Gnustep features a rearranged printing kit, LessTif now supports wheel mice and it builds as Motif 2.1 clone, while Fox had a recent development release too (1.3.12). Some of Fox’s flagship applications include Xfe, ReZound & Goggles which look great in AA (screenshot). Elsewhere, gtk2-perl released a bug fix package of the GladeXML module while the Gnome developers are now discussing the possibility of adding Python (pygtk, part of Gnome Bindings) to the core Gnome release.
GNUstep 0.9.4, Lesstif 0.93.96 and Fox Toolkit Releases
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Linux users will never take GNUstep seriously, compared to GTK and QT.
It’s just too bad that all efforts on creating an GNUstep-based OS have failed.
But then again, there are no OSs based on other environments like KDE or Gnome either.
It would be really nice to see a system with only GNUstep-based applications. And I believe that it would soon be possible, it seems like more and more are discovering how simple and fun it can be to develop for the old NS-plattform
Tips regarding my Fox applications:
1. The Fox configure script still has a defiecieny and in some linux distros it won’t locate the right freetype2 headers and so it would disable AA even when this is passed as –with-xft in the ./configure script. If this happens, you need to temporarily do this:
cd /usr/include
mv freetype freetype-ORIG
cd to your Fox 1.2.x folder you just untarred
./configure –with-xft –with-GL (I think that’s the argument for GL, not sure)
make; make install-strip (the strip installation has two bugs, do a ‘make install’ if you can’t fix it by hand)
mv /usr/include/freetype-ORIG /usr/include/freetype
2. Edit your ~/.foxrc/Desktop file to change the colors to match your Gnome’s or KDE’s colors. That’s how I have them in the screenshot to match the “Gnome Default/Raleigh” theme. My Desktop file:
$ cat /home/eugenia/.foxrc/Desktop
[SETTINGS]
normalfont=”Sans,90″
backcolor=Gray100
#basecolor=#D4D0C8
basecolor=#dcdad5
bordercolor=Black
forecolor=Black
hilitecolor=Gray100
shadowcolor=#808080
selforecolor=Gray100
#selbackcolor=#0A246A
selbackcolor=#4b6983
tipbackcolor=#ffffe1
tipforecolor=Black
3. Make sure you read the README.Opengl which contains a patch for a libtool bug. Then, during compilation of ReZound, you might get a gcc error about TextFont-something. Open that file with a text editor and swap the two TextFont lines that are next to each other inside an #endif statement.
4. Xfe and Goggles compile with the Fox-dev version of 1.3.12 just fine, but not without hacking around their makefile/configure scripts to tell them where the headers live, so better stay with 1.2.9.
5. I found Xfe to be a good Nautilus/Konqueror replacement, it is build around the MS Explorer idea and it works well. Its only drawbacks are a MIME type-mixup bug which won’t allow you to load a known text/plain mime type with a text editor.
6. ReZound works only with OSS atm, alsa support is in CVS. It could do with treaming its very long menus btw.
>so basically.. it’s easier to code in and does more, yet it’s worse because of >steve jobs?
>god i love moronic trolls.
People who have never actually tried Obj-C often look at it and see some small syntactical differences and then discard it as some weird smalltalkish
language not worth learning.
It’s really quite sad, since it’s partly those small differences that makes Obj-C so interesting. But most people are and will continue to be quick to dismiss things that are a bit different compared to what they are used to.
There is this Live CD which you can install and have a system (with almost only) GNUstep and GNUstep applications. http://www.linuks.mine.nu/gnustep/ Give it a try. I hope to have a bunch more apps and the latest versions by next month. Have fun, developping graphical software is only nice with GNUstep.
I’m happy to see GNOME adopting pyGTK as the alternative defacto GUI language. I disagree with the individual who says it should remain a scripting language and nothing more. That’s totally blatant! It should be the prefered RAD language on the platform for several reasons:
1). It is politically neutral.
2). It is free, stable and available today.
3). It is upsettingly easy to learn and program in.
4). It is a RAD tool.
5). It supports multi paradigm programming techniques.
6). No Microsoft, no SUN, no patents to worry about. (Whine all you like folks)
7). It interacts effortlessly well with C and C++
8). There are tons of GTK/GNOME applications already written in it at all levels, not just scripts.
9). It is more productive and as popular as the alternatives
10). It’s GPL–nuff said.
I hope these points keep being hammered on the thread.
> It’s really quite sad, since it’s partly those small
> differences that makes Obj-C so interesting. But most
> people are and will continue to be quick to dismiss
> things that are a bit different compared to what they
> are used to.
I’m obviously biased on this topic, as I’m the GNUstep maintainer for Gentoo, but I have to say, what you are saying is quite true.
As far as GNUstep goes, it’s getting there. The -make and -base packages are post-1.0, and -gui is pre-1.0, so yes, the -gui aspect is a little sketchy at times. I would recommend only certain aspects of GNUstep in critical situations, at this point, but heck, some parts of it, as I said, are pre-1.0, so that should be a clear sign to you about what you can expect.
However, GNUstep, Cocoa, and Obj-C in general, are a pleasure to code in. If you’re just starting out, give Obj-C it’s fair chance: don’t even use GNUstep or Cocoa; write up some test apps for yourself. Play with the language. If you actually grok OOP, you’ll soon see why Obj-C makes C++ look so haxor-ific. Sadly, really understaning OO is almsot impossible when most people use C++ as “lazy man’s C”, as C++ simply doesn’t give you any reason to stick to the OO paradigm.
I may be jumping the gun here, and assuming things about other posters in this thread, but for any C++ over Obj-C fans, can you really off the cuff explain the differences between reinterpret_cast<>() and dynamic_cast<>()? When was the last time you wrote a templated class? Was it fun? Or painful? Finally, the concept of __sending messsages__ v. __calling a method__ cannot be overstated as AMAZINGLY POWERFUL. How can a language be usefully introspective when everything has to be computed at compile time?
Anyways, just tossing some ideas out. I will say that Qt is possiblly the best architectured C++ framework I’ve had experience with; and GTK+ is noble in that it does seem to stick to an OO paradigm, even though it’s written in C. GTK+ v.1 series had Obj-C bindings; I don’t know if GTK+2 series has them (yet).
And just for the record, I’m not a fan of Steve Jobs, Apple, or any corporation, for that matter. 😉
GNUstep seems to be a big wasted opportunity. The developers need to drop their love for the 1988 NeXT GUI, and get modern. Put the menubar where pretty much every other Linux/*BSD/whatever app puts it and create a non-ugly theme. No toolkit is an island.
> GNUstep seems to be a big wasted opportunity.
By what standards?
> The developers need to drop their love for the 1988 NeXT
> GUI, and get modern.
Modernity of programming environment sort of is one of the side goals of GNUstep. Plus, you’re definitely referencing the look of the GUI, which, is of course, customizable.
> Put the menubar where pretty much
> every other Linux/*BSD/whatever app puts it
That will simply never happen. Arguments about SDI v. MDI v. other GUI paradigms (gawh, i hate that word, and i keep using it!) aside, SDI interfaces offer a certain kind of consistent feel of use. I, personally, love it. If you don’t mind SDI, but hate that the default GNUstep menu is vertically stacked, you can get WildMenus, which makes it horizontally aligned, at the top of the screen.
Plus, your assumption that GNUstep is another UNIX like OS, is a non accurate observation. GNUstep can be run on top of Windows, OS X, Linux, *BSD, on top of X11, native windowing, etc. However, GNUstep is meant to recreate the OpenStep4.2 spec, and extend it to maintain decent compatibility with Cocoa/OS X. It’s not _supposed_ to work like a GTK+ app, so I’m not going to say “you’re wrong”, but your statement is flawed.
> and create
> a non-ugly theme. No toolkit is an island.
The package is called Camaelon. It lets you load themes.
a theme bundle can also change the menu to mac-like horizontal style.
PyGTK seems less fun to work with than Ruby-GNOME2. Too bad that they can’t help Ruby-GNOME2 become official as well. 🙂
Just to chime in, GNUstep has RIGS, which is a ruby bridge for Obj-C/GNUstep.
Ruby is possibly my favorite interpretted language.
The Ruby/Cocoa people and the RIGS people need to get together and come up with one package, imho.
>> GNUstep seems to be a big wasted opportunity.
>By what standards?
By the fact that the total number of users can’t be any more than 100.
By the fact that there are hardly any applications, and even less that are worth using.
By the fact that it doesn’t integrate properly with _any_ window manager, not even WindowMaker, which is supposed to be its official one.
By the fact it has (almost) source compatibility with a good chunk of one of the most popular operating systems for desktop in the world. Yet, the number of ported applications is like, 3?
The case for pyGTK against all other bindings is not that it’s python is a better language. It’s because it has mature bindings (ruby’s are not stable yet according to its devs) and more importantly: it has many-many third party apps already. Just make a quick round at gnomefiles to see that pygtk is the most used binding toolkit, second only to GTK+ itself.
No, I am not using python, but reality is, it’s very popular among gnome third party developers, and so in my opinion pygtk (and gtkmm IMHO) should be part of the Gnome platform releases. But the actual core of gnome-libs should stay C.
I generally agree. The performance edge that Python has over other languages is important as well. But GUI code with a nice OO language works so well, that Ruby should be considered also (put it on your radar at least, Gnome devs
. Besides, the Desktop development should become as easy as possible, because the web programming and the competitors are targeting at a very high level programming which is making the current Desktop programming kind of deprecated. I don’t like the OO of Python. About the maturity of Ruby-GNOME2, at least check out its history:
http://ruby-gnome2.sourceforge.jp/hiki.cgi?History
Pretty old (98) and stable project. 🙂
> By the fact that the total number of users can’t be any more than 100.
What kind of fact is that?
> By the fact that there are hardly any applications, and even less that are worth using.
I prefer GNUMail over many other clients, it does thread arc, very usable
and beautiful.
Agenda has the best presentation of events among other clients, imho.
Gorm is the best tool to build UI and with Obj-C’s power, it makes
other tool on other toolkit looks like a joke to me. (sorry)
I also love ILogin, it is the most interesting login panel I know of and
there are other usable apps.
> By the fact that it doesn’t integrate properly with _any_ window manager, not even WindowMaker, which is supposed to be its official one.
Lack of developers to solve some small problems.
> By the fact it has (almost) source compatibility with a good chunk of one of the most popular operating systems for desktop in the world. Yet, the number of ported applications is like, 3?
3? You must be kidding!
> By the fact that the total number of users can’t be any more than 100.
There is no way to prove or disprove this statement, however, I can say with some confident zeal: there are more than 100 users.
> By the fact that there are hardly any applications, and even less that are worth using.
So, considering gnustep-gui is still pre-1.0, and the fact that all my core app needs are covered, except for an IM client, I have to disagree with this heavily.
> By the fact that it doesn’t integrate properly with _any_ window manager, not even WindowMaker, which is supposed to be its official one.
0) GNUstep doesn’t even need a window manager to run
1) GNUstep apps can be run in any window manager (or on Windows, for that matter)
2) WindowMaker + GNUstep apps offer a very NeXT-like experience … not sure what you mean it doesn’t integrate with it, ’cause it seems to just fine over here
> By the fact it has (almost) source compatibility with a good chunk of one of the most popular operating systems for desktop in the world. Yet, the number of ported applications is like, 3?
Well, name some open source OS X apps, and we’ll get to it. “Major” apps, like Camino the browser (or OS X Firefox), AbiWord, etc is pretty much just waiting on Obj-C++ support in gcc (coming in the next major version 3.5/4.0). This is unfortunate that it’s not _now_, but firefox/GTK+ is working just fine for the meanwhile. 🙂
I suppose, overall, I just don’t get the arguments:
– GNUstep isn’t as usuable as GNOME or KDE _right now_!
– GNUstep applications are themeable, but its default theme I personally find offensive!
– GNUstep is mostly source compatible with Cocoa, so why aren’t companies/dev’s creating GNUstep apps!
… ’cause _it’s getting VERY usuable as a DE (note that GNUstep is more than just it’s use as a DE). Themes need to be written, but they _can_ be written, now. More and more first-on-Cocoa apps are getting ported to GNUstep, and almost all GNUstep apps that aren’t base-OS specific, (e.g. for GNUstep on Linux) are already ported (e.g. GNUMail.app).
>Lack of developers to solve some small problems.
Ah, yes, but _why_ is there a lack of developers? Because GNUstep doesn’t know if it is a desktop or a development framework. And as a result it does both shoddily.
>I prefer GNUMail over many other clients, it does thread arc, very usable
>and beautiful.
>Agenda has the best presentation of events among other clients, imho.
>Gorm is the best tool to build UI and with Obj-C’s power, it makes >
>other tool on other toolkit looks like a joke to me. (sorry)
>I also love ILogin, it is the most interesting login panel I know of and
>there are other usable apps.
That’s 4 apps. Agenda is alpha at best. iLogin hasn’t even had a release. I’ll give you Gorm and GNUmail – just.
>3? You must be kidding!
It isn’t many more than that. I can think of TAMsAnalyser, MPlayer and, well – that’s it. I can’t think of any commercial applications whatsoever. GNUmail and Cenon don’t really count, as I believe they went from GS -> Mac OS X, not the other way around.
> >>By the fact that there are hardly any applications, and even less that are >>worth using.
>So, considering gnustep-gui is still pre-1.0, and the fact that all my core >app needs are covered, except for an IM client, I have to disagree with this >heavily.
So you don’t need…
A good, featureful word processor?
A spreadsheet?
A persentation program?
A web browser?
An image manipulation tool?
A database frontend?
A web design tool?
A video editor?
A file-sharing app?
etc…
Well, good for you. However, I regret to inform you that most computer users do, and those are the reasons why they use their computers.
>1) GNUstep apps can be run in any window manager (or on Windows, for >that matter)
>2) WindowMaker + GNUstep apps offer a very NeXT-like experience … not >sure what you mean it doesn’t integrate with it, ’cause it seems to just fine >over here
GNUstep on Windows? False. The backend _still_ isn’t finished.
WindowMaker doesn’t integrate well at all. It’s written in a completely different framework, WiNGS. Fonts are different (at least on the GS live CD), widgets look different.
I hated the interface. It is way beyond non-intuitive.
So, you’ve already mentioned the “GNUstep doesn’t know if it’s a DE or DK. Well, it’s a development kit. The most appropriate thing one could say about a GS based DE is that “This DE is built with the GNUstep libraries.” An odd categorization, and definitely one that isn’t easy to grasp for new users, but like the “what is linux?” naming argument, it does not matter for users, but in this context discussing os news, it surely does matter.
> So you don’t need…
> A good, featureful word processor?
Ha, actually “no”, but I’ll say OpenOffice.org for this one.
> A spreadsheet?
ditto
> A persentation program?
ditto
> A web browser?
Camino or OSX Firefox, ported.
> An image manipulation tool?
There’s PRICE, but if you’re thinking GIMP style, well, the GIMP.
> A database frontend?
Gah, freaking phpmyadmin and phppgadmin work just fine for my needs.
> A web design tool?
ViM works well for me, even when I want to write serious web page (which isn’t often).
> A video editor?
I have little need for this, but it is a fun/useful thing to have; this is a good idea for an app for me to write though, thanks 😉
> A file-sharing app?
Also, a BT client would be fantastic … but screen and btdownloadcurses.py serve my needs for now. Ooh, this is a better idea for an app to write. 🙂
> Well, good for you. However, I regret to inform you
> that most computer users do, and those are the reasons
> why they use their computers.
… see here’s that argument again … “GNUstep sucks ’cause it doensn’t have XYZ type app!” Yeah: _yet_. I’m not saying “Use GNUstep or you suck!”, cause that would be silly and pointless. I’m saying GNUstep is a great framework to write apps in, give it’s pre-1.0 nature. Sure, it needs more apps. I can say that about every single OS out there. _NO_ OS fits my needs perfectly, not even Gentoo with GNUstep, like I have here, but it’s closest to what I personally want.
I can’t say I rightly care if the app I’m using currently doesn’t fully integrate with GNUstep / written with GNUstep libraries. Apps will be soon enough. Dev’s will get to it. I’ve been using some variant of GNU/Linux since ’94-’95, and if I had the attiude you are describing, I’d of switched to Windows long ago, and laughed at Mac’s silly OS7,8,9.
> GNUstep on Windows? False. The backend _still_ isn’t finished
… Windows is the extreme example because of it’s “who’d a thunk it” feeling in regards to GNUstep.
> WindowMaker doesn’t integrate well at all. It’s written
> in a completely different framework, WiNGS. Fonts are
> different (at least on the GS live CD), widgets look
> different.
… WiNGS? You mean the “Window maker is not GNUstep” library? Yeah, no kidding it’s not written in the same framework. 😉 WindowMaker is a Next-like window manager. GNUstep is a NeXt-y framework. Together, you get a mostly NeXT-y experience (or at least, that’s the goal).
So, in the end:
– I get my personal work done via any program I can run on my computer.
– I get to work on and help out an exciting framework to program in.
… I don’t see what the problem with this is. I see what point you’re arguing, I just don’t see why it matters, I suppose.
> I hated the interface. It is way beyond non-intuitive.
… hrm. Thanks for the informative comment.
I, personally, appreciate all critique about this, ’cause it helps me clarify, in my own mind, what GNUstep means to me, and what it could mean to users.
However, “It is way beyond non-intuitive” means what? You didn’t like the SDI interface? Mini-windows frighten you, and give you nightmares?
Nice tips. But uncommenting line 7123 of ‘configure’ seems to enable xft and is arguably easier than moving stuff around.
I love compiling my own stuff (the Slackware way; not the BSD/Gentoo way) but sloppy releases definitely take some of the fun out and put headaches in its place.
(The Slack/BSD/Gentoo thing is not meant to be a flame – I usually have a Gentoo to run and currently sort of triple-boot Slack/FreeBSD/NT in that I have them on this box – just prefer a more hands-on approach.)
Wonder if nedit compiles with this LessTif? I usually install binaries just because it seems so touchy with every other release of either LessTif or OpenMotif.
>… see here’s that argument again … “GNUstep sucks ’cause it doensn’t have >XYZ type app!” Yeah: _yet_. I’m not saying “Use GNUstep or you suck!”, >cause that would be silly and pointless.
I’m not saying GNustep “sucks”. I follow the project with great interest. There is a lot of potential, and it would be shame to see it go to waste not through any technical deficiency, but through poor marketing and a lack of direction.
GNUstep, in its currrent state isn’t suitable for end-users. There also seems to be no clear agreement between those in the project whether GS is a framework and nothing more, or a framework _and_ an envioronment to use those applications within. As I said right now, it does neither well. GNUstep apps don’t integrate well into GNOME, KDE or Windows, making it undesirable to write desktop-neutral apps with. On the other hand there aren’t enough apps for GNUstep to support its own, integrated desktop.
>I can’t say I rightly care if the app I’m using currently doesn’t fully integrate >with GNUstep / written with GNUstep libraries. Apps will be soon enough. >Dev’s will get to it.
“Dev’s will get to it”? Everyone knows that GS doesn’t have anywhere near as much developer support as GNOME/GTK+ or KDE/QT. And it won’t have as long as it has this confused objective. GNUstep has practically no mindshare, and no corporate backing, and it needs it. If GS developers want anyone to use what they’ve spent the last 10 years developing, they’re going to have to change that. Until then, there won’t be enough developers to make GNUstep a viable desktop platform, or a framework that plays nice with other desktops and WMs.
What GNUstep can do is offer the option of placing menus in windows (MDI, is what this paradigm is known as isn’t it?) to make it more flexible when it comes to use as a framework. They can also start working closely with more other projects (freedesktop, perhaps?) to make the framework more standard and do everything they can to get the word out there, write tutorials for sites such as this one, lobby distro makers to include GS packages, do some research on the end-user aspects like usability. I’m aware there probably isn’t enough manpower, but GS wants to be actually used by people, it has to take steps such as these and define whether it is a desktop or a framework, and perhaps split these two objectives into seperate projects, with the desktop teams building for the end-user on the framework team’s work for developers.
To succeed I think GNUStep needs to conform to the recommendations given at freedesktop.org. Things like drag & drop need to work well between Gnome, KDE, Mozilla,.. and GnuStep applications. Could sombody tell me if this is the case in the latest version.
you like ruby, gentoo and obj-c? we should talk sometime
aim me: omnivector
that’s awesom.e please tell me where to get that theme bundle
It is on the front page of gnustep.org that
gnustep isn’t a desktop environment.
Lacking of some applications doesn’t mean
that you can’t use GNUstep applications.
About portability, GNUstep is being used to port
a _lot_ of Cocoa source to GNUstep already and
vice-versa, which I believe you already know that.
> Ah, yes, but _why_ is there a lack of developers? Because GNUstep doesn’t know if it is a desktop or a development framework. And as a result it does both shoddily.
I can’t see your point. Lacking developers on Window Maker
support on GNUstep to fix some small glitchs has nothing to
do with if GNUstep is also a desktop environment or a simple toolkit.
And some people don’t even use GNUstep on WindowMaker and it is
still usable.
>iLogin hasn’t even had a release.
Anyone can pull it from CVS.
wildmenu. I don’t know where exactly to get it, sorry.
Maybe another default look would help. I don’t mean doing a MacOS/WXP/KDE/… clone, but maybe rounded corners, different colors (than now)… not something hard to do, but nice “eye-candy” (just some drops)
First of all, thank you Eugenia for those tips of FOX. It’s a really nice, fast toolkit, but I never got it to use my color scheme.
Now, concerning GNUStep. I am considering puchasing an Apple machine within the next few months, and I’d love got get some feeling for obj-c. However, I really cannot stand the look of GNUStep. I am convinced that GNUStep could look as polished as MacOS X does, but I haven’t seen any screenshots to prove that. Could someone at least give me some screenshots of GNUStep that just don’t look ugly? And changing a few gradients isn’t enough for that.
I will say that Qt is possiblly the best architectured C++ framework I’ve had experience with; and GTK+ is noble in that it does seem to stick to an OO paradigm, even though it’s written in C. GTK+ v.1 series had Obj-C bindings; I don’t know if GTK+2 series has them (yet).
Yes, thats roughly what I think. Qt/KDE is the first development framework that I’ve taken a liking to since NeXTStep. I did write some Objective-C bindings for Qt/KDE, and although the Qt ones worked well enough, by the time you wrapped the KDE classes the libraries were just too huge and slow to load. That was because Objective-C++ wasn’t available, and so I had to wrap the entire C++ api in C first and then call that from the Objective-C bindings.
I was interested to read you comment:
Well, name some open source OS X apps, and we’ll get to it. “Major” apps, like Camino the browser (or OS X Firefox), AbiWord, etc is pretty much just waiting on Obj-C++ support in gcc (coming in the next major version 3.5/4.0). This is unfortunate that it’s not _now_, but firefox/GTK+ is working just fine for the meanwhile. 🙂
Once that happens it would actually be possible to have viable Qt/KDE Objective-C bindings, although I wonder if there would be sufficient people interested to make it worthwhile. But there are a lot of very good non-gui Objective-C frameworks out there (not least the GNUStep Foundation Kit), and a binding would make it possible to integrate them with KDE.
It’s here, see the copyright file for the real url. http://www.linuks.mine.nu/i_debian/wildmenus/
You are the maintainer for Gnustep on Gentoo and you show up here, in public?
Ok, I’m kidding, but seriously, Gnustep is majorly out of date on gentoo it seems.
Like gnustep. I like the way it looks and everything
Very fast on my system as well.
Hey, Mister Ignorant, how dare you belittle Armando’s great work. His new Gentoo GNUstep ebuilds/packages have just recently been released, but since I doubt you bothered to check this before spouting off, you wouldn’t know that.
One or two people asked if there were any screenshots showing that GNUstep can be themed beyond changing gradients. You might find this interesting:
http://www.roard.com/screenshots/screenshot_theme29.png
Note that this is work in progress, and that it will hopefully look more polished when released as a Camaelon theme.
Yes, I know that the new ebuilds are slowly making it out (albeit package masked, or even hard-masked), and I appreciate the work that Armando is doing to get these things back up to date.
Just making a general observation on how bad thigs have been (and unfortunately to a certain extent still are) until the new ebuilds make it through. The current “stable” builds that were in portage until the last week or so wouldn’t even compile. Now we have the beginnings of something great, in unstable, at least…