Comment by christiangenco
9 years ago
I think a lot of commenters here are missing the point and getting distracted by push notifications (who wants a website spamming them with notifications?) and loading screens (hardly a feature).
Apple supporting PWA (Progressive Web Apps) is hugely important because it enables a future where web apps can natively support browser, Mac/Windows/Linux desktop, and mobile iPhone/Android/Windows native mobile with a single codebase of open technologies.
Why is that important? By fragmenting development effort, the overall product isn't as good on any platform.
There's an app I'm making on the side to keep track of your contacts (like a personal customer management system). This needs to store all your contacts offline, because it'd be too much friction to load everyone you've ever taken notes on over the network every time you open the app.
Right now, the only way for me to accomplish that on iOS is to make a native app. This means I had to learn an entirely new technology stack (React Native and XCode), completely rewrite my views, tie everything into my backend, and go through Apple's Byzantine approval process (which I still haven't done because I can't figure out why my app compiles and runs locally but complains about libraries not being linked when I try to archive it to upload to the app store).
This is unnecessary duplication of work that could've been spent writing new features, makes it harder to add new front-end features in the future (because now they have to be added in two places), and adds a huge lag in the time it takes me to push changes to the iOS client (weeks, vs. the seconds it takes to push a change to the web client).
If apple supported PWA, I would've spent my time making the database keep a local syncing copy on the browser (with minimongo or pouchdb), and then every platform would've benefited from faster page loads and offline syncing.
Until Apple adds PWA support, I can't make as good stuff, and people can't use the better stuff.
But as an iOS user I expect you to use the technology stack provided by my preferred operating system. I don't want to use your app if you're targeting a lowest-common-denominator feature set.
When I change my preferred text size through accessibility settings, good native apps respond correctly. If I need voice over support, the operating system knows how to read the view hierarchy to me in a logical way.
When drag-and-drop becomes a thing in iOS 11, native apps will implement that feature well. I think it will take some time for web apps to implement it as nicely (if ever).
There are thousands of tiny details that your web app just won't have. Those details are more important than your familiarity with a tech stack or how long it takes you to deploy something.
You say that:
> By fragmenting development effort, the overall product isn't as good on any platform.
But I would say that:
> By building a web app, the overall product isn't as good on any platform.
I have yet to find a "web app" that I delight in using, though I love many web sites and native apps.
>But as an iOS user I expect you to use the technology stack provided by my preferred operating system. I don't want to use your app if you're targeting a lowest-common-denominator feature set.
I would wager that the average quality of an iOS app written by someone like OP that "has a web experience and _has_ to learn iOS just to work on that platform" will probably be lower than if that person (with web experience) could just extend their web app natively with PWA.
I completely agree with you that the app should feel native to the platform (I actually quit a job a while back because they wanted me to theme our web Android experience as iOS for "consistency across all devices" instead of matching the user's device's design patterns), but there is huge value in giving the devs the tools they need to write the best product they can, and splitting codebases and requiring more work/knowledge/moving parts is actively detrimental to a quality end-product for everyone, unfortunately.
> iOS app written by someone like OP
That's a business choice that they made, and will suffer for. If a developer and business choose to half-ass the native iOS application, there's no reason to believe it would behave any better on iOS when written as a webapp.
> the tools they need to write the best product they can
The best product that can be made will never be made with cross-platform tooling. It will always be lacking. My proof for this? Java programs. TCL programs. Electron programs. There has yet to be an application which uses the same GUI code across applications that is as good as a native application.
21 replies →
>I don't want to use your app if you're targeting a lowest-common-denominator feature set.
You might even use a hybrid app without it knowing. Many apps just need to show some buttons, input fields, images or a map and hit a web service. Brushing ALL hybrid apps off as useless is in my eyes just ignorant.
It's possible. I still use web apps (e.g., Slack client on macOS). But I dislike them compared to good native apps, mostly due to their lack of consistency with the platform and general sluggishness.
If I'm using a web app and not realising it, then I would happily keep using that app. I do not think I am, though.
Also, there are plenty of native apps which are terrible and not consistent with the platform. I do not like to use those either.
7 replies →
This hits the nail right on he head. I can see developers wanting to have one codebase, but as a user I want apps that can take advantage of everything iOS offers. There is no Picture-in-picture or metal in PWA.
> I have yet to find a "web app" that I delight in using, though I love many web sites and native apps.
I wonder how much of that is an intrinsic problem with web apps conceptually, or a result of the various limitations and design fuck-ups of the browser vendors.
The fuck-ups aren't just on the browser vendor site.
The entire front-end ecosystem is currently worse than enterprise Java at its peak and the quality issues stem from all sides.
The best web app I use daily is Slack (wrapped in its native app "shell" on macOS). But it still feels pretty poor.
On iPad it's worse. On iPhone strangely, it's not too bad.
1 reply →
> There are thousands of tiny details that your web app just won't have. Those details are more important than your familiarity with a tech stack or how long it takes you to deploy something.
How many details does an app that doesn't get written have compared to a website?
> I have yet to find a "web app" that I delight in using, though I love many web sites and native apps.
There are a lot of websites I like using. There are very few apps I like using so much I want to install.
You can't get pixel-perfect (including android) between all devices, but to the extent that a manufacturer enables a "better" experience, it gets much closer and more consistent behaviour compared to native apps:
https://webkit.org/blog/3709/using-the-system-font-in-web-co...
...and even if you get pixel-perfect between android and ios, you sacrifice "cultural-correctness" (ie: floating buttons v. top bar v. bottom bar, etc.).
Writing two pixel+cultural perfect apps on two platforms, keeping them in sync, making sure they're not buggy, attempting to share code, attempting to keep them both secure is incredibly expensive. If you don't believe me then do it yourself.
Making a PWA which gets 90% of the way there, and integrates as well as possible with the system (ie: fonts, location, notifications, accelerometer, etc.) is generally _less_ expensive than doing a single native app well, and has the chance to get you 90% of the way there on desktop and your "alternate" mobile platform.
PWA can be incredibly powerful (along w/ manifest.json-style support as android has), and I'm waiting for the day apple catches up to android on this one.
"and even if you get pixel-perfect between android and ios"
Nobody but designers who think too highly of themselves wants this. Everyone else wants the app to fit in with the platforms toolkit. This requires the designs to be different.
"Writing two pixel+cultural perfect apps on two platforms, keeping them in sync, making sure they're not buggy, attempting to share code, attempting to keep them both secure is incredibly expensive. If you don't believe me then do it yourself."
Which is why I don't do it the way you described. I embrace what makes each platform unique.
I think PWAs are fine. I just wouldn't encourage them to be used side-by-side with native apps (e.g., deploying them to the home screen). It creates the expectation that they should culturally fit in with native apps, and they won't.
I'm okay with them living in the browser and gaining the performance advantages, offline support and push notifications.
Theres something about this comment that winds me up but I don't have the time to unravel.
> When I change my preferred text size through accessibility settings, good native apps respond correctly. If I need voice over support, the operating system knows how to read the view hierarchy to me in a logical way.
> When drag-and-drop becomes a thing in iOS 11, native apps will implement that feature well. I think it will take some time for web apps to implement it as nicely (if ever).
All these things the browser should be able to do well, if they wanted.
Given that web apps still fail to do drag-and-drop (and copy and paste) as nicely as native apps — and these things have been a staple of Mac interaction since forever — I do not hold out much hope.
2 replies →
But they don’t.
> But as an iOS user I expect you to use the technology stack provided by my preferred operating system.
As an iOS user I don't expect Apple to mandate your preferences for me
> There are thousands of tiny details that your web app just won't have. Those details are more important than your familiarity with a tech stack or how long it takes you to deploy something.
They are more important to you. They may not be to me if they prevent an app I need being made, or being available cross-platform (much more important to me than it being perfect on any one), or being affordable (to me).
The notion that every app must be the perfect gold-plated 'delightful' experience is corporate marketing drivel. It is relevant to some (people and apps), but not others. We don't need the personal tastes of some precious souls to be mandated for all of us by the platforms we happen to use (today).
I disagree that it's marketing drivel. Wanting everything to be consistent is the entire reason I was driven to the Mac as a platform in the first place many years ago.
Being on a platform where users and developers care about exactly what shade of grey their menu bar icon was, or matching the platform characteristics, adopting system-wide services, making apps accessible, is very important to me.
You may not care, but that's how I choose a platform. It's not marketing, it's personal preference.
I'm going to push for iOS and macOS to develop in this direction by supporting developers who try their very best to make thoughtful and consistent software.
Your argument works against yourself: "cheap cross-platform apps are relevant to some, but we don't need the personal tastes of some precious souls to be mandated for all of us." (I'm not making this argument against you, but try to see how cross-platform is your own personal preference that you are trying to push onto others. In my opinion it degrades a platform even if you don't use any cross-platform apps.)
Could not agree more. How does a PWA use ARKit? How does a PWA integrate with the camera? How about the accelerometer? What about iCloud, Handoff, or any number of iOS technologies? Perhaps Metal or other iOS graphics technology? What about TouchId? Bluetooth?
I also find the comment about needing to learn a new stack “React Native and Xcode” to be ridiculous– no, what needs to be learned is Swift and Xcode.
Far too many “web” developers consider native mobile to be some kind of subset of web development and thus expect to use the same tools as they use for web.
Web is a different medium! If you want to program embedded systems, then the first question isn’t “how can I do this with JavaScript?” They would learn the correct language for the platform, perhaps embedded C. You don’t launch a Linux server and then ask “how can I make this server run Windows? I guess I should write a JavaScript library for that!” It’s ludicrous.
With iOS, developers often just think of it as a “native” website rather than an actual application. It seems like some developers will do everything possible to avoid simply learning Swift and making actual apps that fully exploit the power of the device.
React Native – if that is considered “good” then we have major problems. Facebook applications are horrible at power management; they suck power at phenomenal rates compared to other applications. The smoothness of the UI isn’t as “native” as actual Swift apps coded correctly. There always seem to be a slight amount of glitch in the experience. Facebook has famously avoid actually coding real native apps – from the beginning of their mobile experience they have seemingly embraced doing everything except writing actual Swift or Objective C. It is almost a religious opposition to it – and despite being a multi-billion dollar company, some tiny app studio in Poland could write higher quality apps. It should be an embarrassment, but they’re Facebook so everyone just accepts the status quo of less than perfect. No person here can say that the Facebook apps are perfect. But they should be. They have a gazillion dollars and can hire almost anyone they want, so they have no excuse for anything less than perfection. At the very least get power management right!
There’s always this argument that x-Native is “good enough” – if, as a company you want “good enough,” then keep making apps that conform to the lowest common denominator. If you want to make extraordinary applications that move the needle of quality, then use Swift and build it correctly.
This will likely get downvoted into oblivion because the HN crowd seems to be exhaustingly enamored with React Native, however, regardless of how it’s framed, writing PWAs or using some cross-platform “solution” is a cop-out. It’s lazy and it provides users with an experience that is worse than they deserve.
iOS is better than Android in so many ways, yet developers insist on making iOS apps that are really just cross-platform compromises.
My tiny bootstrapped company is working to release our iOS app, with Android soon to follow – if we can do it, there seems little excuse for actual funded companies to skimp on providing the best experience for users. Those arguing that PWA or x-Native cross platform systems are just as good as actual native, well there is no amount of argument that will change your minds. Which is sad. Rather than trying to make React Native, etc. “better” why not just use what is already better? Why not let users enjoy the full power of their devices instead of writing these average “good enough” compromises. It’s like this ridiculous trend of using Electron or, in the past Adobe Air. Nowhere near the quality of writing an actual native app. Looking at you Slack. Slack is even proud to have made a “native app with web technologies.” WHY? Damnit make a native app with native technologies! Can you not hire two actual MacOS developers? Why should making Electron apps be celebrated? It’s sloppy. It’s lazy. It’s a disservice to users. Why use some Electron-wrapped webpage and just not the webpage?
Every day it seems on HN people are posting about <some language> but very few posts about Swift. Is there some opposition that I am missing? Why must JavaScript be the language of everything? what ever happened to picking the right language for the job rather than trying to force a web peg into a native hole.
By the way, my exact arguments could be made for Android development as well. Android users are also being short-changed by these pseudo-native cross platform “solutions.”
Yep. There are also, as I'm sure you're aware but for the benefit of the less-familiar, plenty of "hybrid" ways to build "native" apps for mobile: https://github.com/auth0/blog/blob/master/_posts/2017-01-11-...
Apple supporting PWA (Progressive Web Apps) is hugely important because it enables a future where web apps can natively support browser, Mac/Windows/Linux desktop, and mobile iPhone/Android/Windows native mobile with a single codebase of open technologies.
Why after over 30 years of experiencing cross platform "write once run anywhere* technologies do developers still think that's the best user experience? Yes it makes life easier for the developer but it's rarely best for the user.
I think this is the crux of the matter. Apple supporting PWA means lower quality apps for its users, and Apple has the market share to demand apps be native code.
I'm not trying to argue that it is the BEST user experience. I'm trying to say that it is hurting small dev shops and startups because they are being forced to learn a completely new tech stack in order to play ball. I could have spent that time implementing new features that users would actually use and in turn improve their business, or in this case, reach and help more people with valuable medical advice.
In the end, Apple got what they wanted. I needed a feature that PWA's can give me - but Apple hasn't added support for them in mobile safari, so I paid the $100 to get access to the app store, and was forced to learn a completely different language.
Yes, the end product has an arguably better and 'native-like' experience, but it took me longer to do and it is lacking some of the features that I could have rolled out if I was able to use PWA's. And it would have worked on Android out of the box as well.
I don't regret learning React Native. It was actually really, really fun. The community is great, and being able to write native apps now feels really good.
But its the principal of the matter. Holding back innovation for your company's own selfish reasons is a shitty thing to do.
Yes, the end product has an arguably better and 'native-like' experience, but it took me longer to do and it is lacking some of the features that I could have rolled out if I was able to use PWA's. And it would have worked on Android out of the box as well.
So am I as an end user suppose to be upset that you were forced to make a better product.
Holding back innovation for the company's selfish reasons?
Back in 2008 they said the same thing about Apple not supporting Flash and Java.
If anyone is being selfish to try foist cross platform apps that you admitted weren't as good, it isn't Apple.
GMail was (and maybe keeps being) a much better user experience than native email clients for desktop PCs.
I can’t disagree more. I find gmail to be a perfect example of why web apps are playing to the lowest common denominator and result in poor user experiences everywhere. Just my opinion though, I realise most people love it.
5 replies →
It's better in that it required no configuration. But from a UI perspective it feels inferior. For example, native clients can just show you a list of all your messages, but GMail still paginates like a late-90s PHP site.
2 replies →
Believe it or not, in my experience, Exchange + Outlook 2016 stomps all over GMail. I find that its faster, searches quicker, and takes up _WAY_ less memory. I don't do any fancy things other than basic email, scheduling meetings, etc so YMMV.
Completly disagree.
I only use GMail as gateway to aggregate my email accounts and synchronize with my Android devices, native mail client.
On Windows and GNU/Linux systems at home, I happily keep using Thunderbird.
4 replies →
I never quite understood the complaint about having to learn a new tech stack to write native apps. As a native app dev, it feels kind of like getting angry that your skills in motorcycle mechanics don't transfer to building rockets.
And this complaint practically always comes from the front end web crew... every other type of developer I've met has zero issues with learning the technologies relevant to a particular platform.
I'm a scientific programmer who just wants to have a few tools available across my devices, mobile as well as desktop.
PWAs are a great idea for me. Getting an Apple developer license is not.
Then don't develop for Apple platforms. We don't want your "apps" anyway.
1 reply →
You can make a a native app, which will always be better than a webapp. As an iOS user, I have no intention of ever using webapps (including things like Cordova apps). If you can’t be bothered to make a iOS-native app with iOS native look, feel and features then just don’t bother.
But it means learning something other than JS!
It feels like many developers are adamant to never leave their comfort zone. Hence, JS everywhere.
Coders like to code. Product builders want to build products with a minimum of fuss.
1 reply →
A native app will never have instant install, or update without download. A native app will never be available on all platforms.
Sometimes a native app is better, sometimes it is worse. It's certainly better if all you care about is fancy animations.
> A native app will never have instant install, or update without download.
So ? Updates happen automatically anyway.
> A native app will never be available on all platforms.
I don’t need it on all platforms, I only use iOS.
4 replies →
Is anyone demanding instant install or updates ?
I can download 95% of apps in under 30 seconds and all of my apps update whilst I am sleeping. And nothing is stopping you downloading new content in a native app which is where the majority of use cases stem from.
1 reply →
As an iOS user, do you ever use websites that are more than plain hypertext, even if they don't try to have an iOS native look?
A website is not an app. Not sure why I need to explain that.
5 replies →
I think for most people, the answer is a resounding yes.
> Apple supporting PWA (Progressive Web Apps) is hugely important because it enables a future where web apps can natively support browser, Mac/Windows/Linux desktop, and mobile iPhone/Android/Windows native mobile with a single codebase of open technologies.
If 'a single codebase of open technologies' is so important then the same argument says Apple should abandon their platforms in favor of linux/AOSP. I'm sure a number of people here believe that (and strongly) but ... (1) Apple has a few hundred billion counterarguments sitting in the bank and (2) the entire industry has greatly benefited from Apple's efforts at pushing their closed platforms.
And if 'a single codebase of open technologies' isn't the be all end all then the argument reduces to "Apple should subsidize the technology I'm invested in". I bet turnip farmers think Apple should buy lots of turnips too.
Reverse the discussion:
I'm a C and C++ programmer with a background in embedded systems. My idea of a development framework is a Makefile, a bunch of headers and .a files. I've got an investment in a lot of libraries I've already written over the years. I want to develop great web applications. Should I feel frustrated that it's not straightforward to use my preferred toolset to build this web app? Should I blame browsers for not accommodating my development preferences? No! I need to bite the bullet and learn JS and HTML.
You need to pick the appropriate tools for the platform you're targeting, get out of your comfort zone and take the time to learn them.
Could always apply to OkCupid. They, crazily enough, build their website with C++.
> Apple supporting PWA is hugely important because it enables a future where web apps can natively support (...)
But it is not in Apple's interest to support these kinds of apps. They make money on the app store, and like to keep developers within the walls of their ecosystem.
You can definitely build that contacts app as a web app on mobile safari. HTML5 appcache lets you take it offline, localstorage gives you 5 mb of storage, web sql or indexeddb gives you 50 mb. You can use something like pouchdb to give you a clean wrapper.
Yes, service workers would be nicer, but you don't have to go native to do an offline app on iOS. I built one years ago, appcache is good enough.
What are you storing in your contacts database that's so huge that it can't be loaded over the network every time?
Are you hand-writing notes and storing the notes as images?
I use Apple/Google's built-in notes fields on the default addressbooks and it works just fine. I can't imagine having huge write-ups on individual contacts unless it was for some business purpose. In that case, I'd move to a dedicated note-taking application anyway.
He is probably concerned about users who cannot use a network all the time or have slow connections. Many phones have LTE but trying to load some modern websites over slower 4G, or even slower 3G, is a nightmare, not just from the size but also if they're not on a proper CDN.
Maybe developers should focus on the subset of platforms that they know they can deliver a first rate experience on, instead of forcing support for every platform imaginable.
It's a sad state of affairs for the web when we see articles claiming that one company's failure to adopt a standard amounts to threatening the web. The open web supports and should encourage competition, not blind homogeny.
I don't think it's misguided for a company to hesitate in supporting a platform which will enable its users to jump ship to another competing company.
> This is unnecessary duplication of work that could've been spent writing new features
I bet that's how Apple feels about implementing PWA.
So put your web app in a Cordova bundle and you can send push notifications. No need to learn react or write native code. Not a huge problem to solve.
Thank you Christian. This is what I was trying to say. I think I could have better job at making my point more clearly. But you got it.
> Right now, the only way for me to accomplish that on iOS is to make a native app. This means I had to learn an entirely new technology stack
It's not the only way - you can create a hybrid app with something like Ionic (obviously this is compiled to a native app at build time, but you never touch a line of XCode yourself).
One of the big selling points of hybrid apps is that you can use your existing Javascript/Typescript skills to create apps that look and feel like native apps.
Downvoter care to comment? Seems a bit random.
It seems like at some point we're just going to end up re-writing Java - But for the web.
"By fragmenting development effort, the overall product isn't as good on any platform."
At the same time, you're no longer focusing on what makes each platform unique, and just giving them all a lowest common denominator. That's not really good either.
"This means I had to learn an entirely new technology stack"
Learning new things is good.
"Until Apple adds PWA support, I can't make as good stuff, and people can't use the better stuff."
This is absolutely wrong, and is just an excuse. You can make great stuff, and people can use it. You just need to put in effort.
You can make a a native app, which will always be better than a webapp. As an iOS user, I have no intention of ever using webapps (including things like Cordova apps). If you can’t be bothered to make a iOS-native app with iOS native look, feel and features then just don’t bother.
Screw new features, using JavaScript and HTML/CSS for apps must die.
The DOM and CSS are good underlying principles. It would be nice if there was an alternative to HTML, like something encoded in JSON.
As to JavaScript, with WebAssembly coming, there may be options in the future.
> something encoded in JSON
Those who do not learn the lessons from XML are doomed to repeat XML.
The DOM is simply an in-memory representation of an XML structure, and any attempts to populate a DOM with JSON (or YAML or ProtoBuffers or...) will simply re-create XML. CSS is a language for writing XML transformations (i.e. XSLT).
Of course, we already have JSON versions of schemas, transformations, xpath, namespaces, and incompatible decoders, so perhaps it is already too late for JSON.
1 reply →
Flow-oriented markup is meant for documents not GUI.
3 replies →