“The GNUstep Windows installer is based on the MinGW system and consists of the basic MSYS and MinGW libraries, other library dependancies and the GNUstep Core packages (gnustep-make, gnustep-base, gnustep-gui, and gnustep-back). The installer installs GNUstep onto most varieties of Windows (see below for tested installations) and sets up the computer to make it easy to run GNUstep applications. It is based on the NSIS installer.”
Honestly, who’s going to use it on Windows of all platforms? GNUStep doesn’t have killer apps to entice Windows users, and apps appear even more alien in Windows than any Gtk+ app ever did. Even on its native Linux environment as a developer toolkit it’s far overshadowed in mindshare by Gtk+/Gnome and Qt/KDE. As a cross platform compatibility layer for OS X Cocoa apps it lags too far behind the Apple libraries and does not seem to be catching up.
The project really seems to be headed for a dead end (or a vanishing niche) to me, regardless of coolness or technical merits. This may sound like flamebait, but I think it’s just a realistic assessment of the market. (Yes, even FOSS projects have markets. Even if they don’t get money, they still need active users and developers to stay alive.)
On the contrary, I think as the Mac is getting more and more popular, and more developers make Mac-only apps (there’s already a ton), some may wish to port these to other platforms, without a complete rewrite.
GNUStep may be the solution to their needs (depending on the coverage of the Mac OS X APIs).
Edited 2007-12-25 01:49
Also, some people are forced to use Windows on their workplace. Until KDE apps can run on Windows (I prefer GNOME, but… whatever), this could be a great addition.
Until KDE apps can run on Windows (I prefer GNOME, but… whatever), this could be a great addition.
GTK2 has long been ported to Windows and the Gimp, Abiword, Gnumeric and Pidgin all run on Windows.
AbiWord uses GTK on Windows? The last time I checked AbiWord had a custom GUI for each platform.
GTK2 has long been ported to Windows and the Gimp, Abiword, Gnumeric and Pidgin all run on Windows.
None of those applications are part of the GNOME Desktop Environment.
None of those applications are part of the GNOME Desktop Environment
Abiword and Gnumeric are the core apps of GNOME Office. I think they automatically count as GNOME apps and the GIMP is included in the list of GNOME projects.
The original poster said:
Until KDE apps can run on Windows (I prefer GNOME, but… whatever), this could be a great addition.
I think he was referring to apps not the complete DE.
Nobody on Windows can use it if there’s no Windows installer. At least now the possibility to use it is there.
The read advantage for everyone is that developers can develop in the comfort of a Unix environment like Mac or Linux, then spend a little bit of time tweaking the program and create a windows version instead of the other way around. Your Mac apps look and behaves native, and there is really no such thing as a Windows native feel (just look at the differences even amongst all the MS apps).
The sad fact is that a large number of people are forced to use windows for one reason or another, and with GNUStep, you can still use Objective-C and Windows users can use your programs.
Some developers like choices even if there isn’t the time to try everything of interest right now. These GNUstep ObjC articles stick in the back of my mind so if I do get a Mac system this will give me one more cross platform avenue back to Windows.
but I wish
1. They helped MidnightBSD with gnustep+etoile ports
2. Improve jigs + Io bindings
3. Improve solaris port.
instead of porting to windows.
But I really understand them since people can start programming on windows until they jump to FOSS Operating Systems.
Something I would like to see born from the gnustep project is a standalone foundation framework (a self-contained gnustep-base library) and Win32 or Avalon or whatever Windows GUI tech wrapper in Objective-C.
I really like the language, but in Windows it’s use is not practical.
I wonder if they’ll collaborate with GNUstep.
Hopefully this’ll enable more developers to work on this project and who knows we may even see a native web browser created for this desktop
i’ve been watching the Gnustep project for ages and despite the old look of the desktop (which looked like a killer app back to the Windowsmaker and afterstep ages), I’m still wondering why they didn’t try to add some kind of gnome/kde backend to their api in order to make gnustep app compatible with the two main linux desktop and more importantlty to incitate Mac os X 3rd parties to port their app to gnustep/(gnome|kde).
It would be far more usefull this way… it would attract more people to gnustep/cocoa….
no one is ever going to developp something for a few hundreds users (gnustep user base), but an app for os x/ linux using gnustep api and well integrated on their day to day desktop ? how cool….
writing this, I realize this is exactly what wxwindows is for..
so forget the whole point, gnustep is brain dead..
If you had ever taken the time to read the mailing lists, you would understand why this isn’t possible.
Cocoa (and GNUstep, by definition) uses a scalable vector based drawing system (postscript) to draw it’s controls. The look can be altered by subclassing the existing classes and overriding the drawRect: method.
In order to use the native widgets we would need to implement something similar to how SWT does it and, as with SWT, it would not be possible to subclass these classes, since they would depend on widgets which are drawn outside of the control of the gnustep-gui library. This is unacceptable.
So, before calling something brain-dead for reasons you don’t really understand, I would suggest you do your research on why things are they way they are.
Thanks, GJC
— GNUstep Chief Maintainer
Despite the fact that I knew, as everybody else who keeps reading news about system used by 123 persons, that gnustep use a new flavor of display postscript, your argument is irrelevant. Just take it as the opposite way, if it’s not possible to use native widget for gnustep, maybe you should try to have a version of gtk using display postcript as graphic backend, and then your argument will vanish like your potential user base when you insult potential user in a public forum.
Anyway, I can’t believe I’m arguing with someone using gnustep and a postcript backend running on top of X server, posting with gtkished firefox.
by the way, I’m using the real OpenStep legacy desktop, it’s called mac os x and at least all cocoa apps look consistant. Call me back when you’ll have something decent and native to use with gnustep.
Djamé
ps : for the others niche os users, I know it’s rude but this sob pissed me off.
Djarne,
Your “solution” wouldn’t fix anything, since it misses the point entirely
GNUstep does not use a DPS server, it uses a DPS/PS like API on top of a backend which is used to do the low-level drawing such as line drawing and font rendering. By low-level I am *not* referring to buttons and text fields.
There are a few different backends:
* Windows – Uses GDI to draw in windows
* Xlib – Uses X functions to draw in windows
* Cairo – Uses the more advanced Cairo library to draw.
Each of them use the line and font drawing primitives to create the widgets.
Gregory Casamento
— GNUstep Chief Maintainer
P.S. A word of advice. The best way to start a civil discussion is not to start it off by calling something that you’ve publicly demonstrated you don’t understand as “brain-dead”. I believe that most “niche os users” would agree with me.
And so what ? what have you demonstrated so far except the fact that of you course you know that something is not doable…
I’ve always thought that the whole point of the whole OpenStep architecture is to provide rich API and brilliant app.
If no one is willing to use your plateform because it doesn’t integrate well with any of the two major unix desktop, telling that the kind of integration people are excepting from the so called most powerfull plateform ever won’t help. More importantly it shows that you’re just sitting on top of your ivoiry tower. You know what ? stay there because You didn’t answer the main question : Why should people use GnuStep if it doesn’t bring any compelling advantage nor any killer apps ?
Djamé
Ps : by the way, don’t P.S on me, you start flamming over because I was refering to your plateform as “brain dead” because there’s just no native app I can use on it. You’re just as relevant as my old Falcon and as I was telling, “brain dead”
Djarne,
There are no ivory towers here and I’m certainly the last to be in one. When I took over as chief of the project a year ago the problem of getting more people to use GNUstep was the main issue I was concerned about.
GNUstep is taking steps towards become better integrated with other UNIX desktops through the use of a theme engine called Camaelon and we are working on another one which is more flexible and extensible as well.
What is GNUstep’s compelling advantage? It’s development environment and it’s API are.
Currently, GNUstep’s Development environment is as good or better than those on GNOME and KDE and has been for some time. Gorm has the ability to be dynamically extended using palletes… glade currently lacks this feature entirely. Many of it’s features are enabled by the language Objective-C, the language GNUstep is based on, which is much more dynamic than C++.
Does GNUstep currently have any killer apps? No, unfortunately, it doesn’t. This is because we currently lack developers. One of the reasons for this is the way GNUstep looks. And, as I said, it is currently GNUstep’s top priority to change that and make it work better not just with environments like KDE and GNOME, but also with Windows.
At any rate, I hope this clears up your confusion. Please feel free to email the list at [email protected] if you have any questions regarding GNUstep. Archives of the list are available at gmane.org:
http://dir.gmane.org/gmane.comp.lib.gnustep.general
Later, GJC
— GNUstep Chief Maintainer