What's with that Node.js part? Even if it is part of the backend it has nothing to do with the app, so i 'll go ahead and guess either the writer is clueless, or they spit out random trendy jargon in their articles.
I agree that the author has no idea about client vs. server, but reporters aren't technical, it goes with the territory. I think Prasad lead the discussion to node.js (the author journo wouldn't know to ask) so that they could talk about open sourcing tools, etc. It was a chance to get a message out.
Is it possible also that node.js is providing faster response times on the same number of servers?
"...LinkedIn just released a brand new iPad app built using HTML5, backbone.js, and underscore.js. The app includes a rich stream of information which is a combination of network updates, group posts and news articles..."
Their method is also how I have built most of my Android apps. You use a WebView for the user interface and then use the addJavascriptInterface[1] method to allow calls from the Javascript code back through to chosen methods in the Java code. You also need to use Java when you want native settings menu and system search functionality.
Making calls from Java back into the Javascript is a bit more painful since you don't get a response back directly, but is sufficiently useful to dispatch events.
You can bundle your javascript, images, html etc in the app assets so it will still work even if there is no networking. Depending on the app network requirements it is also likely you'll want to do the networking from Java so it can manage threads, connectivity, saving resumable state in SQLite databases.
This approach is fairly easy to develop and test (you can do a lot of it in a desktop environment). The drawbacks are that elements of your UI may not appear "native" and that some events are a pain from Javascript, especially dealing with swipes.
I must admit I thought a lot of mobile apps were developed this way, and hence this article being a non-story.
A nitpick on the title, cuz it annoys me. Why would anyone say "You'll never believe X"? The X in question (95% HTML) is actually not that unusual to begin with.
But even if X is mind-blowing, the fact is that the article was written to show readers this mind-blowing fact. And after reading said article, should I still be in the "never believe" camp, making it an unconvincing article? Or should I now believe it, thereby invalidating the premise that X is something "I'll never believe"?
Argh, sensationalism is very annoying. As a web dev who's used PhoneGap, and I'm glad to see it's becoming an accepted approach to mobile apps... But PhoneGap has been around for years.
Regarding responsive design: the best mobile apps/sites will have a specially crafted mobile design, but responsive design is a good solution if you lack the resources/need for a great mobile app.
Unfortunately, a lot of this reads as "Node.js IN the mobile app", not "Node.js backing the mobile app".
It'd be better framed as a triumph for well composed and efficient Javascript in a web view. (In fact, Javascript isn't even mentioned once.)
Perhaps, it's just not a publication I should be reading.
The takeaway should be as Prasad was quoted, "As long as we can make the experience fast enough, nobody can tell the difference. It still feels right."
[ Fully expecting my next client to ask for Node.js in their iPhone app because "LinkedIn did it." ]
Not being an ios dev, I didn't understand the tech they actually are using. Is the article saying that the linked in app is basically some sort of web container (webkit?) and everything on the screen is actually html/javascript being interpreted by the web container?
Didn't google put out a "gmail app" recently, which was basically a web container with all content in html? I remember there being a backlash because it wasn't a true native app. Did that app have a different problem?
Is it feasible to create nice, usable iphone/ipad apps using mostly html/js (but have them look like native apps)/
Go to slide two and you will see a massive amount of text content. Those instances are perfect examples of when a UIWebView is a perfect vehicle for organizing information that HTML and CSS do best. Several apps like this one do similar stuff when presented with marking up text blocks.
Apple's App Store app is mostly HTML too. This is a trend piece pointing out how web techonlogies is seeing more real world uses by big name corporations, though not a particularly new trend.
The App Store app is also aggravatingly slow, just like every other heavily-webview'ed app (Facebook, and now LinkedIn I guess).
Note that desktop apps that try to do this fare little better - iTunes is widely hated by its own users for being bloated and slow, particularly the store component, which is entirely webviews.
In the mobile space right now, these webview'ed apps seem to be heavily sacrificing user experience for developer experience, which is something I cannot agree with.
Right-o, except these webview-driven apps have all the weaknesses of websites and none of the advantages. A website, from the perspective of the user, has the advantage of being entirely frictionless - anyone with a browser can use any website anywhere without downloading, installing, or configuring anything, and it's always up to date. A webview-driven app still gets out of date, and still presents the significant friction of making the user install something... all the while enjoying none of the usability benefits of going native.
It should also be noted that the performance gap between web and native on desktops is at least two orders of magnitude less than it is on mobile. The difference between a website vs. a connected native app on desktop can be something on the order of milliseconds - below user-perceptible time.
The difference between, say, the iOS browser vs. a connected native app, is on the order of seconds, if not tens of seconds.
On a desktop with a broadband connection, the difference between a highly optimized 1KB JSON call vs. a 500KB full HTML-CSS-JS content blob is really a rounding error - below user-perceptible time. On a mobile that translates into seconds where the user is entirely non-interactive - and this is before we even get into the performance difference in rendering and input. Hell, part of the reason why apps have taken off in mobile (where they haven't on desktop) is because of this crucial difference. People hate using websites on mobiles, even mobile-optimized ones. The browser is such a poor experience on most mobiles that people will actively seek out the Store/Market, authenticate, and download huge binaries just to get the better UX.
I don't believe what he's saying about responsive design at the bottom of the article, and certainly the motivation about context and device seems trite. It's entirely unobvious to me that my intent is correlated with my device. Often as not, I slouch around on the couch or on a long trip with my iphone because I'm either too lazy or unable to use my laptop.
The article might make him a disservice though; I'd like to hear him elaborate in his own words about it.
No-one is advocating 'responsive design for apps' so I'm not exactly sure what he's talking about.
It is true that a responsive design for the LinkedIn website might be a tricky-ish project given the current (surprising?) complexity of the 'main' site.
So they used HTML5/JavaScript in a UIWebView with Node.js on the backend? Pretty sweet, I was just remarking to a co-worker how much I enjoyed the Updates section.
How do you calculated 95% of the app is remote? I am hard pressed to come up with an equation to even start with. local code size vs total code size required to run (where the remainder comes from HTML5 download?)
Let's not forget that building a quality HTML 5 interface takes as much effort as building a quality native interface - and you have to specifically build it for a specific platform.
I suppose the biggest advantage is that there's a lot more common code that can be shared between platforms which potentially could cut down development time for developing for multiple platforms.
I'm just going to say it: i don't like it that much.
Don't get me wrong it looks nice and it is a functional app but I instantly feel this is not a native app. The interactions are wrong. The scroll bounce is not correct. There are noticable delays. The back button does a fade instead of popping a viewcontroller. Little things like that.
I'm curious if they integrated Sencha touch with it too. Personally I use Backbone to layout structure, but still integrate it with Ext for specific components.
Does anyone know if LI used any specific Node modules? I'm evaluating Node.js for a project and am trying to get a handle on the thousands of modules available in npm.
"Today, however, Prasad no longer feels the need for the security blanket. “In the tablet version of the server, we’re still using Node, but now the clients are talking directly from the load balancers to Node, there’s no nginx.”"
I thought the headline was just another SEO attention grab, but it's actually factually correct; there's no response to this sort of thing other than bemused incredulity.
