TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Why Native Apps Really Are Doomed, Part 2

82 点作者 ericelliott超过 8 年前

23 条评论

comex超过 8 年前
I went to pwa.rocks on my iPhone and briefly tried out some of the apps, and almost every single one was obviously broken in some way. In no particular order, some of the issues:<p>- No or broken animation.<p>- Wrong dimensions for screen.<p>- Dropping frames (for the Flappy Bird clone - an extremely simple game).<p>- When viewed in-browser, swiping left to right to initiate a &#x27;back&#x27; action causes flicker or a double animation, or is broken (doesn&#x27;t go back to the right thing).<p>- Taps on the bottom navigation get messed up, showing&#x2F;hiding Safari UI (this is unavoidable to some extent but there are workarounds).<p>- Clicking the hamburger button is so laggy on one of the apps it&#x27;s probably making a network request.<p>After &quot;installing&quot;:<p>- Doesn&#x27;t work properly in a chromeless environment (e.g. the Google I&#x2F;O app has a login screen that you can&#x27;t cancel once you click it).<p>- Apple web app metadata missing or broken.<p>- Pressing any button takes you out of the app and into the same site in Safari.<p>- Icons that look bad when cropped to the iOS rounded rect.<p>Oh, and some of them didn&#x27;t work because of missing Safari features, but that at least isn&#x27;t their fault.<p>Even those that worked didn&#x27;t look native; some of them looked like Android apps, which is nice if you&#x27;re on Android. None of the ones I tried installing supported the swipe-back gesture in chromeless mode, which works in essentially every app on iOS not made by Google. (Presumably because the page would have to do it manually, as opposed to in-browser when it gets mapped to the browser back action, see above.)<p>Overall, maybe the browser primitives are there, but libraries for proper app-like UI are clearly not. I&#x27;ll stick with my native apps for now, thanks...
评论 #13003521 未加载
评论 #13004713 未加载
评论 #13011307 未加载
coldtea超过 8 年前
&gt;<i>App store friction is a major obstacle. It takes about 6 clicks to install a native app</i><p>If I&#x27;m at the app&#x27;s entry in the Store (similar to the navigation I need to do to first visit a web app), I can install an iOS app in 1 click of the Get&#x2F;Purchase button (plus the thumb touching for authorizing the sale). I&#x27;m pretty sure it&#x27;s the same in Android.<p>Even better, with the native app I don&#x27;t have to create an account, re-enter my details and suffer all the BS that I have with web based commercial services. Plus, Apple&#x2F;Google already have my credit card details.<p>I also don&#x27;t have to suffer a lowest common denominator platform, or wait for the the web standards AND the vendor to have new native feature access finally trickle to the phone&#x27;s JS engine.<p>&gt;<i>Deciding to install an app is a lot harder than deciding to use a web app.</i><p>Committing to a web app (for a user), and monetizing it (for the software author), is much harder -- especially on mobile.
评论 #13004119 未加载
评论 #13004667 未加载
评论 #13003375 未加载
sgt超过 8 年前
I still thinking developing the first app on iOS makes sense. If the app works out (i.e. people start using it), then I&#x27;ll consider building an Android version. What I&#x27;ve observed is that if the iOS app is not going to work out (because iOS users tend to buy more apps), then it probably won&#x27;t work out on Android either.
评论 #13005202 未加载
brilliantcode超过 8 年前
wow, I can&#x27;t believe some of the attitude here in the comments. PWA is not an all fit solution. It&#x27;s still in the early stages but it&#x27;s pretty good at taking advantage of your browser. It covers a high enough minimum of capabilities that&#x27;s enough to replicate an app like experience. The big pluses are the add to home feature as well Service Workers that lets you run your code after the browser is closed (this part freaks me out and saw something here about sw botnet lol) and SPA pretty much replicates UI of any native app.<p>Think of the chrome browser on Android as another app you are constantly going to compete with in which it&#x27;s capability is getting more and more absurd until it inevitably covers almost all native capability. It&#x27;s only a matter of time and seeing how quickly Chrome moves, I&#x27;m not surprised.<p>I think this is fucking great. I was going to learn Xamarin but now I won&#x27;t have to. That&#x27;s fucking huge. I&#x27;ve been putting off mobile dev for a while because I found working with Java and Swift clunky. So I waited and waited....and now my dream is coming true. It&#x27;s pretty amazing that it&#x27;s Javascript that unites them all. But ES6 is a serious contender...especially that ... operator that expands the JSON objects, it makes working with REST API which almost always involves JSON, it&#x27;s a heaven to work with. Hell, I shunned ES5 and was actually learning clojure and maybe using Clojurescript with Om, but even though I really love clojure es6 seems like the much perkier option.<p>I wonder if there&#x27;s any nice frameworks that can generate some scaffolding and UI kit and stuff for PWA. I&#x27;m learning React &amp; Redux but I personally think Vue.js will outpace it...
评论 #13003634 未加载
评论 #13003891 未加载
klaustopher超过 8 年前
From some technical perspective, the author might be right ... But what he is totally missing is the user side.<p>When seeing an ad for a service on TV (which is where most of users are coming from if you are talking B2C businesses) or anywhere else, every user goes to the app store and enters the name of the service there. I have seen quite a few businesses try the &quot;mobile website only approach&quot; and as soon as they started real world ads (read: not online), they very quickly realized they needed to be on the app stores. Cordova[1] might be a solution for that problem but then you would have to write a Cordova app and not _just_ a mobile web site ...<p>So, completely disregarding the user story, the author might be right. But when your app is trying to reach end users, there&#x27;s absolutely no way around the App Stores in the foreseeable future.<p>[1] <a href="https:&#x2F;&#x2F;cordova.apache.org&#x2F;" rel="nofollow">https:&#x2F;&#x2F;cordova.apache.org&#x2F;</a>
评论 #13004053 未加载
评论 #13004003 未加载
评论 #13009983 未加载
评论 #13004541 未加载
flukus超过 8 年前
&gt; The average user downloads less than 3 apps per month. Half of US smartphone users download zero apps per month.<p>Why is this at all surprising? It&#x27;s not very often I have more needs in my life that an app can helps solve.<p>&gt;60% of apps in the Google Play app store have never been downloaded.<p>Again not surprising. How many of these apps are crappy wrappers for a website anyway?
评论 #13003579 未加载
MarkMc超过 8 年前
The author seems to focused on the improvements to web apps, but ignoring potential improvements to native apps.<p>In particular, if Google were to support (1) Android apps running on Chrome for Windows and Mac; and (2) Android Instant Apps (coming soon [1]); then it would significantly swing the pendulum towards native apps.<p>[1] <a href="https:&#x2F;&#x2F;developer.android.com&#x2F;topic&#x2F;instant-apps&#x2F;index.html" rel="nofollow">https:&#x2F;&#x2F;developer.android.com&#x2F;topic&#x2F;instant-apps&#x2F;index.html</a>
评论 #13004131 未加载
评论 #13026892 未加载
Hydraulix989超过 8 年前
My app needed to be able to do 3D rendering of 200k polygon scenes at real-time framerates on circa-2013 devices with Adreno 320-spec GPUs.<p>Needless to say, a custom C++ engine was the only thing that could cut it (shared code base with native activity&#x2F;NDK on Android and Objective-C &quot;glue&quot; using C wrapper functions on iOS).<p>Even Unity was only hitting 8 FPS on Adreno 320 (with mobile shaders and everything I could think of doing to try to coerce their black box into doing what I wanted it to do)...
评论 #13004995 未加载
jevinskie超过 8 年前
It will be interesting to see how Android &quot;Instant Apps&quot; end up working. Looking at AOSP, bits of &quot;ephemeral-app&quot; support code has been tricking in for (probably) Android O release but I haven&#x27;t been able to find a complete example. Has anyone else had better luck? I never received a reply from Google about an invitation to the Instant App beta.<p>I&#x27;m mostly interested in if the ephemeral apps can use native code. From what I&#x27;ve seen so far in platform&#x2F;system&#x2F;sepolicy, it looks like native shared libraries <i>will</i> be allowed but I&#x27;d be very happy to hear any more information that others have uncovered.
TeeWEE超过 8 年前
Native apps are always here. Your browser is a native app. Your dailer is a native app. Your homescreen is a native app... There will always be native apps. And on top of that higher level languages might work. But its not that progressive web apps will replace native apps. Maybe for some apps. But the current tech stack for web apps will never beat the performance of native apps. Their architecture is build for very fast loading, and best graphics performance. Web still has some way togo, and is inherently bound to http request based app loading. Which will always be slower than loading native code into memory.
partiallypro超过 8 年前
I think native apps are doomed for most smaller companies where the cost just doesn&#x27;t make sense, but there will always be a place for native apps for popular services that need OS level permissions. I do wonder how many native apps for smaller companies that end up costing more money than they are worth, but I&#x27;m sure at this point some companies are invoking the sunk cost fallacy instead of creating a proper web app that would save them money in the long run. Or the &quot;we have to have this because our competitor has this,&quot; even if their competitor is experiencing the same problem. It will be interesting to see how long it takes for that to shake out.<p>On a whole it&#x27;s not doomed, but it&#x27;s a bit silly to have 2-3 dev teams that have to keep platform parity in native apps instead of just having one app that has the same experience across all platforms. Even solutions like Xamarin haven&#x27;t really removed this rigidity with continual delivery and feature parity.
评论 #13004332 未加载
blntechie超过 8 年前
I use HTML apps in place of native apps as much as possible and this is B.S. We have been hearing about web apps replacing native apps in mobile since the original iPhone came out without a native SDK. I understand JS and browser are catching up but they are not close to replacing the native apps anytime soon.
KaoruAoiShiho超过 8 年前
Am I missing something? According to the linked alibaba case study there should be a add to homescreen prompt on it. However, I go to alibaba.com, I get a link to google play or to download an apk, no add to homescreen prompt. Did the experiment fail and alibaba revert to native?
评论 #13002884 未加载
评论 #13004559 未加载
SCdF超过 8 年前
I think the lack of friction is a real and present benefit. And the benefits of native development (eg interaction outside of the browser sandbox, high performance) are not required for a lot of apps. <i>And</i>, it would hopefully kill off all of the pointless individual news sites app that exist for no reason.<p>If Google and Apple coherently get behind the concept, it could be a big deal.<p>There is a lot you&#x27;d need to do right: making sure things like notifications and storage are managed clearly; allowing users to &quot;uninstall&quot; &#x2F; wipe apps that have gotten too large; something to combat and manage the slow sludge that will accumulate with endless &quot;driveby&quot; app installs (imagine if every news website was a PWA with offline content, you go for one article and oh look now you have all of their CSS and JS with offline support stored on your device, and maybe random data [say, how far you scrolled in the article + the article contents for offline support] stored in leveldb, with no way for the browser to know if automatically clearing it will break things for the user); etc.<p>The problem is that Apple have shown no interest in doing this.<p>And I cannot see Google internally organising themselves well enough to be coherent about this either (c.f. the messaging &#x2F; hangouts &#x2F; allo &#x2F; duo disaster).<p>Which would be a real shame.
评论 #13004951 未加载
franciscop超过 8 年前
&quot;iOS has 45% of the US smartphone market, and iOS users spend $1.08 per user per app per user vs $.43 on Android, so obviously, write an iOS app and you’ll come out ahead, right?&quot;<p>Let me guess those prices ($1.08 and $.43) are for the US. Which is probably the country where people actually pay for most stuff, I&#x27;m pretty sure Android users monetary reward differ from the US than from other countries
ivan_gammel超过 8 年前
I don&#x27;t think it&#x27;s correct to blame the technology choice on the failure of some apps, as the author does, talking about percentage of revenues and user retention. Every technology has it&#x27;s own pros and cons and with adequate investment in R&amp;D can be applied to achieve very good UX. More likely, the success rate has business reasons, not technology ones - how the development was organized, what was the business model, how much time did the team spend on UX design etc.<p>That said, what are the real benefits of PWAs, which by design are not fully compliant with platform best practices? That&#x27;s the question that I&#x27;d like to get an answer to, including the comparison of installation success rates, not just exposure of user failures to install native app.<p>I myself have chosen Ionic&#x2F;Cordova as reasonable trade-off between the speed of cross-platform development and quality, but eventually I&#x27;d like to convert my app to the native to get the best UX on each platform.
enos_feedler超过 8 年前
The issue with app store friction is obviously going to disappear. This was in the cards ever since Apple decided to do runtime permission model vs. install time. If there is one thing iOS innovates on is reducing friction (identity, payments, etc). The app store is clear friction. In the next couple of years you will be brought into a native app experience simply by tapping a link.<p>Why isn&#x27;t it here yet? Some infrastructure and technology is needed under the hood. Distributing resource dependencies (code, data, assets, etc) under the hood so tapping a link is a smooth experience is not easy. We will get there though.<p>Google has announced similar technology with Instant Apps. I think with both sides its going to take time for developers to really understand the paradigm and build apps in a performant way for link-driven experiences.
z1mm32m4n超过 8 年前
&gt; Android has 86% global market share.<p>I hear it frequently mentioned that Android has larger market share globally, and as such is a better target than just looking at iOS market share in the US.<p>Is this a fair statistic? Let&#x27;s say you&#x27;re just writing on the first version of your app. That is, it doesn&#x27;t have the full i18n and l10n treatment yet. How do things compare if we still set our sights globally, but we only look at the English-speaking population?<p>My hypothesis is that it might be worthwhile at least when starting out to target iOS just because it has a larger market share in the population that you&#x27;re likely to hit first. Ideally later when your app has proven itself, you&#x27;d expand to different regions, different platforms, etc., but I&#x27;ve only ever seen one side of this statistic.
评论 #13003898 未加载
评论 #13003895 未加载
relics443超过 8 年前
I feel like I&#x27;ve heard this before...
fit2rule超过 8 年前
There are other ways to write cross-platform apps that don&#x27;t require one sell ones soul to the web demon - for example, using a VM of ones own choosing, one can build a host engine on each platform, and then use the same common framework for app development ..
评论 #13005006 未加载
评论 #13010129 未加载
baybal2超过 8 年前
&gt;Eric Eelliot<p>The guy has a trolling black belt
onebot超过 8 年前
Nonsense from a &quot;Javascript&quot; developer.
评论 #13003835 未加载
评论 #13003660 未加载
_RPM超过 8 年前
Use the Fast Mail mobile app for a week, then tell me this again.