A New Survey Outlines the Trends in Mobile Development for 2012

With iOS Apple’s operating system takes the spotlight with 89% and Android at 79%, I think that closes the debate of Native vs Web App.

Oh, but hybrid HTML5 is at 72%. While native apps dominate, many are being built using HTML and Javascript by Web developers.

Pure HTML5 is 6%, but pure native ones is also only 22%. The bulk will be hybrid ones, 72%.

Just to be clear, the question I am trying to get an answer for is "Native vs Web App". I would argue that you are either a web app or you are not, there are well defined properties of a Web App that obviously a native or "hybrid" app doesn't have. Web Apps are the 6% of the mobile space. I don't use any, and I don't know many users who use them.

Please define hybrid? If all you mean is Rich Content in a WebView, it does not qualify as "hybrid" for me. In that case, all the apps I write are hybrid.

Please also define how much of the HTML5 spec is used in the hydbrid model. After all it is not difficult to create a few menu tabs and a navigation bar in HTML, or even make a Web API call. Why do people chose Hybrid over "pure Web" (considering all the advantages "Web Apps" ;). They don't seem to go native just to access the GPS or the accelerometers. Are they?

The report defines hybrid as applications rendered in WebView, but using JavaScript to make some native calls when necessary. The bulk of the code is HTML5. They are not pure web apps because they are packed to be deployed via App Store/Market/etc.
The main differentiation between native and hybrid is if the app uses the native UI or not. Moving away from the native UI is a big step. When HTLM5 will provide access to all hardware, then there will be no need for native code.
Pure native apps will be done by those who will take the pains to write an app that renders in the native UI and looks exactly as the OS manufacturer envisioned apps to look on their devices.

The survey did not clarify how much Web vs. native code were in the "hybrid" apps. This would be nice to know, subjective definitions of "hybrid" aside.

> They don't seem to go native just to access the GPS or the accelerometers. Are they?

This is the route PhoneGap takes, it's mostly a Web solution with native hooks. Appcelerator Titanium, a "native" solution, uses Javascript syntax, catering to Web developers.

HTML5 (or really just "modern" HTML) is attractive because it's great for cross-platform development. It may even become more attractive as bundled browsers become more capable. The bottom line is that Web developers have a viable career in mobile app development.

>> The bottom line is that Web developers have a viable career in mobile app development.

That remains to be proven... It's easy to hope. The reality is that the development of mobile solutions has just become insane with a) the fragmentation of Android, b) the fragmentation of the platforms and c) the Web developers wanting to reuse their skills in in a context where "web" concepts don't fit well and d) the soon to be one billion smart phone/tablet users who can't care less about what you use to build an app as long as the UX is the best possible.

Personally, I am a bit tired to hear the HTML5 drum when all evidence point to the contrary. Even companies like Sencha or Appcelerator offer a 100% JavaScript programming model with no "HTML" inside. Sencha's tag line is "ma' look at my body tag, it's empty". Is that a "Web" app? show me the Web there! Is a Web developer automatically a Sencha or Appcelerator developer? no. Is a Sencha developer a Web developer? not much more.

>> rendered in WebView, but using JavaScript to make some native calls when necessary.
that's fair. I don't build hybrid apps then, but I am not sure this is the definition of a "web app"

>> When HTLM5 will provide access to all hardware, then there will be no need for native code.

How far are we from that?

To get a real pulse of the trends, I suggest that you look at Indeed Job trends:
www.indeed.com/jobtrends/iOS.html -> 0.4
www.indeed.com/jobtrends?q=HTML5+mobile&l= -> 0.14
www.indeed.com/jobtrends?q=Android&l= -> 0.48
www.indeed.com/jobtrends?q=Mono&l= -> 0.012
www.indeed.com/jobtrends?q=Appcelerator&l= 0.004
www.indeed.com/jobtrends?q=Sencha&l= 0.015
www.indeed.com/jobtrends?q=PhoneGap&l= 0.012

Native -> 0.73 (0.88-0.15 for jobs with both Android and iOS)
Web/Hybrid -> 0.17

75% native -> ~25% Web/Hybrid

That sounds a bit more representative and in the 25% how many "native apps" throw in HTML5 just in case? My guess is 30% of the jobs add HTML5 just in case.

It's hard to tweak Jobs numbers. It's a lot easier to say "Yes I think I'll use HTML5 next year"

As you yourself have noted, not everyone is really using the HTML5 spec features. Many are just using HTML.

www.indeed.com/jobtrends?q=HTML+mobile&l=

Though I also doubt the accuracy of the stats if you're searching 2 terms, e.g.

www.indeed.com/jobtrends?q=android+java&l=

In any case, as a Web developer who just wants to be employed, it doesn't matter to me if it's Web or HTML5 per se, but rather I can use Web-related skills such as good old HTML and Javascript.

In that case you'll have to sort out between mobile Web sites and mobile Web Apps. I thought I was generous associating HTML5 with Web Apps, but if you consider that a well crafted job posting would most likely contain a reference to a framework like Sencha, JQuery Mobile, ... the picture looks very grim for Mobile Web App Developers if you add all the job posts containing a reference to one of them.

I can see the same kind of people who told us that REST was great because it was from the Web, making the same empty argument with HTML5, yet everything changes with mobile clients:
a) people don't need to use the same app on different platforms, they bring their client with them, ad they stick to their platform for years
b) mobile devices are bandwidth and power constrained
c) mobile apps are used "in context": walking, in a bar, having some fun with friends, in a meeting, ... and require sn optimal UX each time, that has been crafted especially for that context, not just the one size fits all
d) Monetization strategy: I know that it has never really been a concern of the Web and as long as Web companies can sell your privacy we are good, but now with the knowledge of your location, privacy takes on a whole new meaning
e) Location is now at the center of the user experience: the Web has tried for 20 years to eliminate location (and that good, Amazon for instance gives you the ability to by anything wherever you are). But successful mobile apps have to understand how to use location to the fullest
f) Consumers can actually vote with their wallet now that they have a choice and there are 500+ M of them and I argue for any successful Web-based Mobile app, a native one will emerge that will win the market, simply because they can't compete effectively

(Just to give you an idea of consumer power, in 2003 I was told that SAP had about 15 million seats after 30 years or so in business, in the mobile world, OMG2POP got to 20 millions apps in what 3 weeks)

So I am not quite sure what we are talking about here, syntax, architecture, skills, is there still some people who believe that mobile apps are BAU and the concepts they have been using for 10 years will just propagate seamlessly in the mobile space? That's quite a naive thought process. Web Apps in the late 90s solved some important problems: ability to run on multiple (tethered) platforms, ability to distribute updates easily, ... but it took Google-size efforts to build a decent UX and 15 years later, desktop and mobile Web apps are still way behind their desktop counterpart.

Mobility eliminates 2 of the core advantages of the Web and brings a lot more control in the process of what is running on my device. So why would anyone think end-users will vote for something other than native? Only people who are emotionally and financially attached to the "Web" can think otherwise.