trust my private data to a PWA, because the developer can ship me malicious JavaScript at any time. With open source I can compile and audit code myself (or rely on others to do so for me); with an app store I can rely on a vendor to audit code; with a web app I have no assurance.

Google is a schizophrenic company.

Just watch a few Google I/O tracks, and there is always the future of mobile apps is Android on the Android tracks, while the Chrome/Web tracks are pushing for the future of mobile apps is Web, with an occasional jab at native devs, as if on the room next door they weren't talking about native development.

Then to make it even better, now we have Flutter as the third track.

While PWAs are slowly being adopted by everyone else, Google is the company pushing them.

They even have a Web Site just for PWAs, https://developers.google.com/web/progressive-web-apps/, so yeah politics.

Perhaps their product mismanagement style alone is enough of an indictment for Google to be broken up into smaller companies.

...or there are just many different ways to solve for the massive variety of problems, and developers with a huge variety of backgrounds, and Google is trying to help them all? p.s. I am a Googler

Which is understandable.

What isn't so understandable is having a speech as if on the next door employees of the same company weren't exactly speaking the opposite.

I don't understand how in stage A one can speak how glorious the future of native apps is going to be, while on Stage B we are hearing how web apps are going to win the war against native apps and PWA are the path to it.

That's not even remotely true. We are seeing significant, growing app development on the "hybrid" stack at Ionic, and strong interest in Progressive Web Apps.

How many of top 20 apps in App Store are web apps?

I agree that web apps are faster to iterate but the user experience is guaranteed to be worse.

I've worked on some of the web apps that are usually featured as "see web can do this". I'm not an outsider. I'm just observing how any app that is successful is either native already or they are making it native asap

At the risk of sounding flippant, what the top 20 consumer apps in the app store use and what the rest of the market use are very different things, especially when factoring in the enterprise. I say that as someone building a business and making money from these customers.

At any rate, we have a number of well know brand flagship apps built on Ionic, though they might not have usage patterns like Facebook/Snapchat. By far the most mobile development happening right now is outside of the well known consumer social app space.

Mainly due to the average consumer not being interested in installing an app for something like LetGo or OfferUp (Craigslist competitors). Even Facebook works well in a mobile browser, with push notifications even!

It's like claming that open-space offices are superior because you see them everywhere.

> growing app development on the "hybrid" stack at Ionic

That is not the metric. Is fast-UI, not how many devs are lazy or don't know how do native.

I invest more than 2 months researching the space this year (after several other times before), and seriously, only iOS is fast enough for html/webview.

Android is SUPER slow. I mean, i get your ionic sample app and run it on my test device. I touch once... and wait a century (sarcasm) for it to respond..

That’s a gross exaggeration re: Android touch delay. You are correct though, Android is not as performant as iOS in nearly every category, including JS/web. And no, raw UI perf isn’t the metric for most of the market. Many teams are successful with the stack and deciding to use it, considering it to have perfectly acceptable performance for their needs considering the benefits of using the web stack (as we can attest to working with many hundreds of thousands of developers using Ionic), hence it not being a “failure.” With PWAs emerging, expect the growth of the web on mobile to heat up.

I don't mean touch delays, I mean the android UI get stuck for so long to be non-usable for any than simple stuff.

> web on mobile has failed.

This definitely isn't true. The web is very much alive and flourishing on mobile.

Using web technologies for _everything_ on the phone however, is a mistake. That is not good.

In my experience it doesn't matter what you use to make an app with as long as it fits the use case. I've seen great web apps and horrible ones. I've also seen great native apps and ones with horrible experiences equally. It's all a balance act of using the technology you know best and which covers your needs best.

I still remember how Symbian introduced the Web Runtime for web based mini-apps.

Just today I had to kill Chrome on my device for the WebGL demo that was posted earlier. Yet it does OpenGL 3.x just without sweating.

Reply on news.ycombinator.com