Poll: Which platforms do you develop for?
I see a poll on similar lines "You and mobile development" asked two years back[1]. This would be an interesting exercise, given the rise and fall of platforms since.
[1]https://news.ycombinator.com/item?id=3014502
64 comments
[ 2.5 ms ] story [ 156 ms ] threadEdit: Added Embedded/RTOS.
We don't currently support Windows but we want to keep an eye on it.
(I don't disagree that strongly. It's very ambiguous.)
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.
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.
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...
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.
http://quickhist.onloop.net/Android=57,BlackBerry=3,Chrome%2...
Example: http://hnlike.com/hncharts/chart/?id=6603807
its not really surprising considering the content and philosophy that Linux and Web are rating so highly - but its nice to see anyway. :)
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 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 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.
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? :) )
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.
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.
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.
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.