Phoenix Is Trying To Be An Open Version Of Apple's Swift

[wizard69]

It is interesting that GNUStep never generated any mind share in the open source world. I'm not sure why either.

Yawn,
I've been following GNUstep for about 5 years now. No progress made on making a usable desktop (come on, where is the System Preferences plugin for network-manager or hell ifconfig, and a plugin for PulseAudio/ALSA. 0 progress on implementing Tables and SimpleWebkit ), half rotting applications with no focus around what they're doing and an API that frankly is useless outside of OSX, unless you want to run an application that looks out of place on every current desktop.

There is a problem with that? Seriously I run all sorts of different apps on my Mac built with a bunch of different tool kits. Looking out of place isn't a problem if the software is useful.

Sadly it didn't have to be this way. With a more concentrated effort on getting the simplewebkit implementation and Vespucci up to snuff, and someone refactoring terminal.app to support tabs (something OSX has had since version 10.1 at least.) GNUStep could have made a pretty attractive development/desktop environment. But as it stands, it's not overly useful, and running individual apps as standalone, I think I'd rather see apps rewritten with a GTK or QT interface.

What would be the point the world already has GTK and useless QT apps.

It's a shame because the potential of GNUstep is so much better than the other APIs but the lack of development on the end user front means the API will almost never be used outside of weird projects like phones and obscure applications.

Hey the only way to make it better is to contribute. I don't expect it to ever take off though. The problem is pretty simple really, if the likeness to Apples software isn't enough to engage developers there isn't much more that can be done at the moment.

Interestingly Swift offers potential here to pull in more developers to the GNUStep project. In a nut shell a lot of people don't like Objective C.

[name99]

Do you know that for certain? Or are you speculating?

I think it's a reasonable assumption. It's the way I would bet.
Apple has SPECIFICALLY stated that Swift code right now is in a state of flux, and they make no promises about backward compatibility. This is precisely so that they can change things that don't work well or turn out to have been bad ideas. (This has already been done with some fairly aggressive changes to the semantics of arrays.) A schedule was also given for how long they expect this period of churn to last, but I can't remember the details.

It is this expectation of change rather than any sort of malice that is, IMHO, preventing them from releasing something today.

One thing that I think is obvious, and which vitiates any open source attempts, is that there is one glaring hole in Swift as it has been announced today, namely the plan for parallelism support. A language coming out in 2014 cannot be taken seriously without this, and while blocks and GCD cover one small subset of parallel patterns, there is a huge range of other important parallel styles, from SIMT to parallel for to pipelines, along with the other machinery one occasionally needs (barriers of various sorts, sync primitives of various sorts, marking variables as "shared" or not, etc). I cannot believe that Apple is unaware of this and does not have a plan, but they presumably have good reasons (which may range from lack of manpower to not wanting to overwhelm developers to requiring particular hardware [eg HW TM, or HSA]) for not yet making this stuff public.

(There are other possible holes in the language that may or may not be considered important and needing to be fixed before release, for example some sort of *built-in* database communication/modeling language rather than simply passing strings [which the compiler is blind to)] to APIs.)

[alcalde]

Why would you say that. It is like saying Go or D or Haskell are a waste of manpower. I may never use any of those but that doesn't mean other wont.

Besides Swift gas a lot to offer the development world.

It's really a mixture of Python, Rust, Go, with a few other things from PHP etc. and static typing. We already have things like that with Genie, Boo, etc. I'm not sure it has anything to offer the general development world. It has a lot to offer the walled garden of iOS development because it's much more simple, readable and modern than Objective-C. The rest of the universe has Python, Ruby, Go, etc. and doesn't really need Swift.

[mdias]

The rest of the universe has Python, Ruby, Go, etc. and doesn't really need Swift.

Sure, except Swift is not competing with the rest of the world. Only with Android which has a virtual machine uncapable of working with an unsigned int for example... Only God knows why google chose to keep android Java-only...