Gnome 3.0 is a long way off still and I have already written a wishlist for Gnome, however this time around I thought I write some specific feature requests for the next major Gnome release (coming in more than a year from now). These are applications that I would like to see included by default on Gnome 3.0 for a more full and integrated experience.
Post a Comment
-a modified file selector. The new one is OK but I still cant open hidden folders.
-better network integration into nautilus. Its gotton alot better in 2.6
-a gnome equalivlant to kcontrol
-intergrate totem into nautilus
-MORE APPLETS!!! I cant get enough applets 
-better performance from gnome-terminal and gedit
Anyone know if dashboard might make an appearance in 3.0?
Actually, you are kinda late. You are talking about "gnome system" feature requests, while this article is about "user applications". 
Your comment should have being placed into my older article, the "Gnome wishlist": http://www.osnews.com/story.php?news_id=5361
>-intergrate totem into nautilus
That would be a mistake. However, there is already a Totem plugin for Nautilus that lets you view still movie thumbnails inside Nautilus (not play them though). Do you have installed that plugin?
Here goes my list:
1. A good P2P client like what KDE has in apollon. Need one integrated with GNOME.
2. A good integrated Bit torrent client. I don't want to have java installed just to use limewire (P2P) and azureus (BT).
3. A good IM client that also allows me to send files while chatting. Gaim does not let me do this at present.
4. A good easy to use photo album generator. Yes I am a camera junkie and like to post my photos on my private site for my family and friends to see.
5. An easy to use help file creator similar to Robohelp. Yes, I am a Technical Writer by profession.
That's all folks. I am happy with GNOME 2.6 on Gentoo but will be more happy if I see the applications I have listed above.
A Desktop Environment is an environment under which applications run. It does not have to include those applications you mention. The view that Gnome should come with many of the applciations you mention show that for you a DE is just like Windows. Why grouping everything together and create a mess just like Windows ? Why also kill innovation by bundling application into a DE ?
Let GNOME focus on being a DE. Let people develop applications. Whether those apps run under GNOME or another DE is not important. I run KDE as desktop, with apps using several look & feel: KDE, GNOME, X, Java, etc.
I prefer quality applications, even with a slight diffeent look, rather than one blob of a lesser quality.
Your article shows a Windows mentality. GNOME being a UNIX-based system, it must keep a UNIX philosophy: small is beautiful, focus on doing one thing right, and be modular. See Raymond's book if you do not know UNIX's philosophy and please do not try to impose Window's philosophy in a UNIX world.
Your article also shows a tendency to divide the Linux DE world. Many of the applications you list and say should be part of GNOME, either alreay exist as independent applications running perfectly under GNOME, or as applications using QT for instance. So why reinvent the wheel and not reuse the good things that for example are already bundled with KDE and work so well under GNOME ?
GNOME is a Desktop Environment. Period.
This article is tendentious as it aims to divide the Lunix world at a large, fueling the Desktop Environment war, and trying to impose the Windows philosophy in the UNIX Open Source community.
My KDE comment was there but there are always KDE comments on GNOME stories and vice versa, I still had some suggestions for GNOME.
Also, DE's developers are very responsive, for example the usability study that was published which outlined clearly what things need to be improved was immediately embraced. Now the menu is better split up and the toolbars have started being cleaned up.
I don't care if you use KDE or not, I was just asking you to give some suggestions to the project. Also, if you "gave up" on it say why so people can improve it and of course I'm sure you are not so close minded that your current decision is going to remain the same forever. Hopefully you will, if your suggestions are addressed reconsider, so by making your suggestions known, perhaps filing bug reports and wishes you are helping both desktops get to their goal faster.
System features, eh? Oh well someone else proably brought some of those points up. 
Do you have installed that plugin?
I can see the thumbnail, I was kinda hoping for an embedded preview. Like in Windows 2k/XP.
And your comment is nasty and unfair and you don't understand that Gnome and KDE are not simple window managers anymore, they are both frameworks and environments It comes a time that these DEs become big and mature and more is required off them.
You see DEs as you described above. I don't.
I see Gnome as a framework and collection of software with apps that are TESTED to work together with each other. This is what gives VALUE to the DE in the eyes of the OEM: the full solution. The ability to have a system that you "plug n play" on your product and have all basic apps and other frameworks already in place.
I see XFce as a DE in YOUR definition (being a pure DE). But I see KDE and Gnome being more like in my definition. 4-5 years they were just DEs as well, but times change and OEM providers need this pre-QA'ed, full solution. It adds value to Gnome/KDE.
I always see things from the business/user point of view (both the OEM and Gnome side). You see it from the point of view of a geek.
I see and want evolution on how things work or should work and you just want nothing to change. It's as simple as that.
I agree with most of the above, however this seems more like a wishlist for Gnome 2.10 rather than 3.0.
Take a look at the difference between 1.4 and 2.0 and compare it to the difference between 2.0 and 2.2. Gnome 1.4 is a completely different animal than the 2.x series. Granted, that was an intentional rewrite, but the fact remains that major version numbers represent major changes, and minor version numbers are evolutionary, like the addition of new apps.
For Gnome 3.0 we need to stop thinking about how Gnome needs a media player and a photo manager...these have already been solved. Rhythmbox is slated for 2.8, as is Totem and Evolution, and if it's not slated for inclusion in the platform, at least these apps exist (like Gaim...just about all Gnome users run Gaim, though it's not supported). There is no shortage of quality applications for linux, and the situation just keeps getting better and better. So if that's what we don't think about, then what DO we think about?
Major enhancements. Core technologies. I was hoping to see the community standardize on a managed environment for 3.0 (be it mono or parrot or to a lesser extent, java) but the issue seems to be at a standstill with everyone content to straddle the fence. This goes hand-in-hand with a really kick-ass, easily-accessible development framework. A full SVG desktop with next-generation functionality in GTK (OSX and Longhorn aren't sitting still on the eye-candy improvements...GTK has to evolve to incorporate some really cool stuff that makes use of accelerated hardware as that seems to be the way things are going these days). Some fruits of this proposed mozilla-gnome alliance, hopefully resulting in XUL bindings for GTK. Inclusion of gnome-volume-manager to finally put Gnome on par with WinXP and OSX in this department. Verify that Seth really is still working on Storage and see what can't be done about plugging it in to Gnome. A new kickass menu with a real menu editor. The list goes on and on and on.
So please, don't request applications to be included. Yes they should be included, and they will be included, but the more pressing issue at hand is "what do we want the next major instantiation of Gnome to be?"
It's both. A DE should provide a straightforward mechanism for interprocess communication and a solution for session management and integration. It should also provide a set of applications that are known to work reliably based on the framework it has laid out.
Really the two philosophies should be split into two different areas of Gnome: core system and applications teams. Not sure if that's the way it's done already (I'm afraid I don't know much about the internal structure of the team), but it would make a lot of sense for these two areas to be handled separately and have the teams work together to provide one another with what they need.
Eugenia please don't mix PhotoOrganizer and ImageViewer that are two different typs of programs.
If I could report abuse for calling my opinion "nasty", I would. But apparently common users can't do that against your posts, which is very unfortunate :-(
Also it seems that for you the DE is the platform. It is not. Linux is the platform Eugenia. The DE is integrated into a layered approach, and yes can be see as a framework, but it runs on a wider framework called Linux.
Accept you can be wrong sometimes Eugenia :-)
I expect to see a *lot* of apps in Gtk# finding their way into a few of these areas. F-Spot, blam! (nicely usable RSS aggregator, given it's age), MonoDevelop, etc are all showing lots of promise, and proving how useful mono is in rapidly/easily developing applications. I'm hoping that people's fears about this "evil MS tech" won't blind them to a nice new language and set of bindings.
>If I could report abuse for calling my opinion "nasty",
Your comment WAS nasty in the way it was written. It didn't read like a disagreement, it read like an attack. And so I replied accordingly.
>Accept you can be wrong sometimes Eugenia :-)
Why? I am not wrong on this one. I accept that XFce is a DE in the *traditional* meaning. Gnome/KDE are not and they should not be. Not anymore.
I think Gnome needs a real GnomeOffice, the think that is called GnomeOffice are just some apps but they don't play together.
And I think Gnome needs somethink like "Gnome Graphics Suite" or "Gnome Studio" where the graphic oriented apps worke nicely together.
With
- the main Programms:
GIMP
Inkscape or Sodipodi
VIPS
- and the tools:
autotrace.sf.net/potrace.sf.net frontend
SANE frontend
gphoto frontend
hugin - Panorama Tools GUI
Image viewer (gqview)
Image organizer (gThumbs/F-Spot)
This is a nice list, but this could be done within the 2.x framework. The features I would like to see for 3.0 are:
- Multiple programming language support (C++,Java, Mono and others)
- Storage
- HAL
- (The best) Support for mobile devices.
- Gnome 2.x applications must work on Gnome 3.x
- ZeroInstall from Rox
And gnome 3.0 shouldn't be done in a hurry. Continue the 2.x cycle and when the features for 3.0 are ready, release it. People don't care if the release number becomes: 2.12 or something. What they want is a Gnome that is productive.
You blasted Eugenia with "This article is tendentious as it aims to divide the Lunix world at a large, fueling the Desktop Environment war, and trying to impose the Windows philosophy in the UNIX Open Source community."
Your statement is utterly ridiculous.
Some of these are reiterated from the article, but that's only because they're really important.
1) Get Gnome to use alsa or jack (which in turn would use alsa), then get totem and rhythmbox to build against gstreamer. Solve the media issues by providing a solid framework. This is what gstreamer is supposed to do, unfortunately it's really confusing with so many ./configure options and not a lot of documentation.
Getting them to work embedded in mozilla/epiphany is another biggie.
Oh, by the way, work is underway to abstract the mozilla rendering engine (gecko) so building a full mozilla won't be needed for epiphany. Not only that, but apps that use gtkhtml like yelp and evolution can just start using gecko
2) Get GnuCash ported. This would be the crown jewel of the Gnome desktop, IMO.
3) Ditch all transparency hacks like nautilus select icons, transparent panel, etc. Make way for true alpha-transparency
4) Add right click menu for entries in the main menu, so that you can create new shortcuts and move them around. This is especially important with the new spatial nautilus, which has no address bar. I have to open up a folder, type 'CTRL-L,' then type "Applications:///" in order to add shortcuts to the menu. How usable is that, Eugenia ;P
5) While I'm on gnome-panel, make more changes happen automatically without a restart. New users will think they need to restart X (I open a terminal, killall -q gnome-panel). Aside from menu changes not updating automatically, I changed some settings in gconf-editor which didn't go into effect until I restarted Gnome completely.
6) Integrate hardware detection via HAL/DBUS
I would like to see a better and much integrated "sticky notes" program which will be much enhanced than what it is right now. I would prefer it can be emailed and saved in several open source formats. I think also they will to be less on memory foot print and offer other services like integration with your email client, personal management tools and other applications easily. Plugins can be easily made for that too.
I have not checked out GNOME 2.6 so dont blame if some of these features are there in that too.
I plan to embark myself in the development of such tools soon under MONO once I get acquainted with C#.
I have to have a presentation program, capable of reading Powerpoint files, as part of the office suite. Criawips made a silent release recently, so I'm hopeful ...
I already have Evolution, Abiword, and Gnumeric -- Criawips (or an alternative) is a necessary piece of the pie.
... Python should be an official binding -- no excuses.
We really have to dump XMMS. Its the only app I can think of with gtk 1.x dependencies.
Beep [http://beepmp.sourceforge.net/] is pretty nice(included in dropline gnome) but is still unstable.
This is mostly being solved by the Programming Bindings platform for Gnome. Which should continue to develop, and add more languages to the list.
But can I make a java class and call it from c# without getting a headache? Making it something like:
Java:
package foo;
class useless{}
c#
Using foo.useless;
etc.
And no I dont want it to be CLI or bytecode only.
A new software map (a GNOME apps page) seems more important so that people interested in a coherent look are able to get an easy overview of all those optional apps available. This is also interesting for developers to decide where some more work is needed.
All these apps might differ in GNOME integration or HIG'nification right now but its the users choice and the projects developers attitude may change in future.
This way, the GNOME project might be able to propose a seperate package for all major application areas, for example Office, Multimedia, Internet, Science, etc., in the near future.
As a sidenote: The divide of the Linux DE world is already happening due to users vote, IMHO. Having a bunch of different file selection dialogs within your desktop (for example) is irretating and un-productive.
What we need the most, is a new IDE and an MSDN type API document system. IMO that should be number 1 on the list. It's still far too uphill for new developers.
A little off-topic... but I'd like to report that in DarwinPorts have been recently committed many Gnome 2.6 ports ( http://darwinports.opendarwin.org/ports/?by=cat&substr=gnome ): GNumeric, Abiword, GIMP, Inkscape and the other mature applications are, in my opinion, the real valid point of GNOME and their integration in Mac OS X Desktop (through X11 Xnest) is really enjoyble.
There is a apps.gnome.org project going on. The page is somewhere out there and looks pretty good. Quite a few applications are still missing, though, but this should be easy to fix now that the setup is there.
@ThanatosNL
Here stuff in menus etc are instantly updated with help from fam. And I can also right click in menus and add/remove items. Dragging stuff around doesn't work, though.
I really liked this list as well as a lot from the previous one. Though I would like to think of this being doable within the 2.X cycle, at least parts of it should be.
One thing that bothers me is when using QT based applications, they feel so much out of place. This should be improved. Even though the needs those currently fill will be fulfilled by Gnome apps in time, there may well be some great application using QT that won't fit in.
Hopefully the work done on Suse/Ximian will lead to better integration.
And I don't see how you can stand the default look. There are others out there that are simple and look so much better.
There is going to be a GNOME 2.10, 2.12 even. 3.0 will likely be somewhat incompatible with 2.x (ABI/API incompatible perhaps?). And no, it won't be like 1.x -> 2.x. That was a complete rewrite.
I know this has little to do with application suggestions, but I really want to be able to change the gtk icons from within the theme manager.
There is nothing wrong with a theme having a default set of icons, as long as I can override this in an easy way.
Thing is, KDE and Gnome both have guidelines for how applications should be (look, feel, behave). The Gnome HIG is pretty good and with more refinement, devs should be able to create any of those applications listed (and more) with good integration with the target DEs. This will give us a more level playing field and allow different apps of the same class to 'compete'. If you bundle apps with the DE distribution, it becomes more painful to switch apps.
The upside of doing that, of course, is that you 'encourage' people to not reinvent the wheel. But is this necessarily a bad thing? Sometimes a bundled app is not good enough but because of faulty design, it's not easy to fix properly. But it's still part of the bundle!
Really, it should be left to the user to choose apps that they want. Not everyone will want the apps that Eugenia wants, for example and I don't think they should be forced to download and install those apps - and all because it was bundled into the DE distribution.
I think it was a good idea for KDE and KOffice to be separate projects. KOffice still is very integrated with KDE, though. This separation makes the individual projects more flexible and nimble - they don't have to wait on each other as much. You only need to agree to HIG and API and it should all jive. What if an app needs another app for a certain function? A good example is KDE's Ark. It will function with the generic archivers (most unices will have tar and gzip) but when you try to process, say, a .zip file and you don't have zip on your system, Ark will say so. Then, all you need to do to get the functionality is the install zip - and presto, Ark works with zip files.
Isn't this how we want things to work? The DE sets the standards, HIGs and APIs and projects that want to target these can subscribe to them. The only time you don't want this, is if you want to limit competition. (M$?)
A further example, the CD burning thing on Gnome. The point is any CD burning app should be able to integrate with nautilus! The user can then set the default. Not all app developers will place the same importance to the same list of features.
This level playing field and choice is what we should keep. Let the linux distros decide the exact bundle they want - and the distro user can still choose to install another app instead later on and still have the same degree of integation. Compare: Windows comes with IE. We can still install Mozilla. But can we remove IE? No. That sucks.
Do we want this on linux/BSD?
Hi.
Most interesting article... Anyway, what I think Gnome lacks is a good help system. Or does anyone use it at all? I recently browsed through the help system in my Gentoo Gnome 2.6 system and I was struck by the lack of documentation in the help system.
As we expect people to migrate from other platforms to Linux and perhaps to the Gnome GUI, it must have a very extensive and high quality help system with plenty of information to aid the user. Integration with other Gnome apps into the help system also would centralize the help system. Selecting help from the main Application menu would display all the help available, and selecting help from the app would move the user directly into the help section for that specific app. Perhaps a troubleshooting navigation system too integrated into the help system would give the user a better help section experience.
Anyhow, that is just my 2 cents
/nexim
>You blasted Eugenia with "This article is tendentious as it
>aims to divide the Lunix world at a large, fueling the
>Desktop Environment war, and trying to impose the Windows
>philosophy in the UNIX Open Source community."
>Your statement is utterly ridiculous.
I think its not, Gnome is a window file manager you can look that up on the Gnome website. Its also a framework for other applications. Why should everything be intergraded into Gnome?
A financial suite? why? you can download it yourself or choose to install it with the distro you use. I rather ahve a lightweight fm wich i can extend to my wishes but not because some person thinks its necessary to put in a p2p client of finacial suit per default.
Gnome applications should be separated from the constraint of the host operating system. Gnome applications should not depend on what package format, file layout or hardware architecture the host operating system is running under.
The next DE, whether it be KDE or Gnome should follow what mozilla-firefox is doing. Provide a cross platform base (mozilla-firefox) in windows, Linux, *bsd, etc... and allow users to download and install whatever apps (extensions) they choose to install. The apps should only rely on the base(i.e. Gnome KDE) without worrying about the underlying host operating system (SuSe, Windows, FreeBSD, OS X). This will eliminate the common installation problem that users face when they deal with open source software.
In order for this idea to succeed, the Gnome team must provide an IDE that not only can compile code but also create an a Gnome specific package that any user can install.
"The GNOME project provides two things: The GNOME desktop environment, an intuitive and attractive desktop for end-users, users, and the GNOME development platform, an extensive framework for building applications that integrate into the rest of the desktop."
I would be all up for a Gnome-application-pack and/or Gnome-application-support-pack that provides all of the different applications mentioned in the article and the different comments above, while the second pack would provide all the items that help applications integrate easily into Gnome.
I do believe Gnome should keep to the above quoted outcomes and be separate from the "install everything by default" mentality.
Just my 2cents.
Omega, I agree with you.
Eugenia: I disagree. The open source community should not try to reinvent the wheel, but eg. try to integrate KDE/QT applications into Gnome, so they look and feel like they are Gnome applications.
There are too many resources needed to make 2 different Instant Messaging apps (one for KDE and one for Gnome). I think the community should focus more on standards - like the ones at freedesktop.org - and less on reinventing the wheel.
For instance: I really like the KDE app Kile (for LaTeX). Should I make my own Gnome app, just copying all the features from Kile ?
Note that none of these things are at all related to "GNOME 3.0", and the sane ones at least are almost certain to happen before we have a major number change. There are *no* plans at all for GNOME 3.0 apart from knowing that we will break API/ABI compatibility (and even then, it will probably just involve removal of APIs deprecated during the 2.x series).
This would be more appropriately titled as, "My Wishlist for GNOME", without invoking the dreamy, totally abstract "GNOME 3.0". :-)
Just to say that syncing a pocketpc device is possible.
See :
http://synce.sf.net/
I sync my ppc everyday with evolution (using multisync) over wifi and it works great. I can also browse my files, install apps etc..
Most interesting article... Anyway, what I think Gnome lacks is a good help system. Or does anyone use it at all? I recently browsed through the help system in my Gentoo Gnome 2.6 system and I was struck by the lack of documentation in the help system.
There are problems with the help system, and with a lot of the documentation, and there are people working hard to fix those problems. As with any free software project, it will get done faster if you help.
I'm curious exactly what problems you have with it. We can't very well fix problems we aren't aware of. We rely heavily on feedback from users. Unfortunately, the type of people who run betas and submit bugs are generally not the type of people who often read help manuals.
As we expect people to migrate from other platforms to Linux and perhaps to the Gnome GUI, it must have a very extensive and high quality help system with plenty of information to aid the user. Integration with other Gnome apps into the help system also would centralize the help system.
The help system is already very centralized. There's also an effort underway to develop a new system on freedesktop.org that will be shared by KDE, GNOME, and any other desktop that chooses to use it.
Selecting help from the main Application menu would display all the help available, and selecting help from the app would move the user directly into the help section for that specific app.
This already happens, to a point. What's mostly needed is a more topic-oriented organization of the help files, which will allow applications to point to specific topics, depending on what you're doing.
Perhaps a troubleshooting navigation system too integrated into the help system would give the user a better help section experience.
Certainly. There are a lot of cool things we could be doing. And most of it requires a good deal of infstructure work. It's not the most popular thing to hack on. If you like doing this sort of thing, you should join in. We like new contributors.
IMHO we need a good media framework, not only for GNOME, but also for all other apps, Gstreamer, Mplayer and jack are the best candidates, the best solution should be a fusion of them.
A financial suite? why? you can download it yourself or choose to install it with the distro you use. I rather ahve a lightweight fm wich i can extend to my wishes but not because some person thinks its necessary to put in a p2p client of finacial suit per default.
I agree. Not even Windows integrates MS Money, Visual Studio, or Office. Even Smaller apps like Picture It! or MS Reader aren't included (heh, yet for some reason Windows Movie Maker is. Go figure). I know financial software is popular, but not that popular.
Yet I do think a Browser/Media Player/Image Viewer/etc., should be included/integrated. Apps like that are there to "view" a vast amount of content.
They're "neccesary", at least in most cases. For the most part, GNOME has it covered I think, except with rich/media playing.
Not even Windows integrates MS Money, Visual Studio, or Office. Even Smaller apps like Picture It! or MS Reader aren't included (heh, yet for some reason Windows Movie Maker is. Go figure).
Uhh..Before someone misinteprets that as saying "because MS doesn't do it, Gnome shouldn't", it's not what I meant. I'm just saying that there's "popular apps", and then there's "necessary" apps -- usually viewers of some sort.
Just a quick note on modem dialer, does Gnome really need that ?
Modt modern distros ship such a toole ,e.g. redhat-config-network
which makes it a nobrainer to configure network devices including a modem.
http://www.osnews.com/story.php?news_id=5361
"Gnome needs to enter the 21st century and create a new theme and also modify GTK+ to support new features that take theming to the next level. I am not talking about yet another mediocre theme like most on art.gnome.org, but something "wow," something that can draw new users like a magnet and be clean and professional at the same time. First impressions do count and looks too. Unfortunately."
Gnome has always looked better than the window managers I've actually use, so did it matter? no.
However I use kde now and like it =), and yes, it's possible to make it look good aswell.
This is mostly being solved by the Programming Bindings platform for Gnome. Which should continue to develop, and add more languages to the list.
But can I make a java class and call it from c# without getting a headache? Making it something like:
Java:
package foo;
class useless{}
c#
Using foo.useless;
etc.
And no I dont want it to be CLI or bytecode only.
This is possible, but you can't have your cake and eat it too. There's a tool called ILKVM(sic?) that converts your bytecode to CLI. I don't know what you mean by you don't want it to be CLI or bytecode only. It has to compile down to some common bytecode.
RE: devs
Gnome really does need a first class IDE. Anjuta isn't there yet, but to be fair Kdevelop was started in 1998(IIRC) and took nearly 6 years and a complete rewrite to be usable. Monodevelop could possibly fullfill this role if people implement parsers and some other plugin work for other languages.
I tend to agree with the poster that said the DE is not about apps. Sure, having a official media player, IM app is cool, but pretty much orthogonal to Gnome 3.0 as a system. IMO, it's much more important to work as hard as possible on issues like dbus/HAl, maybe an official runtime, for the Gnome 3.0 system then worry about apps that can do their own thing without really worrying about infrastructure issues. Please core Gnome people, I know Mono is not politically-correct, but if it turns out that if it's officially Royalty Free, make that the official runtime. If not then use Java. This statement by Havoc Pennington does not look promising. "In the meantime though, to me it looks increasingly like we need to go down the road of Python, XPCOM/UNO, GObject introspection, and other half measures;"
Half-measures? This is just a recipe for disaster that will most likely haunt Gnome for years and years to come.
My main wish for GNOME is that it's UI guidelines are blended with KDE's, so that apps written for either DE have a consistent feel. That would effectively double the number of apps you can run in either environment without turning them into an inconsistent mess.
That on it's own would go a long way towards making Linux as pleasant to use as Windows. At the moment neither DE has all the apps I need, the inconsistent GUI you get when mixing KDE and GNOME apps is one of the main reasons I really hate using Linux.
There are some efforts under way to make that happen. Take a look at gtk-qt for example.
http://www.kde-look.org/content/show.php?content=9714
http://www.kde-look.org/content/show.php?content=11264
or OpenOffice with the KDE Native Widget Framework:
http://dot.kde.org/1082652256/
1.) Why would OEMs prefer a "complete" desktop with applications? Wouldn't OEMs rather take distributions?
If it's just about making other people's life easier for deciding which applications to use, then GNOME should have a "prefered applications" list, I don't see the point of adding the applications to the desktop project just for this...
2.) The fact that Mono won't be in the core plattform project is a huge reason IMO not to add all applications to the desktop project. For example I'm not a fan of Rhythmbox as part of the desktop, when Rhythmbox isn't really integrated anyway and there is a very nice alternative based on Gtk#.
I really don't see any good reason to add every useful application to the desktop project. Surely the fact that finished desktop products should ship with those applications is not a good reason alone. I find it rather gross that an article like this is written without even acknowledging that many people in the GNOME project do not share this kind of thinking. If they would, we'd probably have Galeon in the desktop a long time ago and then replaced with Epiphany later (for example). I'm not sure if GNOME would have been better off with that philosophy.
I don't really understand this one. C# is free am I right? Some libs provided by Microsos aren't, but we don't use them to develop Gnome apps, are we? If there are some wonderful apps written in Gtk# I would certainly not hesitate with their integration into a future Gnome.
And the Python thing: It is not included in Gnome, but it is included in every distro I would think of. So what's the point?
Voluntary developpers should not be limited with tools they should or shouldn't use. They will not code their apps in languages they don't like. Why throw their work away?
Ive heard of lkvm before, but my idea is different. Just imagine: compile a java class in windows(with j2sdk) and use it in c# on linux! So what I want is: a shared backend for the frameworks, and perhaps also for other languages. There is already a product for this on the market: http://www.jnbridge.com/. It would just be awesome to have something like that for the Gnome platform...
I agree that including best-of-breed applications into Gnome, as long as they comply with the HIG and play nice, is a good thing. However, I do not like KDE's approach where they include too much. Gnome does not need three text editors for example. One application for each of the tasks they wish to integrate is sufficient.
I hope that some of these applications do make it into Gnome, but in the process of doing so, I hope they do not hamper Gnome's simplicity or its elegant usability.
I also hope that Gnome remains manageable as a project, and the inclusion of these applications does not hamper its development and growth. In my view, Gnome is becoming the most advanced and most usable GUI on the earth, period. Let us hope it stays on that path.
This article really disappointed me, I agreed with almost none of the points in it. The point that really got my blood boiling though was the ALSA remark. Why should gnome focus exclusively on this Linux-only technology ? Gnome is NOT linux only, people use it on Solaris, FreeBSD, etc and these platforms still use OSS!
I thought this was the entire point of the Gstreamer framework BTW, putting an abstraction layer between apps and backends like OSS and ALSA.
I think that along with these aplications, gnome, as a plataform, needs better development tools. Mono will be an advance.
Excess of dependencies mades install linux (kde/gnome) apps a really terrible pain. The way Mac OS X does it's really interesting... We need something like that (and not a dependencie hell, with rpm), working in any distro.
CD Writer. Well, the best is K3b. Gnome needs something full featured like that.
F-spot - would be really good if it was integrated in Gnome (yes, I know, it is a mono app, and I'd like to see gnome 3 built on mono).
A new Control Center (what's about Eugenia design it?) :-)
A ssh gui client, similar to SSHClient for Windows Workstations. The one from www.sshtools.com is good, but not as good as one native and full integrated client would be (it's a java app).
Build Epiphany without any dependencies of Mozilla.
A Chat/IM client that suport audio/video/file transfers and fully integrated with the environment (like evolution address book).
A really good midia player, with plugins to my web browser and fully integrated with my system.
A better file manager, like Path Finder.
A samba Browser, like smb4k (nautilus is not that good with this task).
I think you are being a touch unfair there, Chris. The situation with Kate/KWrite is in transition, because of internationalisation efforts. There will be two final editors though, KWrite for notepad/wordpad work, and Kate for programmers to edit code with. Presently, there are issues with right to left language support, which is why three are required.
In terms of other KDE applications that get frequently bundled, there are only single applications for classes:
Kopete - IM
KMail - E-mail
KNode - Newsreader
Konqi (FM) - File management
Konqi - Web browsing
Kdevelop - IDE
K3B - CD burning
As you can see, the KDE project includes applications typically one per area of endeavour. There are some cases where transitions between applications (for example, when discussing audio/video players) make it essential to package two applications until a situation is resolved, but by and large, there is just one.
Where Eugenia is running along the edge of corrrectness, is when you consider that while a number of applications (such as Kate) are in the KDE Base package, others are typically farmed off to other packages. When there is a KDE release, the core packages get updated, but not always the non-core ones, and those, similarly can follow their own release schedules, like KDE PIM is, right now.
Because of the separation of elements in this way, it is possible for distributions to pick and choose the packages they install, and the applications. The KDE project says "These are KDE-based applications that we think are the best in their field for KDE... Give them a try if you like.". Eugenia would prefer to see GNOME as a bundled, inseperable package with all applications and everything bound in. The truth is, GNOME as a project is, and always has been a GTK-based panel, system manager and set of frameworks for applications. Applications themselves are GTK applications, and they happen to all look the same under GNOME, but that doesn't mean that they are part of GNOME. In KDE terms, everything is tightly coupled due to component reuse and inheritance, including structures, frameworks, messaging systems, etc. But with GNOME, they are very loosly coupled. There are plenty of GTK applications fulfilling Eugenia's needs out there, but she wants to bind it together in a way that disagrees with the core philosophy of coding GNOME and GTK applications.
Just get a distribution such as Mandrake or Debian! They solved Dependancy hell YEARS AGO!
Nautilus what's the difference between what you're suggesting and what mono provides? From what I can see you're suggesting a common format that .NET and Java can both be translated to. IMHO this is a silly idea, as now there is not just one level of abstraction between the bytecode and the processor, but TWO!
IKVM is essentially a java bytecode loader into the mono environment. It makes your java code run within the mono framework. What more could you ask for??
You're right with the headline.
Gnome needs something similar to Konstruct.
I know there are some things (cvsgnome, garnome, etc) but they don't work.
You know, distributions like Suse don't support Gnome and it's difficult to try a new version without pre-build binaries or without an easy to use compile-script.
A better solution would be distribution-independent binaries of course.
What's this about "core philosophy"? Take a look at the GNOME Desktop and Developer Platform dependency trees some time. It is not at all what you've described. :-)
Konstruct was based on an early version of GARNOME! If GARNOME has not worked for you, ask for help. You can also use jhbuild if you want to build from CVS.
SuSE does support GNOME, although in current versions it is not so great. That is changing now that the Ximian desktop team are part of SuSE. :-) The http://www.usr-local-bin.org/ project provides polished, recent versions of GNOME for current SuSE systems, too.
"Gnome is like KDE, its missing all the good features that Windows already has. No easy application installation or removal."
Ok, I think I'm just going to become the official OSNews Linux package management myth buster.
Repeat after me: package management is better on Linux than it is on Windows, so stop pretending it isn't (or stop using Redhat!)
The reason this poster probably has trouble adding and removing software is because he uses Redhat or some related RPM-based distro, which uses the same kind of system Windows uses. Installers are wrapped up into compressed files, shit is thrown all over your system, and you are left with the task of cleaning up the mess.
In Debian and Gentoo, with their respective dpkg and emerge systems, dependencies for programs are handled, and installation and removal is MUCH easier than it is in Windows.
Personally, on my Windows box, when I have to remove a program I have to do: "Start->Settings->Control Panel->Add/Remove Programs."
Then I wait approximately 60 seconds while Microsoft queries its own registry to compile the program list.
Then I click "remove" and have to watch progress bars and occasional complaints that my *.isu files don't exist, or what have you.
On my debian box, when I want to get rid of balsa, for example, I just do:
apt-get remove balsa
And whammo, it's gone.
Doesn't get much easier than that.
Plus, if I want to install balsa again, I just do
apt-get install balsa.
And whammo, I have it back. How would I reinstall Thunderbird in Windows after removing it? Oh yea, I'd go to www.mozilla.org, download a thunderbird binary install, run the setup program and then it'd be installed.
So, what was that, Linux lacks good software installation/removal, you say? Oh yea, that's right--THAT'S PURE FUD.
Argh, when will people learn.
That's the point. Distribution-independent binaries. We need that to gnome and to any other linux app. We need an international standard to build linux distros, maybe dictated by OSDL. As I said before, the way it's done in MacOS X is one of the best/easiest ways of doing that.
Like Eugenia, I see linux/gnome/kde as a product, from a corporate/user point of view, not as a geek toy.
We need address these things to compete with Windows in corporate environments.
And more, we need a standard plataform, not a lot of peaces put together. We need an integrated environment, from the base system (kernel), to x server and the desktop environment.
Gnome is NOT linux only, people use it on Solaris, FreeBSD, etc and these platforms still use OSS!
I thought this was the entire point of the Gstreamer framework BTW, putting an abstraction layer between apps and backends like OSS and ALSA.
Jack would be a better cross-platform choice. It's really high quality.
About gstreamer, gstreamer is only used for multimedia apps. You still need esd to play sounds. Eugenia (correct me if I'm wrong) was talking about, for instance, the sound that plays when you start up Gnome. What would be nice is to have Jack replace esd, and have gstreamer just use jack. Getting audio to work with Gnome right now takes work--I have audio fully functional (i.e. gnome sounds play for wm events throu esd) but volume control doesn't work. It's because I use devfs. I don't use devfsd to clutter up my /dev dir, because I use devfs to clean up /dev in the first place...I ought to just try udev
From the Jack webpage:
JACK was designed from the ground up for professional audio work, and its design focuses on two key areas: synchronous execution of all clients, and low latency operation.
Ever notice how out of sync esd can be? Solved
"Like Eugenia, I see linux/gnome/kde as a product, from a corporate/user point of view, not as a geek toy."
Then you ought to be able to grok why what you're proposing is a phenomenally difficult thing to achieve... It's not going to happen in the short/medium term, so please get used to it and be happy with our success anyway. :-)
The question is: Why do we need to deal with dependencies? In my point of view, dependencies must not exist. The most used libraries must be on the distro. If you build some app and that app needs extra libs, put it together with your package. That's simple like that. About Debian and Gentoo: They are not targeting corporate/home users, but they are geek tools/toys (no ofense here).
Now, look at how easy is to install/remove apps in MacOS X and in Windows. That's is like it must be. If not, linux will always be a server OS/geek toy. If I had the money, my desktop would be a Mac and not linux because of these issues.
I agree 100% with the guy who said the way to go is encouraging new developers, and to do that, Gnome must have a good IDE (MonoDevelop is heading on that direction) and a good MSDN-like website (which as far as i know, it doesn't exist yet).
Victor.
Yes, I know. It would happen just if someone with the money and structure takes the work and build it by himself. Well, it's not a good idea to do that today, unless you do like Apple: build the OS and the hardware. Your money (return) will come from selling hardware and extra apps for your OS.
Maybe Novell will do something about that, but I don't believe they has the need resorces to do that. By the way, they are a Network/server comphany.
It can be done, yes,it can be. And I think that it will not be done by the community, but for some corporate with the resorces (and intere$t) to do what must be done.
Good apps are built with good tools. MonoDevelop looks promising but there are a lot of people out there that wants a visual tool (rad and case tools).
There is need to improve the toolkits too. That's time to have standards, stoping to break thinks all the time (just look at glibc, for one small example). There are a lot of issues to solve.
One issue with Gnome is that it was not designed to be a desktop plataform. Why not to stop doing a lot of small things and sit down, design the APIs, the entire desktop as it must be and then work to build it with high quality?
The question is: Why do we need to deal with dependencies? In my point of view, dependencies must not exist. The most used libraries must be on the distro. If you build some app and that app needs extra libs, put it together with your package. That's simple like that.
Simple, yeah, but recursive dependency handling is a better solution because by nature it will mean smaller binaries, which translates to faster load times, more disk space, less shared code among binaries, and faster updates. With most libraries (think OpenSSL for an exception), you don't have to recomplie software built against them when you update them.
The FLOSS world of software moves fast--we need advanced frontends to package managers (like portage is to ebuilds, apt is to dpk and rpm, prt-get is to pkgtools, slapt-get is to .tgz, etc) to keep up.
GNOME includes a developer platform to build applications with. If we threw it away, sat down and "designed it just right", we wouldn't have any software to ship. Incremental improvement is a very important part of the process.
I think that is ok that the base OS has some dependencies. But, for the apps, that kind of apps that you go to any store and buy, lets say, an photo manager, an music organizer, well, evolution, it's necessary that it's easy to install, without any "external" dependencies. When you will update it, well, download just a single package and it's done. The way it's done today is really painfull. I have to donwload dozens of files instead of one single .dmg file, for example. The way it's done today don't save any disk space and give me a lot of headaches to get a simple app working/updated. And, no, ebuilds, apt, urpmi and others are not a solution if you are out of the geek world.
Wow, this guy wants some stuff that he may not see for a while. Some of this stuff is pretty specific. I see the video editing want, because cameras are becoming more and more popular.
The video messaging only sort of works on proprietary messengers.... Most of the problems are networking and protocols though (firewalls and such).
is that so hard to provide some .h wrappers for gtk1.x series and one or two .so in order to recompile without retouching any line or code or even running them without recompiling ?
list of gtk1.x apps I can't live without :
xmms
lopster
xmule
gtktalog
gcombust
grip
cantus
and there's more and more...
every 3 years, desktop leaders folks want us to change every one of our habits because it's good for us....
I've read in joelonsoftware.com that microsoft spend a lot of money in order to achieve backward compabilities, it's normal, in real life it's called respect to the user who sticked for 3 years into a desktop, who buy distro, soft and who is not able to still run them once he changed his distro.....
What's the problem with just running GTK 1.x and GTK 2.x apps side-by-side? You're perfectly capable of running both sets of libraries concurrently.
I don't like your "I still have to use K3B" Eugenia. IMHO gnome should focus in make a better integration of kde and gnome, like the KDE guys are doing with openoffice. K3B is good enought, integrate it in gnome...
Unfortunately, between GTK+ 1.x and 2.x, it was indeed too hard. However, we have committed to API/ABI stability throughout the 2.x series and (despite Eugenia's creative article headers) we have NO plans for an API/ABI-breaking 3.x release - it is unlikely to happen for a while. The big 1.x to 2.x break was done so that we would have a better opportunity to maintain stability for a longer time.
I don't like your "I still have to use K3B" Eugenia. IMHO gnome should focus in make a better integration of kde and gnome, like the KDE guys are doing with openoffice. K3B is good enought, integrate it in gnome..
Actually that is a very valid point. I use K3B in Gnome as well. In reality, if you can get GTK+ and QT based linux apps to interact as well as possible, everyone will win.
While I prefer Gnome overall, there are some really awesome KDE based apps that I use regularly.
Hopefully freedesktop.org, and maybe Novell, will help facilitate this kind of "win/win" interoperability. Sure Gnome and KDE can still compete to win, but they can create an atmosphere of co-operation too, to take the Linux desktop to the next level.
good points!
"i think that along with these aplications, gnome, as a plataform, needs better development tools. Mono will be an advance."
correct. hope the legal (if there really are those) issues can be solved
"Excess of dependencies mades install linux (kde/gnome) apps a really terrible pain. The way Mac OS X does it's really interesting... We need something like that (and not a dependencie hell, with rpm), working in any distro."
for anything else than the core packages - autopackage.org
...
"A new Control Center (what's about Eugenia design it?) :-)"
definitly!
but because the distributions also are adding tools which are not part of default gnome, it needs to be modular!
"Build Epiphany without any dependencies of Mozilla."
mozilla.org is working on that already, seperating gecko from the rest.
"A Chat/IM client that suport audio/video/file transfers and fully integrated with the environment (like evolution address book).
A really good midia player, with plugins to my web browser and fully integrated with my system."
!!!
"A samba Browser, like smb4k (nautilus is not that good with this task)."
i read somewhere in a review that samba-shares are now automatically mounted in gnome 2.6 - is this really true? can't really believe it - does it work nicely?
in general, i wish gnome to become less esoteric and that they are focusing more on fixing/improving stuff than focusing on adding new shiny stuff, or at least 50:50.
and pls stop changing things which worked well and were heavily used (like the "non-spatial" nautilus and the "extract here..." - option - is still miss it:-((!), shoving it down the users throat!
ps: my number-one app-wish is a combined web- and filebrowser which windows and kde have for ages! but alas, because it hasn't arrived yet, it is somehow against some gnome-guidelines;-)?...
No it's not the same, to be more clear on things, this is what I want in Gnome: http://www.jnbridge.com/usecase.htm.
"integrating" gaim-vv with gnomemeeting and evolution, also consolidated contact list?