25 comments

[ 2.7 ms ] story [ 54.2 ms ] thread
Their discussion of how Apple is trying to cripple HTML5 is rather laughable. Apple's the main driver, along with Google now, behind WebKit, the main mobile HTML5 browser technology.
Not to mention how the iphone originally only offered webapps.
I used to work with a guy who is now the CEO of one of those "we can make your app work on any platform" (...because we use Javascript/HTML/CSS etc) companies.

They are making a very successful business out of selling this product to big, very well known brands and businesses through their ad agencies. These are the guys who control the cash, and these are the guys who are still getting high off of native apps.

This kind of shift has got to filter through so many layers of people - although I agree, I just think its going to take a while.

I guess the (other) important thing is to remember is that the best approach/product doesn't always win.

(I'm going to read the article now, FYI)

First sentence: "Over the past two decades, the mobile industry has become increasingly stunted by fragmented protocols, standards, and regional differences."

If the last few years have seen a stunting of the mobile industry, I'm very excited to see what HTML 5 is going to do :P

What was the mobile industry two decades ago (in 1991)? I'm pretty sure that the mobile industry has converged heavily on ~3 major platforms. I remember many closed, underdeveloped mobile platforms in the 90's.

This whole article seems uninformed.

If the evangelists from the article say html5 will overtake native in 2 years, then is the likelihood more along 5 years? And who's to say it'll use the web for distribution? Consumers aren't using Google to find apps, they're using the App Store.
"Native apps are faster for some operations"

s/some/all/

s/all/some/

If you need significant CPU cycles or access to non-local resources, the web apps will probably be faster.

Any XHR that a Web app can do, a native app can do as fast or faster.
Apparently "all" doesn't include "loading", because one of the things that aggravates me most about apps on my iPhone is spending 3-5 seconds looking at a splash screen just to access one bit of data.
That's absolutely not true.

I had the misfortune of working with Outlook on a virus checker infected system recently. That would take minutes to open. In the same time I could open a browser, open Gmail (or indeed the outlook web interface) and send an email, easily.

Same with MS Word - misconfigured printer support on the same system would force it to time-out waiting for a printer before I could use it.

In theory, native local apps are fast. In practice, with real world computers they often suck.

My feeling is that these fantastic assumptions of HTML5 being the best thing since sliced bread and trumping all the benefits of native support assume a vacuum and lack of innovation in the native client space. Even in this vacuum, HTML5 browsers and embedded views have a long ways to go before they can compete with current native feature support, performance, etc. from a consumer's perspective as far as I am concerned.

I have a strong desire to see more HTML5 based applications going forward, but I do not see them overtaking or marginalizing native applications in any substantial way within 2-5 years. I also feel we can obviously expect many developments, improvements, and features to evolve in the native application space that will create further incentives to produce native implementations and that we can not assume that HTML5 implementations will be able to keep pace with them.

The running motif of Apple = Joker, Google = Batman is hardly a good way to get an intelligent conversation going...
This article is in dire need of editing. It's too long, and the metaphor is too stretched (and dated) and the author goes off on a tangent to rag on Apple when that's really not core to the subject.
The problem w this article: The killer app for any computing platform from the dawn of time has been games -- now try making a full featured video game with HTML5. I'm sure you can do something, but to really tap the potential of the platform you have to native.
I built an isometric strategy game in HTML5 and had to give up because of lousy performance in most browsers (barring Chrome).

Maybe my code sucked, but I'm not a complete amateur either...

Well you can build a strategy game now with old fashioned HTML - in fact I use to play Space Empire as a door and that was all done with text!
I know that - I did just that in 2005 and it still runs, albeit with a tiny (~200) community.

What I meant was a game with fancy pixel graphics.

comparisons like are really quite trite.. the team I'm on makes heavy use of both technologies mentioned in the article and right now we have much better results with native code when it comes to mobile applications.. but it really depends on what you need to do

also I have to say viewing it as a competition is pretty ridiculous to me

"Native" software is not going to be killed by hosted web apps, as it still offers users one thing 3rd-party-hosted SaaS never will, full control of the data it generates and manipulates, an empowerment that hosted services are unable to match due to financial and political pressures.

That said, it is rather obvious that the use of software which offers the option to store its data exclusively on equipment owned by the user (software which may increasingly be based on HTML5 languages and frameworks as well), will certainly be much more limited as bandwidth becomes more available, simply thanks to the convenience of not having to manage storage.

This is the kind of analysis you end up with when you assume an exponential change will continue essentially unbounded.