I started building a desktop app last year, but realized that many features I'd come to expect from using reddit/HN (like spell-check, and familiar keyboard short-cuts) weren't there, and I'd have to re-implement each one to mimic the web experience. (Do native apps for mobile get these for free?)
Webapps have surreptitiously become my standard for app features.
On the Mac, you can enable continuous spell checking by calling -[NSTextField setContinuousSpellCheckingEnabled:YES]. On iOS, continuous spell checking is on by default.
If you want to put together really simple interactions, then a web page will get you off the ground fast. The problem is when you get to the details. Harder things become more difficult, or just impossible.
For example, I don't think there's any way to control a browser's spell checker, beyond turning it on and off for a given field. (Correct me if I'm wrong.)
In a native app, I have implemented functionality so the spell checker would ignore text with a certain pattern, but spell check the rest of the input.
I used to think the solution to the web’s gotchas was a framework to conform the web to a native app mentality. I built and deployed web apps with these stacks. I would never do it again. These frameworks build abstractions on top of abstractions. In every non-trivial app I worked on, the top abstractions leak, and you still have to hack HTML and CSS.
Web technology is great for many things. Replicating a native app experience is not one of them.
Why would I want a native app when Web apps are so much more awesome? After using Google Docs with multiple users I can't go back to Word (which is only available on systems for which I have license), Google Spreadsheets with multiple users (which again would also need a license).
There are certain apps like Photoshop, After Effects and Lightroom that still need native. But most of the rest, not so much. I still use a native code editor but I can see that disappearing soon.
I suppose one problem in discussing this is the image of 'native' app each of us has in our heads. Is it 'iTunes'? No longer need that. Google Play or Amazon Cloud music let met access all my music everywhere. Is it Office? For some users they probably still need it but most of them will have a better experience with something online.
Is it iOS apps? Which ones? Games? There's a few high-end iOS games. Most of the rest are available as web apps though or certainly reasonable clones. Is it all the native web substitutes like the G+ or Facebook app, the Twitter or Forsquare app. The Yelp, Amazon or IMDB app? All of those are a waste of bits. Frustrating because in their quest to simplify they always leave out important features. Waste because their mobile websites provide the same features.
The only thing hobbling those apps are a few more browser features and ability to put them in the app stores even if they happen to just be bookmarks.
> Why would I want a native app when Web apps are so much
> more awesome?
Because web apps are not awesome. Multiple users? Ok, that's a feature, not something very web specific.
Google docs always seem very very brittle to me.
Windows 8 also provides a consistent set of APIs across all languages supported: JScript, C#, C++ and VB.NET. There's nothing you can do in C# that you can't do in JScript on Windows 8, with the possible exception of DirectX.
sproutcore is probably the most complete framework to deliver desktop class apps. But I find the combination of Backbone.js and the Handlebars.js template engine very effective as well. They feel lighter and not so set in stone.
I`ve been using Backbone, and one thing I forget all the time and bothers me is when you use Jquery plugins on elements that get re-rendered from templates (then you have to initialize the plugin on the element again, etc - for example, something like the Jquery UI Slider).
The render method is the right place to hook things up with the DOM. Although I trust the garbage collection is up to par when it comes to forgotten DOM elements, I've recently started implementing destroy methods on views that deal with more complex jQuery plugins. Some plug-ins implement destroy methods themselves, but in some places I just unbind the few events manually.
Try breaking up your render method into smaller chunks, and only re-rendering content that needs replacing. You can even leave the el if it exists already, and just update text or form fields.
Agreed! AngularJS makes the code so concise and easy to read! I'm even considering adding specific support for it in my LIVEditor project (a live html/css/js code editor).
Oh demo… So it usually goes as this:
great demo, should try -> hey building some PoC is so easy -> hm, how do I get around this corner case -> hm, how do I implement this particular thing -> damn, how do I debug that? -> fuck it, we ar going native.
That might be your experience, but I don't agree as for Angular. We are building a "desktop application" using AngularJS just now. The application is a visualization of the status of multiple systems over time - nothing very complex, but still it is a real application. I have not yet been disappointed by Angular - I love being able to build my model, and then bind all my views to the model and all the views update automagically whenever one of the views update the model.
ExtJS from Sencha is lightyears ahead of any other JS framework as far as a polished UI toolkit goes. Its not free if you're not free and open source yourself, but its worth every penny.
I agree, this article feels like it was written in 2006. While the analysis of Obj-J is spot on, does anyone actually use it anymore? I would never consider it a viable framework, especially with the proliferation of top-notch, extremely active community-driven projects, like Backbone, Ember, Knockback, Spine, Angular, and so many others.
There appears to be a lot of FUD in that article which is unjustified or loosely cited. For example the following paragraph:
"Standardization is slow, so proprietary platforms are the first to solve problems. Flash, Silverlight, and JavaFX offer desktop like frameworks but carry technical flaws: the iPhone doesn’t support them, vendors may break things at any time, and the runtimes are often slower than JavaScript."
Most of it is rubbish.
However, I'm a firm believer that if you're going to build desktop class applications, you do it in a desktop class runtime, not JavaScript.
If you have Java installed on your workstation, click here to launch ArgoUML (UML editor). This is what I expect from a "desktop class application" from deployment to usability:
No offense, but your aesthetic sensibilities suck if you think this argo UML is a desktop class application in the realm of natively built desktop class apps.
These apps belong to what I refer to as the "just barely good enough" class of desktop apps. Doesn't look native, doesn't feel native, aesthetics are strictly academic stuck in the 1990's and usability is just north of nominal.
This is a prime example. Fonts, control spacing, inconsistent sizing, 8-bit aliased bitmaps ... it looks like crap compared to a native app like OmniGraffle.
No offense taken. Perhaps a bad example aesthetically I agree, but otherwise the experience is good.
Please note that Argo is a heavyweight engineering tool with metadata, not a pretty picture generator. It's an order of magnitude more powerful with respect to semantic modelling.
What it looks like is less important that how it works in that sector.
EDIT: also regarding native appearance. Every web site looks different - should have a common stylesheet for everyone? No.
>If you have Java installed on your workstation, click here to launch ArgoUML (UML editor). This is what I expect from a "desktop class application" from deployment to usability
Is this a joke? That would be barely tolerable for 2000.
Slow, clunky, non-native feeling, the bad kind of cross-platform, a UI beaten with the ugly stick and very little thinking about usability issues.
I've updated the preface to say as much.