I write iOS apps for a living, and appreciate what Objective-C and (especially) Cocoa have to offer, but why would anyone choose to use Objective-C when there are so many other great alternatives available?
To consider just one example of the pain Objective-C developers like me have to deal with on a daily basis, why would I want to write:
Relative runtime speed or compatibility with an existing Objective-C code base that you want to easily call from your web app are two possibilities that spring to mind.
This isn't to say that conciseness of expression isn't a useful attribute, but I don't think it's so utterly critical to web development that the any tool suited for the job must provide such things as concise dictionary syntax.
You use the best tool for the situation, and I can at least conceive of situations in which I might care about the benefits that come with using Objective-C et al more than I care about how verbose dictionary initialization is.
> I can at least conceive of situations in which I might care about the benefits that come with using Objective-C et al than I care about how verbose dictionary initialization is.
Fair enough. I'm interested to hear more about what these situations are.
For what it's worth, my point isn't limited to dictionary initialization. That example, though, does represent a lot of what makes Objective-C painful for me. That list also includes its extreme verbosity; the need to box and unbox native types when storing them in containers (or passing them to performSelector: and friends, and other uses); the necessity of separating header and implementation files; requiring forward declarations of methods to avoid compiler warnings; simply adding an ivar as a property requires touching code in a minimum of 3 different locations, and potentially across 2 files (@property statement, @synthesize statement, call to release in dealloc (until ARC), and, pre-iOS 4.0, actually declaring the ivar in the class definition); and, of course, manual memory management, which includes worrying about fairly esoteric edge cases like retain cycles.
None of these things are hard, including memory management. But it is code that has to be written, and more code means more bugs. In almost all other mainstream web languages, the computer takes care of these things for you, as it should.
Tack onto all of that the fact that if you're running Objective-C/Cocoa as a web app, you're probably on your own in terms of finding solutions to problems from the community online -- at least compared to mainstream languages -- and it feels to me like it's hard to justify the tradeoff unless you have a very specific need.
Objective-C's incredibly verbose syntax for array and hash manipulation is my number one pain point with the language. A little syntactic sugar would go a long way here.
Unfortunately XCode's definition of word characters doesn't include $, so cmd-double-click on an identifier like $dict doesn't work. Anyone know a fix?
I've actually enjoyed using Objective-C and Cocoa when used for Mac OS/iOS development (the dynamicity of Smalltalk data structures with the control flow of C), but I don't think I'd ever voluntarily choose it outside that context. Too verbose, too fragile.
I'm not convinced that Cocoa and Objective-C are the best approach for anything.
The only reason I use them is because they're what you have to use to make good apps on OSX/iOS.
There are debates on both sides when it comes to things like ruby-vs-python-vs-node or rails/django/aspmvc and reasonable people can take opposing views on those.
I don't see much of a case for Objective-C/Cocoa being a better language and platform than any of the others except for the fact that it's the only one we've got. (And yes, maybe I'm being unreasonable on this)
The verbosity of the Objective-C language (and related Cocoa libraries) and manual memory management are the biggest reasons.
It's not a bad environment and has a certain weird elegance to it, but I would venture that very few people would use it if not for the fact that it's the defacto development environment for doing OSX/iOS development.
It's C, for heaven's sake. ;) Compared to any dynamic / memory-safe language in history. Oh, I don't know--Perl, Python, Ruby, Javascript, Lisp, Haskell, oh, and Java, I s'pose.
I can't bring myself to think of C as being fragile. Fragile is exactly the word I use for Perl, Python, Ruby, and Javascript. Lisp and Haskell, and maybe Java, are the only ones on your list I would say are more robust than C or Objective-C. I guess it's a difference of opinion, or maybe just bad experiences with other people's code.
Besides that, you mention dynamic-ness and memory-safety: these are exactly the two places where Objective-C is miles ahead of C! I don't even know what to say.
Like Appcelerator's Titanium? [1]
Doesn't exactly create very "classic" Cocoa apps, but considering the App Store, Lion etc, that seems to be the way that Mac Apps are going anyway, i.e. looking like you were born on the web or on an ipad won't hurt your sales that much, if at all. (Not too fond of that myself, but well…)
Ohhhh, yes! This sounds awesome, I've been meaning to write something like this myself. Essentially, a node.js equivalent, but for Cocoa. Of course, I'm way too lazy to add that to my work-load, but I'm very glad somebody did.
Now, if only I could use this in an Amazon EC2 environment. I guess with OS X Server supporting virtualization now with Lion, it might be a possibility?
I'm thinking about the fact that neither JavaScript on V8, nor Cocoa are equipped with HTTP networking. So the framework supplies that part. A big part of node, when used for a web app, is that it supplies an HTTP server. Same with this.
Cocoa (and CoreFoundation) actually make it pretty trivial to write HTTP servers already. I was completely surprised by how painless it was when I did it recently. Not quite as easy as Node.js, but on the same order of magnitude, IMHO.
Yes, that's essentially what WebObjects was. However, WebObjects today is a mess, and it's written in Java.
Also, we have a ton of conventions today, with regards to MVC development, REST APIs, etc. that we did not have in '95. Which means that therefore WebObjects is not suited to modern web development.
The only company still using WebObjects in production today is Apple.
Your statement that Apple is the only company still using WO has no basis in fact. There is even an annual conference[1] for WO developers, in which Apple has zero involvement.
While I do like the thought of using CoreData and code I've written for other apps on the web, the following sentence immediately made this thing much less interesting:
"WebAppKit runs exclusively in garbage collection mode to minimize the memory management burden."
I've never liked garbage collection, if you find memory management a burden, you're doing it wrong. And now that Apple introduced ARC even Apple recommends against using the GC.
My question would be: Can those web applications can only run on Safari with Mac OS X?
If yes: Isn't this contrary to what a Browser wants to do? To Run Web Apps in any Browser.
If it cross compiles down to html/js/css it's just fine but the Readme isn't giving much information.
Not really , Cappuccino focus consists in bringing the whole Cocoa/Objective-C style to front-end development (actually you can run Objective-J on the server too). On the other hand, WebAppKit goal seems to be using Objective-C in the backend
Cool. Anything similar for *nix and Windows? I'm fairly certain you can embed WebKit in GTK - in fact, I think that's functionality exposed by the API.
I'm not sure to what extent it's supported with other language bindings, though.
I write iOS apps for a living, and appreciate what Objective-C and (especially) Cocoa have to offer, but why would anyone choose to use Objective-C when there are so many other great alternatives available?
To consider just one example of the pain Objective-C developers like me have to deal with on a daily basis, why would I want to write:
when most other mainstream web programming languages allow me to write something like: