Poll: Which platforms do you develop for?
64 comments
'Web' is a bit ambiguous here. It could plausibly mean either writing http services predominantly used by native apps or client-side browser development, which are really quite different.
The original question was better, I think, because it was clearly restricted to which client platform people are targeting.
I'm surprised that the ratios within purely client mobile platforms haven't really changed that much, apart from a few new marginal entrants. Given what I've seen from freelance offers and job ads, I'd have expected the iOS-Android ratio to have shifted further in favour Android than it has compared to two years ago, when it was basically all iOS.
The original question was better, I think, because it was clearly restricted to which client platform people are targeting.
I'm surprised that the ratios within purely client mobile platforms haven't really changed that much, apart from a few new marginal entrants. Given what I've seen from freelance offers and job ads, I'd have expected the iOS-Android ratio to have shifted further in favour Android than it has compared to two years ago, when it was basically all iOS.
What I often hear is that Android has the bigger installed base, but iOS converts a lot better.
There's probably the matter of dogfooding... many developers in the US and Europe are huge Apple fanboys. That's my subjective opinion and I wonder if there's a way to validate it (mobile browsing metrics on Hacker News? :) )
There's probably the matter of dogfooding... many developers in the US and Europe are huge Apple fanboys. That's my subjective opinion and I wonder if there's a way to validate it (mobile browsing metrics on Hacker News? :) )
Maybe I'm a bit biased as a more seasoned iOS dev, but I think that the entire development experience for iOS is a lot more fun than Android. This runs the gamut from using Xcode and contributing to the community, to launching apps and talking to users. Android, although open source, isn't easier or a whole heck of a lot of fun to program for. Not every Android class is documented (just yesterday I had to read the source code file for AnimatorPath because no top level documentation was available). One step of the myriad of steps to get push notifications working is to copy a string of characters between two colons of the URL bar in the browser. C'mon Google. If you want to make it better for developers, start by making it easier.
Nobody really talks too much about how much "fun" it is to develop for their platform of choice. I think this is a huge factor - one that outweighs most of the objective reasons for using one daily.
Nobody really talks too much about how much "fun" it is to develop for their platform of choice. I think this is a huge factor - one that outweighs most of the objective reasons for using one daily.
Android users in their mass are less willing to buy apps (pay for something they cannot touch and that is available for free via "alternative sources"). Especially in developing world, and that is a lot of the Android market share
There's plenty of metrics to shop that iOS users use their devices as mini computers, despite Android market penetration.
I personally would point to the fact that iOS is a app launcher first and foremost. Sure... Apple would love everyone to use the App Store, iTunes and first party apps... but they're the same pixel sized icons, can be buried in folders and their are numerous 3rd party replacements. Also, iOS Safari was the first real mobile web browser.
I personally would point to the fact that iOS is a app launcher first and foremost. Sure... Apple would love everyone to use the App Store, iTunes and first party apps... but they're the same pixel sized icons, can be buried in folders and their are numerous 3rd party replacements. Also, iOS Safari was the first real mobile web browser.
This just strengthens my view that the open web is the future, although a lot of people will tell you the exact opposite.
Given this poll is on HN, where there appears to be a heavy bias towards start-ups, MVPs, and so on, I would have bet my house that the biggest response here by far would be Web or Web-related technologies.
Whether the wider industry would follow the same pattern is a different question. I doubt anything like as many people here develop embedded software, but I bet most of them have a washing machine, microwave oven, digital watch, car...
Whether the wider industry would follow the same pattern is a different question. I doubt anything like as many people here develop embedded software, but I bet most of them have a washing machine, microwave oven, digital watch, car...
Volume doesn't always mean it becomes the industry standard. (Obvious, but hear me out)
Embedded has never been very capitalized precisely because nobody cares what the software on their microwave is doing, as long as it isn't burning their food or starting fires.
There is a push right now to get some more openness into embedded devices: Nest thermostats, smart watches, and the like, but each of these must compete with dumb devices that can be made cheaply and just work.
Embedded is (to a first approximation) a red herring.
Embedded has never been very capitalized precisely because nobody cares what the software on their microwave is doing, as long as it isn't burning their food or starting fires.
There is a push right now to get some more openness into embedded devices: Nest thermostats, smart watches, and the like, but each of these must compete with dumb devices that can be made cheaply and just work.
Embedded is (to a first approximation) a red herring.
Heavily skewed. Primary audience here aren't embedded systems programmers and the like.
Web has always been the most appealing platform, because it is free, open and easy to develop for. Most other platforms are proprietary, the tech is more narrowed down and developer needs more resources to start creating apps.
Oops, I clicked "Linux" before noticing the "Web" option (I build web apps which run on Linux).
If they are specifically for Linux then that is a correct choice. Web is not really a platform. My PC and tablet don't run "Web", they run a proper OS with a browser on top.
To me, a platform is how a user interacts with your app. An iOS user interacts with iOS apps, and Windows user with Windows apps. Web apps are to be used by web users, regardless of the OS their browser runs on. So, it's fair to say that someone who develops an app who's front end runs on the web, is developing for the Web/HTML5 platform, not for the what ever OS your backend happens to run on.
You may think about your self as a Linux developer, but you are developing for the web platform, because that's where your users are.
You may think about your self as a Linux developer, but you are developing for the web platform, because that's where your users are.
What if it's a Windows app with a seamlessly integrated web-browser - something like Awesomium? What about NaCl Chrome apps? HTML5 is a language, maybe a technology but not a platform. Web is probably something different as well as it can't exist without some other actual platform running underneath.
Windows App or Linux App that just happens to use HTML5 interface, but is installed and runs locally is a Windows or Linux Platform App. A website is necessarily a "Web Platform." Otherwise, you would have to argue that Google or Facebook primarily target Linux as their platform.
Picture this. An alien lends on planet earth and asks you a question "What does this Google company do?" What do you answer:
a. At Google they develop a search app that run on Linux?
or
b. At Google they develop the most popular WEB search engine.
Picture this. An alien lends on planet earth and asks you a question "What does this Google company do?" What do you answer:
a. At Google they develop a search app that run on Linux?
or
b. At Google they develop the most popular WEB search engine.
A Linux appliance that happens to have a Web UI is not web platform.
I'd think, if you are building the server side components you can click Linux and if you're doing frontend (HTML, JS, CSS) you'd click Web...
I disagree. I would say 'Linux' means writing apps specifically for Linux, using the Linux API etc. I write, amongst other things, web apps that run on Mac OS in development and Linux in production. I would tick 'web' in this case. Otherwise I would have to click 'Linux', 'Mac' and 'Web'.
(I don't disagree that strongly. It's very ambiguous.)
(I don't disagree that strongly. It's very ambiguous.)
Thanks! It's a shame this isn't the default view of the poll results on HN.
I thought that but I think it would influence voting
A few graphs would be nice there...
Bar graph so far:
http://quickhist.onloop.net/Android=57,BlackBerry=3,Chrome%2...
http://quickhist.onloop.net/Android=57,BlackBerry=3,Chrome%2...
Cool, what do you think about adding the chance to set a poll id directly?
Example: http://hnlike.com/hncharts/chart/?id=6603807
Example: http://hnlike.com/hncharts/chart/?id=6603807
Not that useful if it's not updated often enough.
I develop an operating system running on bare metal :). It's sort of an appliance, but I don't think Xeons really qualify as embedded.
I was thinking about how popular Go is and looking at example code which to me is fairly clean. Anyway Go seems popular and looks cool and I think that is the next thing I am going to learn.
But the point is, what I am doing with Web applications for the most part is making rich UIs. So I was thinking, since Go compiles so fast, why not just download Go source and escape the browser, and get faster execution etc.? I mean I still think ToffeeScript is the best thing ever but Go is not bad at all.
So what I was thinking of making would be some kind of P2P-based DHT thing, totally decentralized of course, using UDP, and you can use it to distribute whatever, but two main things would be something like a web site but it would just be markdown and a tarball with Go code. Then you would display the markdown (maybe let them embed svg) and a button "Run Application" and sandbox the Go code with something like this https://github.com/zond/gosafe. Maybe have a way to tag sites or some kind of data-oriented alternative to domains. Point being we are using the web to make it convenient to distribute and access sandboxed applications (even the front ends for e-commerce stuff are becoming thick clients in a lot of cases).
It would be nice if we could break out of the browser and also weren't subject to the limitations of JavaScript. And I wouldn't miss HTML or CSS at all to tell the truth. And the other part of this is that the whole server-based model we have needs to go and be replaced with data-oriented networking.
But the point is, what I am doing with Web applications for the most part is making rich UIs. So I was thinking, since Go compiles so fast, why not just download Go source and escape the browser, and get faster execution etc.? I mean I still think ToffeeScript is the best thing ever but Go is not bad at all.
So what I was thinking of making would be some kind of P2P-based DHT thing, totally decentralized of course, using UDP, and you can use it to distribute whatever, but two main things would be something like a web site but it would just be markdown and a tarball with Go code. Then you would display the markdown (maybe let them embed svg) and a button "Run Application" and sandbox the Go code with something like this https://github.com/zond/gosafe. Maybe have a way to tag sites or some kind of data-oriented alternative to domains. Point being we are using the web to make it convenient to distribute and access sandboxed applications (even the front ends for e-commerce stuff are becoming thick clients in a lot of cases).
It would be nice if we could break out of the browser and also weren't subject to the limitations of JavaScript. And I wouldn't miss HTML or CSS at all to tell the truth. And the other part of this is that the whole server-based model we have needs to go and be replaced with data-oriented networking.
It would be interesting to see how RWD replaces native applications on mobile platforms. This is something what we do at http://codedose.com to make our corporate intranet web applications accessible from smaller factor devices without charging clients for native app development.
I wouldn't say it replaces them. It's like saying that Toyota replaces Mercedes - just different quality/price niches.
Only a minority (< 10%) of web developers develop for Chrome Store/Firefox Marketplace. Why would that be? As far as I can think, the advantages of putting on these stores are clear and there isn't much friction to adapt the app for them. Am I missing something?
I have not met anybody outside of the development community who has installed an 'app' from the Chrome or Firefox stores. I suspect penetration is very low.
The download numbers for the popular apps aren't bad imo - https://chrome.google.com/webstore/category/popular. Still your hypothesis is an interesting one. Any possible reasons why this might be the case?
Only a minority (< 10%) of web developers develop for Chrome Store/Firefox Marketplace. Why would that be?
One of the major advantages of developing for the Web is the platform neutrality. You can rise above OS wars and browser wars, and provide something -- one thing -- that almost everyone can use whatever other choices they make.
Given the above, why is there any advantage for a developer in having a platform-specific store for web apps, and why would any user want to access web-based software via a store instead of just going to the site? What is the value proposition here?
The only thing I can think of is marketing, but I'm not convinced they offer much benefit there. Despite being a regular user of both Chrome and Firefox since they were new, not to mention a professional web developer, I don't remember ever hearing of Firefox Marketplace before this thread and only just read up on what it is. Although I'd seen references to Chrome Store when starting Chrome, I had immediately ruled it out as clutter that didn't belong in a browser and hid as much of it as possible so I could go back to visiting web sites. Certainly no client has ever asked me about using either of these channels for a real project.
Of course this is all just personal anecdotes, so it's possible that I'm an outlier. However, it seems like these channels aren't exactly Apple's App Store where most people get all the software they run on a given device. So I come back to asking why any web developer would develop for Chrome Store or Firefox Marketplace. What's in it for them?
One of the major advantages of developing for the Web is the platform neutrality. You can rise above OS wars and browser wars, and provide something -- one thing -- that almost everyone can use whatever other choices they make.
Given the above, why is there any advantage for a developer in having a platform-specific store for web apps, and why would any user want to access web-based software via a store instead of just going to the site? What is the value proposition here?
The only thing I can think of is marketing, but I'm not convinced they offer much benefit there. Despite being a regular user of both Chrome and Firefox since they were new, not to mention a professional web developer, I don't remember ever hearing of Firefox Marketplace before this thread and only just read up on what it is. Although I'd seen references to Chrome Store when starting Chrome, I had immediately ruled it out as clutter that didn't belong in a browser and hid as much of it as possible so I could go back to visiting web sites. Certainly no client has ever asked me about using either of these channels for a real project.
Of course this is all just personal anecdotes, so it's possible that I'm an outlier. However, it seems like these channels aren't exactly Apple's App Store where most people get all the software they run on a given device. So I come back to asking why any web developer would develop for Chrome Store or Firefox Marketplace. What's in it for them?
The changes required to adapt to these stores (I can speak for Chromestore) are minimal, i.e. one needs to write a manifest file. It's like adding a channel to improve the app's discovery. It certainly seems to provide a lot of traffic if your app falls in the top of popular category.
The changes required to adapt to these stores ... are minimal
The changes from what, though? I've just looked through both the Chrome Store and Firefox Marketplace sites, and it's still not clear to me what they actually offer to users, or what their business model is, or in Chrome Store's case how developers even get involved (assuming they can figure out what they are getting involved in or why they care).
Both sites seem to be unclear about almost everything, including what they do or why I should care as either a user or a developer. So it's not really surprising that most web developers here aren't targeting them.
The changes from what, though? I've just looked through both the Chrome Store and Firefox Marketplace sites, and it's still not clear to me what they actually offer to users, or what their business model is, or in Chrome Store's case how developers even get involved (assuming they can figure out what they are getting involved in or why they care).
Both sites seem to be unclear about almost everything, including what they do or why I should care as either a user or a developer. So it's not really surprising that most web developers here aren't targeting them.
Results of this poll as a bar chart in real-time(almost): http://hnlike.com/hncharts/chart/?id=6603807
it interesting to see which subset of developers are most common at hacker news.
its not really surprising considering the content and philosophy that Linux and Web are rating so highly - but its nice to see anyway. :)
its not really surprising considering the content and philosophy that Linux and Web are rating so highly - but its nice to see anyway. :)
For those who upvote Linux, could i know what kind of software you make? I rarely hear people who build desktop app for Linux.
What makes you assume that those who upvote Linux are writing desktop apps?
There are certainly people who work in QT or GTK on desktop stuff, or even those who work in Java and target all desktop platforms, but I would posit (based on my own experience) that many more work in the kernel on device drivers or on the CLI writing utilities.
There are certainly people who work in QT or GTK on desktop stuff, or even those who work in Java and target all desktop platforms, but I would posit (based on my own experience) that many more work in the kernel on device drivers or on the CLI writing utilities.
People might be thinking of scripts here. I did when I first responded.
There must be a better way to do this. I can't un vote and we can't see the breakdown in a pie/bar chart.
Maybe future polls should just ask HN users to use something like http://strawpoll.me
It's impressive really: Linux has(almost) more votes than Windows + Android together
NetBSD
What ever the bill-payer wants.
Java
VMS
No MVS or z/OS!?
Web
Linux
for web browser generally
BlackBerry!? - does it still exist?
Let me know if you would like to add a choice.
Edit: Added Embedded/RTOS.
Edit: Added Embedded/RTOS.
[1]https://news.ycombinator.com/item?id=3014502