Hacker News Reader

The reviews I've read about the existing Firefox OS phones, which have reportedly cost well over $25, have not been very encouraging. Pretty much all of them have pointed out that the hardware is quite poor, and this results in a rather bad user experience. Even considering how rapidly the cost of mobile hardware has been decreasing, what will prevent a similarly negative user experience on these cheaper Firefox OS devices?

While the ideological benefits of HTML5/JavaScript/CSS "apps" are often touted, we've yet to see anything truly remarkable done using such technologies. At best, we've only really seen clones of existing Android and iOS apps, but these generally feel less responsive and less capable than their native counterparts. If Firefox OS only supports these kind of non-native apps, how is it supposed to truly compete with Android, iOS and the other mobile platforms that support these HTML5/JavaScript/CSS "apps" just fine, but also offer far more choice and capabilities to developers in the form of real native apps?

And one other thing that I rarely see considered by the advocates of Firefox OS is how we're seeing more and more used iOS and Android mobile devices ending up in developing nations. They tend to be quite affordable and usable, even if they are somewhat outdated. Why would people in developing nations pay for a new Firefox OS phone, when they can get a used Android or iOS phone that may very well perform better than the Firefox OS phone, while offering a wider selection of apps, while also costing roughly the same?

Can you address these concerns?

How is a Java code running in a VM any more native than web technology?

I wouldn't say that it necessarily is. But Android does offer the NDK and the upcoming ART runtime, which get far closer to native code than what we're seeing from HTML5/JavaScript/CSS.

ART is just a replacement runtime for Davlik. It's still no more "native" than a JavaScript VM. The world "native" has been bended to mean practically nothing at this point, and instead is used to differentiate "good" platforms from "bad" ones.

JIT

I believe these phones are aimed at two market segments:

1. Folks will buy these instead of computers - many of whom don't have computers today. Meaning, their first computer will be a mobile one.

2. Cost-conscious/value-conscious people that would prefer a cheaper new phone over a used used phone (even if slightly cheaper from the perspective of the developed world).

So market (1) doesn't really need killer web apps. The apps are already there; the developed world consumes them via PCs, Macs and tablets.

Market (2) is interesting, especially if new Androids come down market, but I would not underestimate the impact that latent demand from these mega-markets can have on HTML5 apps, many that btw don't even exist today as native apps because they are more relevant in a developing nation.

FirefoxOS has lower resource requirements than Android - for example this $25 device has only 128MB of RAM. FxOS (an it's app ecosystem) is still very young (1.3 just came out, and the first release came out at the end of 2012). I'm positive that we'll continue to see dramatic performance improvements as well as feature parity with the dominant platforms, along with the app ecosystem continuing to mature.

On the app developer side, making apps requires no specialized development environments or knowing anything beyond web technologies (HTML/JS/CSS), which dramatically lowers the bar for making apps.

I'm not sure what counts as "truly remarkable", personally I am pretty impressed by games like Cut the Rope on my FirefoxOS phone. I also carry and regularly use an outdated Android device that has very similar specs, and it's not particularly responsive. More apps are available, but the average quality is not very high.

I could write a whole essay on this topic, believe me... so lets go with the short version.

First is that most of the phones out there are running Firefox OS 1.1 and this version is not as snappy as Firefox OS 1.3 which is probably the next version those phones will upgrade to. Firefox OS 1.3 is much faster and solved a lot of bugs.

HTML5 is the native thing in Firefox OS. All apps and parts of the system itself are built with it. Thats how good it is when we build Telephony stuff in Javascript.

The thing is that Firefox OS is also the easiest platform to develop for. Its much easier to develop for web technologies than it is to learn ObjC and Java. That being said, lots of new developers are coming on board are still learning the best practices so we see a lot of apps that could be better just like you see a lot of pages that could be better on the web.

The Gecko engine and its WebAPI provide the developer with all that is needed to create good quality apps but since this is our first year, some of our third party developers are still learning the trade. Its getting a lot better.

There are 8 million web developers in the world. They could all be shipping apps on Firefox OS if they wanted to. Its free and we provide all the docs, also, its quite easy to develop for Firefox OS first and then pass thru Cordova or similar tech and deploy on other platforms.

We believe that the web will win in the long run and that a "web first" mentality will emerge where people will opt to go with an HTML5 solution instead of ObjC/Java/C# threesome.

I invite you all to explore our rich APIs at https://developer.mozilla.org/en-US/docs/WebAPI and see how every time we feel that something can't be done in HTML5 we try to come with a proposed standard. Its this type of evolution of web standards that moves the web forward in all platforms such as desktops, laptops, TVs and more.

As for used cellphones being sold, that doesn't really count in the larger scheme of things. Specially when dealing with Apple. As an exercise try picking an iPhone 3GS and installing random stuff from the App Store... its very hard because developers in those platforms are always developing for the new shiny and forget old customers with outdated hardware very easily. Now try installing an old web browser, something like the same Firefox that was available at that time and browsing the web. It will work because thats how web stuff is made, its made to keep working or degrade nicely. This means that what we call "open web apps" not only are easier to build but they can be used by more people with different hardware and thats why we have the world wide web and not some strange silo such as compuserve...

Think of yourself in five years: What type of world do you want? A world where web standards and engines have evolved to the point where you can do all you wish as a developer without moving away from web technologies and you're free to run your app in different platforms and systems or a world where you need to ask "pretty please" to a company in the north region of california every time you want to ship an app?

Sorry for making this long...

> First is that most of the phones out there are running Firefox OS 1.1 and this version is not as snappy as Firefox OS 1.3 which is probably the next version those phones will upgrade to. Firefox OS 1.3 is much faster and solved a lot of bugs.

You know, we've been reading the exact same excuse about Android since the 2.x versions yet here we are at version 4.4 where even Google's own apps still exhibit noticeable lag and stutter in scrolling and other seemingly simple operations. User interface performance should be one of the topmost priorities for a modern OS, not an afterthought.

Considering that even my desktop computer with a quad-core processor and top of the line graphics card can't render common HTML5 web apps at a smooth 60 frames per second, how is a $25 smartphone going to be up to the task?

There was a really good writeup about the decision process in the early iPhones where Steve Jobs ordered the UI in its own process so that the user gets instant feedback on interactions even if the feedback is a simple loading spinner. It gives the perception of speed even if a modern android phone is timed to load the same application faster. There are ways to try and achieve this in web apps but Apple's early decision to use that approach was certainly brilliant. If you call a bank and a human answers immediately...telling you to please hold...it would be a far better experience than the current press 1 for English.

> Considering that even my desktop computer with a quad-core processor and top of the line graphics card can't render common HTML5 web apps at a smooth 60 frames per second, how is a $25 smartphone going to be up to the task?

Of course you wont play crysis on these phones. these phones are not tailored for those triple-A games full of 3D stuff but they are pretty good for 2D casual gaming if the developer knows what it is doing.

HTML5 is not a magic bullet. It doesn't solve everything. Of course there are domains that are better suited for other languages. The fact is that HTML5 is the only toolkit that gets us from computers to mobile platforms, smarttvs, videogames, smartwatches and more. Basically its the most common runtime everywhere and its also the easiest. Don't think of Firefox OS as "just a mobile platform" like Android. Think of it as a champion or an extension of the world wide web ecosystem. Its picking what we like from the web and evolving it until it is great on mobile devices as well but without losing the objective of keeping it interoperable with other platforms just like you can open HN on Firefox on Mac OS X or on Midori on Debian/ARM, the interoperable part is key.

The phone with the weakest spec I have here is the Alcatel OT Fire. It has 256mb of RAM, 1x Core and yet I can play games at 60 frame a second in it. This, of course, just work if the developer knows how to code for a hardware with such constraints. If a developer code his app on his i7 with 32gb of RAM and dual graphic card without ever thinking of lower hardware then he or she will probably face difficulties when porting to entry level mobile phones.

HTML5 mobile gaming is still evolving a lot. Check out http://phaser.io for a powerful FOSS framework that is able to deliver great performance on entry level phones.

Achieving 60fps is not possible in all occasions but achieving interoperability is better. I'd rather have a poor person pick a smartphone for the first time, learn just enough HTML to build a little app that helps his community and start changing things then having the same person get into debt to buy the latest iOS thing and then be unable to deliver the app for his community because Apple didn't like it.

I am not talking about games at all. I am talking about simple things like scrolling list views. As touch interfaces are very imprecise by definition, a smooth framerate is absolutely necessary for a usable interface. It's very hard to click or even find the right item in a janky list. And that's what I'm getting at: my desktop computer can't seem to do it in HTML5 apps, so how can a 25$ phone?

It's nice to think about ideals like a common runtime, but this ideology starts to fall apart from multiple directions once you think about it. I already mentioned the performance issue, but what about things like app design?

Interoperability is a bad thing for user interfaces. Unless those mobile phones, smart TV's, video games and watches are all magically going to switch to a Firefox OS monopoly and its interface and guidelines, we're in for a future of switching between apps and finding ourselves having to relearn basic functionality in each one of them.

I don't know about you, but I already face this uncanny valley daily using different web apps in my browser. There's always an unsettling feeling since every app behaves slightly differently. If I right-click this or that, does it show the browser's own context menu or something from the app? If I do this or that, does it behave as I expect it to do?

Using a user interface (be it HTML5 in the browser or a GNOME app in my KDE desktop) that isn't designed for the HIG of the platform I'm using feels like walking through a minefield.

And frankly, normal people don't give a shit about interoperability, openness or any of that stuff we might consider important. They care about a device that is fast, simple to use and provides the apps they want. That's a reality that Firefox OS needs to face.

Thanks for highlighting the least emphasized but the most important aspect of the operating system, which is being completely free and open. Too often I see people criticizing Firefox OS because it looks slow and the current hardware is inferior to whatever device they have now, but to me, that's completely besides the point. The market needs an open mobile OS that respects the user's liberties, and it's only because of Firefox OS that I finally went for my first smartphone.

The OS definitely has a really long way to go and I've noticed quite a few bugs and plenty of missing essential features, but I'm optimistic about its future, provided the developers acknowledge these issues. I think people also need to be made aware that as development continues, they will be able to install the OS on other more powerful devices if they fall outside of current low end, cheap smartphone market.

On a side note, thanks for sharing the link to your guide because I've been looking all over the place for a good resource to help me start learning how to develop web applications.

> The OS definitely has a really long way to go and I've noticed quite a few bugs and plenty of missing essential features

What are the (significant) bugs and missing features? I had guessed that, being in production, the basics were finished.

There are too many list, but I will name a few. If you want a thorough list, head over to the Firefox OS section on Bugzilla.

-Default applications, such as Settings, will intermittently go completely blank. If you zoom out to list all applications, it will look normal, but when you zoom back in, the screen is completely blank.

-Group messages (MMS) are barely functional. You can receive them and initiate them in an unorthodox manner, but you cannot partake in group conversations, nor are they properly displayed as a single thread.

-You cannot add import .ics files to use in the Calendar application. You cannot create recurring events. It's as basic as a calendar can get.

-The clock application doesn't display world time.

-The OS will claim that a SIM card is not inserted when you power on/reboot the phone, even though you can still make calls and send texts. You need to enable and then disable airplane mode in order for it to acknowledge the SIM card (necessary for the Usage application).

-The browser doesn't support extensions. So far, you have tabbed browsing, and you can clear your cookies and your history. That's it.

-Since I last updated my phone, it returns to 100% brightness every time I wake it up from standby. This means that I essentially use the display/brightness settings as my home screen so that I can lower the brightness every single time I want to use my phone (this might be ZTE's fault, but I didn't have this problem before the recent update).

Even my old feature phones from many years ago support some of these features, like proper group messaging. Some of these bugs and feature requests have existed for nearly 2 years on Bugzilla. Nevertheless, I'm enjoying the phone, and I love it when the Marketplace gets updated with new applications to check out. I'm really looking forward to the 2.0 update, and I hope that ZTE rolls it out to the Open C, unlike how they completely abandoned support for the original Open.

I have used the OS and I can confirm your headaches, and maybe even worse (because I've used 1.1 and it was far more bad than the current release version).

They released it too early. It wasn't ready. It still isn't ready, from what I've seen. I'm aware that Android was in a pretty bad state in its own v1.4, but when they release a mobile OS today, it has to put up to expectations set forth by others. No one is going to say "it has so many bugs and far worse functionality than Android/iOS, but it's okay because it's only v1.4" simply isn't going to fly because users expect what they're accustomed to.

And the choice of technology is a little problematic. How can you ever implement Skype, for instance, in Javascript? (But with acceptable performance, naturally.) I guess they have to improve asm.js support astronomically in order to enable such use-cases.

I've developed for FxOS (you can find an app I made on my Github) and it's been really fun and liberating (I have an Android phone so I tried developing for it but I don't like Java, especially the way Android uses the language, and it's been more or less a horror to work with, sadly) but there is a lack of standardization, and supporting docs and building blocks are still not up to my expectations, which is exactly where Android (and I presume iOS) shines. Then again, Javascript is not a bad language. With first class functions, for example, I consider it rather powerful and easy to work with. (But easy to fall into the inefficiency trap, I suppose.) Myself, being a guy who has been writing C/C++ and PHP for scripting it was a great break from routine.

> How can you ever implement Skype, for instance

This is why http://dev.w3.org/2011/webrtc/editor/webrtc.html exists.

I was talking more about video encoding/decoding in real time as Skype uses a proprietary protocol and encoding scheme IIRC.

That is what webrtc is for too, without the "proprietary" part.

My point exactly. Skype is a big P2P network using proprietary schemes for just about everything. They don't seem eager to interoperate with open standards in the slightest.

  > The market needs an open mobile OS that respects the
  > user's liberties

How do you know that?

"that we can all use shared technology that is interoperable instead of doing this 1990s flashback of Obj-C here and Java there thing"

1. JavaScript and HTML are themselves a 1990s flashback.

2. We want a choice of technology, not being pigeonholed into one "good for all" thing. JavaScript sucks, and HTML sucks equally.

Thanks! I'm sure no one before has even thought that "Javascript sucks, and HTML sucks equally", but here you are, enlightening with your rock-hard arguments and all.

You need Windows and a license to develop for Windows Phone. You need Mac to develop for iOS. Android is more open, but you still need pretty hefty computer to do anything Java.

But anyone can make webpage. Anyone with Firefox can run the simulator, and pretty much anyone with text editor can make FxOS app.

Yes, not having native access sucks for some purposes. I dream of making FxOS version of OsmAnd - offline OpenStreetMap application, but I still don't think it's possible on FxOS. And yes, I'm sure that anyone would rather use Haskell or Python or Ruby or Clojure or anything other than Javascript. But HTML5 and js are the price of accessability of FxOS.

Asking developers to buy licenses and tools is like asking students to pay fees and buy books. But telling developers to only use/target one language is like telling students to think the one and only "correct" way. This is the current message of Mozilla: "JavaScript and HTML are the languages of the web and there is nothing you can do about it. Use/target or go away."

Yes, but those are only available to some (maybe rich, maybe Western, maybe others) people. Mozilla apparently thinks that being open to every person is more important than being open to every programing language.

How are they imposing the choice on you? Android imposes on you the choice of Java. Apple imposes on you the choice of ObjC. Mozilla wants to make it easier to get already skilled developers (I'm guessing a lot of people are more familiar with webdev than with Android or iOS development). In my view, having technology which can work on a wide variety of different devices lowers segmentation thereby helping the system as a whole.

Javascript and HTML5 might not be ideal for the job at hand, but you're welcome not to use them. Just like you are welcome not to use Java for Android if you don't like it.

I can use any .NET language (C#, F# supported well) to target Android and iOS apps _natively_. I can use many (any?) JVM language (Scala, Clojure etc.) to target Android without transpilation. There is a very real choice.

"Javascript and HTML5 might not be ideal for the job at hand, but you're welcome not to use them." - how can I build a Firefox OS or even a just a web app without transpilation (which is like being a 3rd class citizen)?

I would argue that compiling to asm.js isn't exactly transpilation if the Javascript engine is trained to recognize asm.js and specially execute it, optimizing for speed. What is "3rd class citizen" about that? You can draw to a canvas from C code compiled to asm.js if you desire, or manipulate DOM through a thin layer of JS abstraction, if that's your wish. Sure, it may be clumsy because currently there are no good libraries and frameworks covering those use cases, but you can't blame the technologies for that.

I think you have really strong opinions :-)

I don't know if you're aware of emscripten, I believe you know some of it because you spoke of transpilation. Basically if you're coding in C/C++ or something that can go thru that route then you can pass it thru emscripten and achieve up to 1.5x performance on the ASM.js based output.

This is a really good way to reuse code that was "native" on the web. As of now, you can't use .NET or JVM based languages since you don't have the VMs for them and IMHO trying to pass the VMs thru emscripten is a recipe for nightmare.

The main thing is that everything has its pros and cons. There is no silver bullet. Developing with Swift for iOS appears to be a really good thing. The language is very modern and beautifully integrated but it only works on iOS. Same thing goes for JVM based languages, not all platforms are shipping JVMs (Android, Sailfish, BB10, who else?).

If you stop and reason for a bit you will notice that the most powerful integrations and languages are usually tied to a single platform or compiling to some runtime that was compiled for those platforms. Its really hard to create something that works everywhere.

Languages are just tools. When you build a beautiful shelf no one asks what was the brand of the hammer used. Lots of programmers tend to mix the process with the solution. Yes the .NET languages are wonderful, specially F# but how can you empower people with it? How can you make something that people can afford and how can you make it simple enough for them to use.

There are better languages than Javascript and it doesn't matter what is the current pinacle of language design, someone will replace it some day. HTML5 is currently the only toolkit that works web wide and is not controlled by a single company. There is power in standards and also continuity.

Just imagine for a bit a poor community classroom. Pick a smartTV from a government program, pick cheap tablets and maybe phones. This vision is getting more common than people think specially in the communities that are close enough to cities but not a part of it. Now what is easier: to pick developers to deliver to all those platforms regardless of OS using HTML5 or create separate teams to build stuff using whatever is the most elegant language on that platform?

Firefox OS is not about the process, its not about competing with Android and iOS, its not about pitching JS versus any language, its about creating opportunities and serving people that are currently not served by the other mobile platforms....

"its not about pitching JS versus any language"

vs

"flashback of Obj-C here and Java there thing"

?

While this is awesome, the bigger problem in India is that data plans are often prohibitively expensive, and the people who can afford data plans prefer swankier phones (since the phone is a status symbol). But this is a good start, so best of luck :)

Is Indian cellphone provider market monopolized? Isn't there a chance that with higher demand competition will lower prices for data plans in India?

Yes, the Indian cellular market is competitive and aggressive.

There is certainly a good chance that data plan prices will lower with higher demand.

Data plans are not too expensive in India. One can get 1GB plan for 3G data in Rs.139-Rs.160 ($2.35-$2.7) depending on the service provider. A lot of people are not using data plans as they don't have good phones. Most of the cheap handsets in the market are very slow and lack features that needs data.

Just back from a trip to India (from there originally), and don't quite agree with you that data plans are expensive. There is a lot of competition, and both talk time and data are quite affordable IMO. I pre-paid for 1GB of data and ~4 hours of talk time for just 250 INR (about $4).

Pretty much everyone in India - rich or poor - has a mobile phone these days. However, unlike the service, phones cost just as much as they do elsewhere so the poor usually end up using crappy feature phones with limited data support. A $25 smartphone with full web support would make an incredible difference IMO.

Presumably, a thin client will transfer more data than a thick client. Potentially, making that cost significant.

Though maybe it has some clever caching system that reduces this?

I wouldn't call FirefoxOS a "thin client" (if that's what you meant) - it has a concept of "packaged apps" in which the app and all its assets are packaged and downloaded to the phone, and loaded from there. The built-in "certified" apps (like Phone, Music, Calendar and so on) don't require loading anything from the network at all, for instance.

You can think of FirefoxOS apps the same way conceptually you'd think of them on Android/iOS, except they are always written in open web technologies (HTML/JS/CSS).

App developers do have the option to provide "hosted" apps to the store (which is basically a regular web page but without browser chrome), however even in this case writing HTML5 apps that work offline is totally doable using existing cross-browser standards.

it really depends on the app developer. He can use packaged apps and go full offline or use HTML5 AppCache and go offline with annoyances :-)

In any way, the developer is in charge of how much bandwidth he uses.

Well the poor don't necessarily end up with bad phones, there are a few indigenous phone manufacturers like Micromax [1], Maxx Mobile [2], Spice [3], Intex [4] etc. that have carved out a market for themselves in the past few years. These brands manufacture phones with features quite comparable to "good" smart phones and are comparatively very affordable. So, unless the consumer is brand conscious, its possible to get a smart phone with decent features and data support at affordable prices.

In fact, Spice and Intex are the manufacturers that Mozilla has partnered to produce the $25 smart phones.

[1] http://www.micromaxinfo.com/mobiles/smartphones [2] http://www.maxxmobile.in/mobiles/smart-phones [3] http://www.si2imobility.com/spicemobiles/ [4] http://www.intexmobile.in/Product.aspx?PCatId=3

You are forgetting about the cost of living and average income. While 250 INR ($4) is not that big an amount for you. Spending that much amount for just 1GB of 3G data may not be feasible/easy for a lot of people this phone is aimed at.

Like someone mentioned above, if you can afford the data plans, you would never go for this phone. The biggest hurdle I can see for this phone will be providing access to the internet.

How common are free wifi hotspots? I assume the FF phone will still have wifi.

New govt plans free wifi in public places. Would be a boost to smartphone industry. http://articles.economictimes.indiatimes.com/2014-06-09/news...

The phones have WIFI

Data plans are getting affordable now. State carrier BSNL offers 1GB for Rs 139. Thats roughly 2.341 USD

I'd like to develop for Firefox OS, but what are my options if using terrible languages like JavaScript are a non-option for me?

Use a decent language like JavaScript. You are correct that terrible JavaScript is available and widely used, but all of the platforms support both, so you may as well use decent JavaScript instead. :-)

Sorry guys, HTML is faaaar from being proven technology for the apps. And it will stay that way, sorry to dissapoint you.

On the other hand: why this obession to have mobile apps built with HTML and JS. Why not fight for desktop, if you are so inklined. That does not make it any less stupid, I am just curious about the reasons.