GNUstep

[TurboRascal]

Has anyone tried comparing Cocoa with GNUstep and how hard would it be to adapt Cocoa LCL to use GNUstep?

They claim their goal now is an implementation of Cocoa, not OPENSTEP anymore, so I was wondering...

[Phil]

A link where they say that?

And the point of this would be what?

Note there is no such thing as "Cocoa VCL".

Thanks.

-Phil

[TurboRascal]

A link where they say that?

http://gnustep.org/information/aboutGNUstep.html

And the point of this would be what?

Creating GNUstep apps?

Note there is no such thing as "Cocoa VCL".

Oops, I meant "LCL" 

[marcov]

And the point of this would be what?

Creating GNUstep apps?

Well, and the point of that ? :-)

[TurboRascal]

Creating GNUstep apps?

Well, and the point of that ? :-)

I give up on that

Well, actually there are people creating GNUstep apps out there I thought it might be another target for Lazarus, a relatively easy one in case it is really so similar to Cocoa as claimed...

[Phil]

Well, actually there are people creating GNUstep apps out there I thought it might be another target for Lazarus, a relatively easy one in case it is really so similar to Cocoa as claimed...

When you make an LCL app, you're programming against the LCL, not against the underlying widgetset. So having a GNUstep widgetset would not provide any benefit (and probably numerous drawbacks) over, say, using the Win32 widgetset on Windows or GTK2 widgetset on Linux.

The other case would be creating pure GNUstep apps and not using the LCL at all. But then you would need compiler, form designer, IDE, etc. On Mac, we can use FPC and Xcode and thereby utilize the full feature set of Xcode to create Cocoa apps. Something similar would be needed for other platforms.

A more likely scenario (although still in the realm of science fiction) would be if Apple releases an Xcode for Windows. If you have Windows, install Safari or other Apple app, then look at Apple's files under Program Files\Common Files and you'll see many DLL's with familiar OS X framework names (Foundation, CoreFoundation, etc.). Although no Cocoa, which suggests that they used Win API for the UI, not Cocoa. The point is that Apple already has most of the common frameworks ported to Windows. The question would be what strategic advantage an Xcode on Windows would provide as the world shifts to iOS and other mobile OS's and away from desktop OS's.

Thanks.

-Phil

[TurboRascal]

I wasn't actually talking about using Cocoa itself (especially in a possible cross-platform way) at all, just about how easy would it be to "abuse" the existing LCL to create a GNUstep version

One reason to use GNUstep would be that it itself is crossplatform, so you get a similar look and feel on different platforms. Of course, that way it would be possible to also create apps that blend in with other GNUstep apps when needed.

And of course the programming in Lazarus is done against the LCL, that exactly would be an advantage - a developer who knows Lazarus/Delphi could participate in a GNUstep based project without knowing much about the GNUstep itself. Also those apps could be, in a usual Lazarus/LCL fashion, made to work with another widgetsets.

However, it seems the interest in GNUstep is too low for any of this to happen

[Phil]

I wasn't actually talking about using Cocoa itself (especially in a possible cross-platform way) at all, just about how easy would it be to "abuse" the existing LCL to create a GNUstep version

Well let's think about that. The Cocoa widgetset is at alpha stage and the Carbon widgetset is still buggy and unstable. So how "easy" would it be to build a widgetset on top of an incomplete framework that has no compiler support?

Thanks.

-Phil

[TurboRascal]

Well let's think about that. The Cocoa widgetset is at alpha stage and the Carbon widgetset is still buggy and unstable. So how "easy" would it be to build a widgetset on top of an incomplete framework that has no compiler support?

That's a good point too.

[felipemdc]

Well let's think about that. The Cocoa widgetset is at alpha stage and the Carbon widgetset is still buggy and unstable.

I don't think that the Carbon widgetset is unstable. I use daily Lazarus compiled in Carbon and it works fine. I also write very often LCL apps and test in Carbon and they work fine.

As for GNUStep, as marco said there isn't much use for it. GNUStep is useful for people that write apps in Objective C and want to port them to other platforms. For LCL apps you don't need that, you can use LCL-Qt, LCL-Gtk2, LCL-Win32, LCL-Wince. No need to redirect through GNUStep in other platforms.