On the Etoile news blog, Quentin has posted an installation guide for the Etoile desktop environment. “Here are some quick instructions to build and install EtoileFirst you need to build and install latest GNUstep svn version (not the latest release), then you can build and install Etoile svn version too.”
I’d love if GNUstep and Etoile accelerated in development. This would be a good step towards working on a platform gearing developers to use the GNUstep API. I like GNUstep as a project, but I don’t like the plain and bland look. Quite simply, I want GNUstep to look like OSX.
Meh. A skinnable system that could look like either OSX or NEXTSTEP would be ideal to me. I happen to find the NeXT UI one of the most elegant ever created, but I also feel its wrong to force everyone wanting to use the APIs to agree with me.
Quite simply, I want GNUstep to NOT look like OS X.
But it is themeable.
I’d like a skin that looks like MacOS 8/9
I feel this is a good direction for their project to go in. To those that are unsure of the difference between Etoile and GNUstep, it is that Etoile is a desktop environment and GNUstep is just an API.
This distinction is important, as I noticed some comments in reply to the original article did not understand this.
Let each focus on what they do best and progress will be made as long as things do not stagnate.
This is an installation guide? Omg – an next time we will learn how to press the power button.
Why was parent modded into oblivion? This guide is really basic compilation how-to with simple order and two packaged cvs snapshots…
BTW, it might be my incompetence or using Ubuntu Feisty but I had to mess with exporting PATH, LD_LIBRARY to get it going.
as long as I can customize it to my liking, theme it, and configure it any way I want.
I definitely will not accept an OSX clone because OSX is not usable GUI because it’s not customizable.
Same goes for Gnome btw.
Out of the desktop GUI’s only KDE meets my requirements (and just barely).
This one looks quite interesting and seems to have a lot more momentum than Afterstep. Wonder why I never heard of it until now…
Perhaps because the information is in the FOSDEM 2007 presentation of Etoile?
http://wiki.gnustep.org/index.php/Main_Page
http://wiki.gnustep.org/index.php/FOSDEM_2007
http://www.etoile-project.org/etoile/mediawiki/index.php?title=Main…
Trunk SVN:
svn co svn://svn.gna.org/svn/etoile/trunk
I’m personally going to see if it builds well on Debian Sid amd64.
I really hope that this takes off, because this could be the best free desktop of them all.
I was posting a couple of months ago on the GNUstep list that I would like to see the GNUstep developers change focus to work on a desktop, rather than just settle for catering to a lot of desktop environments so they can run GNUstep applications.
I still stand by my point, because I think it’s a waste of time, but this project seems to be going in the right direction, helping GNUstep to establish itself as more than just an anonymous runtime environment like Yellow Box.
This project could be the primary platform to a free, MacOSX compatible desktop.
“I really hope that this takes off, because this could be the best free desktop of them all.”
Maybe it can be the best free desktop, but it won’t be the most used free desktop; this “fight” will be between Gnome and KDE.
I really like the basic concepts of GNUstep, the API, and Objective-C. Allthough I prefer Gtk + C (over Qt + C++), I think I can like ObjC. I’m really thinking about using GNUstep as a development platform. By the way, I do not develop mainstream software for Joe Q. Average, but scientific and administrative software that is usually used by well educated individuals. They don’t look for flashy colourful screens with dancing puppies and squeaking buttons, they want a professional solution that helps them solving complex tasks. In my opinion, GNUstep can serve well here.
At home, I’m currently using WindowMaker (on a FreeBSD system), a very capable window manager based on GNUstep. Especially its speed, the excellent keyboard support (i. e. for Sun USB keyboard) and the “stay out of my way and let me work” concept are appealing to me. But maybe I’m a bit differnt… 🙂
“I was posting a couple of months ago on the GNUstep list that I would like to see the GNUstep developers change focus to work on a desktop, rather than just settle for catering to a lot of desktop environments so they can run GNUstep applications.”
As you might have read from the comments, there are opinions going into the other direction: GNUstep should concentrate on supporting the two major desktop environments KDE and Gnome, because they are standard in Linux world. But I agree with you, GNUstep developers should concentrate on developing and evolving the GNUstep system itself, adding modular default applications and, finally, making all the users who are hungy for eye candy happy. (By the way, I like GNUsteps clean default look, it’s not that bad.) Oh, did I mention? Speed, low hardware requirements, low library consumption.
“I still stand by my point, because I think it’s a waste of time, but this project seems to be going in the right direction, helping GNUstep to establish itself as more than just an anonymous runtime environment like Yellow Box.”
I complete agree. I’d like to see some native GNUstep based applications, so WindowMaker (as a “result” of GNUstep) is no longer “just” a window manager. It would be great to have a default file manager, default media player, default this and default that (you know the stuff I would mention), but always modular, please. 🙂
“This project could be the primary platform to a free, MacOSX compatible desktop.”
Why is there the need to have a “MacOS X replacement”? I really don’t understand it. Allthough GNUstep would be a great platform, if I could write a program for GNUstep that could be crosscompiled to run on MacOS X, I do not neccessary think there’s a need to have it vice versa, or, to have a sort of “MacOS for Linux”. Actually both KDE and Gnome can be given the look and feel of MacOS X, so why to have another one? Feel free to reply and explain.
At home, I’m currently using WindowMaker (on a FreeBSD system), a very capable window manager based on GNUstep.
It’s actually only emulating the look of NeXTstep, but it’s written in plain C. It is in no way based on GNUstep and in fact conflicts with certain aspects of GNUstep.
Why is there the need to have a “MacOS X replacement”? I really don’t understand it.
Not a replacement, but a supplement to run on standard PC’s. The part about being free, means something to certain people and this can be an important aspect, if they don’t want to use OSX.
Actually both KDE and Gnome can be given the look and feel of MacOS X, so why to have another one?
No, they really can’t. KDE and Gnome suffer from many design flaws that GNUstep and a suitable desktop like Étoilé has a chance to fix.
In the early days, NeXTstep proved to have great potential as a very serious desktop, easily being 10 years ahead of the competition at the time, but the expensive hardware and software limited the widespread use of NeXTstep.
Then GNUstep came along and opened up to a possible free version. GNUstep’s far superior underpinnings and the fact that it’s modeled after an existing and well described framework that is known to work very well, makes it a great platform for writing applications and a good desktop environment with. A desktop that can leap way ahead of KDE and Gnome.
“[WindowMaker is] actually only emulating the look of NeXTstep, but it’s written in plain C. It is in no way based on GNUstep and in fact conflicts with certain aspects of GNUstep.”
As I understood, it is supposed to be the official windows manager of GNUstep… at least they share aspects according to the user interface.
“Not a replacement, but a supplement to run on standard PC’s. The part about being free, means something to certain people and this can be an important aspect, if they don’t want to use OSX.”
Yes, I see. But it’s a certain way to reach this goal.
“KDE and Gnome suffer from many design flaws that GNUstep and a suitable desktop like Étoilé has a chance to fix.”
I just assumed this to be, I do not use KDE nor Gnome on a daily basis, but as far as I know KDE from live system CDs, much can be done to make it look like MacOS X in some aspects, but it surely is not a complete replacement for the UI elements (such as menu bar on top of screen).
“Then GNUstep came along and opened up to a possible free version. GNUstep’s far superior underpinnings and the fact that it’s modeled after an existing and well described framework that is known to work very well, makes it a great platform for writing applications and a good desktop environment with.”
I’m aware of this potential, so I’m really considering using it, as I pointed out before.
“A desktop that can leap way ahead of KDE and Gnome.”
At least if it offers nearly the same functionality (tools for common operations and system administration, networking support etc., just have a look at the comments in the article linked), and because of GNUstep as a viable base, I think it can offer these functionalities with less hardware requirements than Gnome or KDE.
I just assumed this to be, I do not use KDE nor Gnome on a daily basis, but as far as I know KDE from live system CDs, much can be done to make it look like MacOS X in some aspects, but it surely is not a complete replacement for the UI elements (such as menu bar on top of screen).
Mimicking a look is relatively easy, implementing all the functionality and getting the same feel is something very different.
You just have to compare the Dock clones available for Linux with the real thing on Mac OS X to see what I mean. At a glance they can look almost identical, but start using them and the major differences become apparent as you hit inconsistencies and important missing features. Then there are the applications, you’re missing out on 90% of the Mac OS X UI when not using applications that are designed for it. Overall the copy of Mac OS X possible on KDE/GNOME is almost entirely skin deep.
Of course the purpose of GNUstep isn’t to provide a copy of the Mac OS X GUI on Linux. What it can provide is a desktop that’s more customisable, which can look and feel like OS X if that’s what you want. Whatever it looks like, it has the potential to provide a highly consistent desktop, with nice NeXTSTEP derived features that other Linux desktop environments lack.
Unfortunately it only really provides those benefits when running a set of GNUstep applications, and at the moment they’re rather thin on the ground. With luck GNUstep will increase in popularity and the software situation will change over the next few years.
Mimicking a look is relatively easy, implementing all the functionality and getting the same feel is something very different.
I’d like to point out that it’s important to understand that GNUstep and Étoilé really is about the whole thing: All layers of the system above kernel level up to the user interface. The kernel itself and the graphics system, AFAIK, is intended to be independent of GNUstep.
A desktop wouldn’t be GNUstep based, if it was a mere skin on a desktop, or even the look and feel. Emulating it is not enough. Building Étoilé on GTK+ is not enough.
It’s about how GNUstep is designed from an existing well-tested standard called OPENSTEP, a template if you will and offers a single, well-designed API to create programs with in Objective C. Objective C isn’t used, because the GNUstep maintainers have an Objective C fetish, but because the standard says so. The standard doesn’t describe how the GUI should look, which is why MacOSX looks different from the original NeXTstep.
The design gives you a ton of leverage emphasizing importance on things being complete on a fairly low level. Very complete CPU independent trivial functions are extended from Objective C. When you want to handle strings of text in any way, use the NSString object or one of its many subclasses. Objective C in itself is a very thin layer above plain C.
You have easy to use arrays, memory handling, even network communication, powerful distributed objects and such things that you could do in plain C if you wanted, only the API takes care of all the boring stuff, like making your code CPU compatible with both PowerPC and Intel and making your data interchangable between applications across clipboards, drag’n’drop, as inputs to services, etc.
On a bit higher level, MacOSX for example offers an advanced set of functions called Core Data, to handle loading and saving of data within your application as well as handling your data like a regular database design. This way, you don’t have to worry about data persistence or building trivial data handling functions, like adding, updating, removing entries in trivial lists and such things. I believe this is being implemented in GNUstep now as well: http://gscoredata.nongnu.org/.
Less footwork for the same results.
Doing such complete work on this level, gives potentially far more leverage, than implementing a fancy IDE on top of a poor foundation. I guess this is why GNUstep already has a fairly complete IDE (and there’s even another one coming), despite not having many apps yet.
Also, what can be seen as an advantage, you have not many choices in GNUstep as to how to create applications. There are bindings to various languages, such as Ruby and Python, but they are still bound to the original API. It’s not like Windows, where you may have a plethora of different APIs (MFC and .net, for example) of varying quality, where they started building a new one as soon as the old one didn’t hold up anymore). It can be hard to select a good one.
In GNUstep, there is only one real choice, there always has been, and you can get going from day one. There aren’t any arguments or wars about, which way to do apps is best in GNUstep. No (or at least very few) ambiguities and a very high degree of reusability.
This is also why most MacOSX developers have quickly migrated to Cocoa and XCode to take full advantage of the system. An unfortunate side effect is that ported applications can’t be as good as a real “native” one, unless you do significant redesign for the platform or rewrite it from the bottom up.
etoile was integrated with NetBSD.
Most of these projects pretty much develop on Linux. This is the same case with Gnome and KDE. Most of it works but serveral componenets require linux features. Like the thread mentioned, it seems many devs are using Ubuntu. I hope they make it a requirement that there is feature parity and portability to the xBSDs.
Many devs are using debian or ubuntu, that’s true. But at least me and david are using FreeBSD.. 😉
How good is GnuStep as a tool kit in rendering complex glyphs such as Unicode for CJK and Indian languages? It is very much a prereq for those of us outside the Latin script languages. Its the main reason why many devlopers here stick to Qt and Gtk – because of the bi-directional and multi-lingual rendering framework.