Gnome has been plugging away with its 2.X series from quite some time now, updating every 6 months on a predictable schedule making incremental improvements with each release. During this period they have kept their API stable and have refrained from making fundamental changes to the project. The developers have acknowledged that at some point in the somewhat near future, they will break from this series and begin work on a new series that removes some of the old cruft and changes some fundamental approaches in how people use Gnome. Here are a few of my suggestions for what would help Gnome 3 a revolutionary leap forward.
pretty good read.
The mono/gjc/java part I find most interesting. Something certainly needs to be done about this. I’m all for everyone getting behind and pushing the same thing if possible. I’m not an expert, but from what I’ve read gjc doesn’t have the cloudy patent future that mono has. Wish they’d gpl java and end all this garbage so we can get moving. I Don’t blame Sun I guess, just sucks.
“In the interim, the Gnome project has accepted python as a mainline language, but none of these [Novell,Sun, Red Hat] companies are actively writing new tools for Gnome in python, as far as I know.”
First show me where it was decided that Python is Gnome’s interm solution. Python isnt even included in Gnome Desktop release and is just now being included in Developer release.
Second, everything Red Hat does is written in Python including the new Sabayon management tool.
“This C#/Java fragmentation cannot continue unchecked. Novell is writing great apps in Mono, and Red Hat and Sun with their Java implementations.”
Red Hat hasent written 1 desktop app in Java. They are a bunch of hypocrites. They push Java but they still arent using it themeselves.
“What if Microsoft decides to stop Mono by enforcing vague software patents? What if Sun decides to do the same to gcj?”
Sun offers TCK and a patent grant on Java implemenations that are qualified. MS dosent offer that with .NET.
Conclusion: Ximian/Novell are irresponsible idiots.
Like, unbreaking gnome-vfs so apps like gedit can edit remote documents? GNOME devs can’t make a simple VFS system work – which the KDE devs seem to have no trouble with – how are they going to implement a new ‘Storage API’?
Or, fixing nautilus so it doesnt risk deleting files when you are renaming them and the view refreshes? – this is just a pet peeve of mine, but using this file manager means risking the safety of your data
Or, speed it up and reduce memory consumption – I know work is underway on this.
And a menu editor? – you know, make it so users can actually control what they see in the various menus scattered around the GNOME desktop e.g main menu/Places etc.?
And the ‘spatial’ paradigm – what good is it if the GNOME components e.g. EOG don’t implement the ‘desktop metaphor’ consistently – every time i view an image i have to adjust the freaking info pane, depite the fact the image window will scale the image to fit – settings in GNOME core apps do not ‘stick’ like they should if the desktop is to be ‘spatial’
Don’t get me wrong. The concept of metadata-driven desktop search is very good. I just find terrible the concept (that seems to be pervasively creeping in GNOME IMHO) that “the user must think what this desktop developers think”.
The only filesystem characteristic that would ever be presented to the user is the physical volume
When saving, the only choice the user would have to make would be which physical device to save to.
I hope you’re kidding. I want to save my files where I want to, the way I want. If a user want to rely on the GNOME storage search, well, there’s nothing transcendental to do: let the user save everything on /home/user, without building any folder. But as a user I must have the choice to do a 100-folders-deep home filesystem, if I feel the need to.
Second, this breaks any compatibility with non-GNOME apps. When I backup my filesystem (say,to a DVD-ROM) and I open it on a non-GNOME system, what do I do with the horrible mess will appear with files all in the same location?
Third, what do I do when, for example, I have to save different versions of the same file with the same name? This is by no means exotic. When I wrote my graduate thesis I had all LaTeX versions of my thesis in different folders. When I code I have all versions in my home directory in different subfolders with version number written. They would simply clash if I save ’em all on the home directory.
The file storage concept you present is cool, but remember of users that want control on their system. If I didn’t want that control, I’d rather use Windows or MacOSX.
“I hope you’re kidding. I want to save my files where I want to, the way I want. If a user want to rely on the GNOME storage search, well, there’s nothing transcendental to do: let the user save everything on /home/user, without building any folder. But as a user I must have the choice to do a 100-folders-deep home filesystem, if I feel the need to. ”
Since this idea is about layering search on top of the existing filesystem, it would be trival to have an option to retain the current open dialog for advanced users who have invested time and energy with their existing heirarchies. Most users have a lot of trouble with the current system, next time you are on a inexperienced users’s system check out how cluttered and flawed their home directories are no matter what OS they use.
“Second, this breaks any compatibility with non-GNOME apps. When I backup my filesystem (say,to a DVD-ROM) and I open it on a non-GNOME system, what do I do with the horrible mess will appear with files all in the same location?”
You’d have a list of all the files you created. How is that a problem? I’m not sure how longhorn’s winfs or apple’s next-gen system will interpret it.
I don’t want a beagle file opener. I don’t even care that much about beagle per se. What looks far more intriguing is the implicit query version that they are currently calling dashboard. It would just be a slimmish bar on the right hand side of the display that always had information in it pertinent to what you were currently doing. Here’s a reminder picture:
http://www.nat.org/dashboard/rewrite.png
Now consider that any time you open up a file dialog that dashboard pops open the most relevant matches for you. Just pick the one you like and drag/drop it to the file dialog. If you want to refine it by hand then pop open beagle and search some more by hand.
Michael
Make it faster, consume less resources and simpler. All these self-important time wasting resource hogging eye candy INutilties and DISservices must go.
Make it so lean and mean it makes people feel they just bought a new computer on which they can get their work done faster so they can get on with their lives.
Oh well, a guy can dream right?
When oh when will menu editing work properly? its NEVER worked quite right, ever.
Mac OS X uses ALT-TAB for changing applications and ALT-~ for changing windows with in an application.
Something simular in GNOME would be nice. Currently ALT-TAB works the same way it downs on MS Windows; which is take me to the next open window (which could be in any application).
Hopefully they can get menu editing working again during the 2.X timeframe instead of waiting until Gnome 3 to do it.
In the meantime there are options, like http://www.realistanew.com/2005/03/18/gnome-menu-editor/ and others.
this java deadlock is a pain, gcj is’nt good enough and sun are’nt GPLing java. java is definietly the way forward its such a beautyful language. i don’t blame sun either for GPLing it they have some valid arguments. what we need is a comprimise of some sort!
it would be trival to have an option to retain the current open dialog for advanced users
If it’s trivial, plan to implement it. Don’t scare people with worlds like “the only choice”.
next time you are on a inexperienced users’s system check out how cluttered and flawed their home directories are no matter what OS they use.
Sure, but you can’t be absolute judge. What is flawed for you can be perfectly OK for that user. This is exactly what I stated before: GNOME developers seems to think “oh,see how are bad these users, let they do things the way I think they should do”. But you’re not my mom, nor my teacher.
Note I don’t think your is a bad idea. I just think the user must have the choice to:
– rely 100% the smart storage you plan
– rely 100% on the old folder system
– rely on both following the needs
Your smart storage beats the current system in many aspects, but has drawbacks nonetheless. I think the best thing is to let the user choose and make clear to the user they can choose. And don’t start the litany “choice confuses users”. Choice has never confused anyone, not at this level anyway. For people confused by choice, there are always defaults: these people won’t touch them. Simple.
You’d have a list of all the files you created. How is that a problem?
The problem is that such a list makes no logical sense, while a directory hierarchy makes sense on every system.
That’s not what GNOME is for. If that’s what you want from your desktop, use fluxbox. GNOME is a desktop environment; it has an intrinsically different purpose from what you’re asking for.
Gnome has features, that is not my major gripe. It has apps, something in pretty much every category. My gripe is PERFORMANCE. GnomeTerminal for example is incredibly lethargic compared to Xterm. Yes I know these apps are doing much more (Pango etc) and I am willing to put up with some loss of performance, but come on guys, this needs work.
That’s not what GNOME is for. If that’s what you want from your desktop, use fluxbox. GNOME is a desktop environment; it has an intrinsically different purpose from what you’re asking for.
Bah humbug. Fluxbox is just a windowmanager, a component. I want a filemanager, media player etc all minimal and fast.
Its been done before geos, amiga, beos.
People have been fooled by ms/apple propaganda. We need efficient tools not a “lifestyle” product.
Nice writeup, but what’s with the epilepsy-inducing “NO” image that popped up in three places on the page? (Using firefox with adblock; no proxy or anything else.)
Given all the trouble with Java and C#, plus the impression that open source never innovates, perhaps now is the time to develop something that OS Community can own itself. No, not python (it’s OO is not the nicest), we should have a modern easy to use language, fast, platform independent, rich runtime and like .net easily linkable to other languages. Plus it should have an easy and intuitive approach to building GUIs (more like VB and Delphi than some of the tedious approaches currently in use).
<<this java deadlock is a pain, gcj is’nt good enough>>
Why isnt GCJ good enough? It can already run:
Eclipse
Azureus (1 line fix)
RSSOwl
Tomcat
…
Classpath already has ~80% of the JDK 1.4 core class libraries.
<< java is definietly the way forward its such a beautyful language.>>
No…just no. Never have I heard someone describe Java as beautiful.
<<No, not python (it’s OO is not the nicest), we should have a modern easy to use language, fast, platform independent, rich runtime […]. Plus it should have an easy and intuitive approach to building GUIs (more like VB and Delphi than some of the tedious approaches currently in use).>>
Ruby?
Nothing new, really, but a nice summery of open issues.
I missed some more fundamental considerations, though: Why should GNOME add another language, for example. Isn’t the binding package sufficient. Why?
Does a user really need F-Spot and the like in the desktop package? Wouldn’t it be the distributions job to care about letting the user install the wanted apps easily?
I also missed some considerations about negative effects: Let’s say, a very succeful application is developed in Ruby that may be needed by the desktop, will Ruby be added, too? When does adding languages stop?
Similarily, the MacOS Style top bar: Should and can all applications change their behaviour? Will third-party developers indeed change it, or will it result in an inconsistent mix for lots of users?
I’m glad that GNOME 3 is still a few years away.
The problem with the gtk+ bindings are that they’re mapping to a procedural api. Python, C#, Java, Ruby, etc…raise the abstraction level certainly, but at the end of the day there are many C’isms to these APIs.
What I would like to see is a specification for gtk+ 3.0 so that all these OO languages can follow.
This specification would allow for variance, but hopefully most language bindings would follow it for the most part.
So basically, we would have two APIs, the straight procedural gtk+ api, and a new OO api specification.
But JVMs can’t run CIL.
What about PHP? I’ve never worked with a programming language like C++ or Java, but could one use PHP for developing GNOME apps? There is that GTK-PHP project going on, perhaps it could be supported more?
>Ruby?
Perhaps, but I think we would have to go for a completely new language otherwise we’ll have continual arguments about which one is better!
Im a GNOME user and I’ll suggest the same I suggest to KDE:
Work in the basics first.
Speed, Redrawing, Memory leaks and known bugs in the GNOME DE.
GNOME is nice already, I like it and I use it, I only want some annoying bugs to be corrected.
There’s no reason to wait around until 3.0 to fix bugs. Work on those is already active and I’d expect 2.12 to fix many such problems within a few months.
Michael
By far the most important thing for the Gnome devs to work on in both the 2.x series and the upcoming 3.0 release will be maximizing speed and minimizing memory usage. The KDE devs have managed to to it with their 3.2, 3.3, and 3.4 releases, and so to can the Gnome devs with upcoming releases.
Gnome’s biggest weakness is it’s memory usage and speed. It’s not monumentally bad, but could definetly use some serious optimization.
Another important thing is to streamline the framework, and make building Gnome easier. Pactrick Volkerding dropped Gnome from Slackware because the framework is too complicated to build (plus there’s the Dropline Gnome project).
Menu editing would be nice, too. 🙂
ok.. I listen to all these posts and rants about memory foot print, choice of languages, this app, and that app when it comes to gnome or kde. But what it really comes down to is no one is thinking outside the box. All these ideas seem to be everyone tweaking the same ideas over and over again and its kinda sad and depressing. We have hundreds to thousands of open source developers who are really, really talentend people – who could make a system so mindblowingly simple, powerful, and elegant … and its going no where it seems. Yes we are making progress, but sometimes I feel like we’re going in reverse. I guess I am just frustrated by the state of computing – we could be so much farther ahead… the technology is here. Whats holding us back?
… sorry … venting … depressed.
The official Menu Editor for GNOME is
http://manny.cluecoder.org/packages/gnome-menu-editor/
Support for it has recently been added to gnome-menus and gnome-panel.
Cheers.
> *IT’S NONE OF YOUR DAMMED BUSINESS WHERE AND HOW PEOPLE
> STORE THEIR PERSONAL DATA*!!!!
Please stop writing in capital letters. Otoh this is indeed a great idea and should be followed. People don’t know anything about files, they are totally confused by the Unix filesystem paradigm and feel lost. People shouldn’t and don’t really want to dive digger than 3-5 folder if all. They want to quickly access their pictures, soundfiles and other stuff and not dig through dozens of files. The more is hidden from the user, the less confused this person gets.
I totally agree with Jay. Please people convince me that I should switch to GNOME from MacOS X, which is already very stable and fast enough, has no annoying bugs and is superior in things like iTunes, iLife, soon-to-come Spotlight. Solve this problem: why I should use gnome.
> why I should use gnome.
You don’t need to. It’s meant for people who can’t run MacOS X because they can’t afford the Hardware/Software.
Add some of the features and speed from gnome1
> Add some of the features and speed from gnome1
What features ?
And you can’t improve speed if you add more features and technology to a plattform.
Since Gnome 3.0 is still probably a good two+ years away.
You can probably expect to see Mono everywhere for 2.14 at the latest. Mono of course would bring Beagle with it (its kind of the whole point).
Expose like features would require either Xgl or Luminocity. I think we can expect to see those around 2.14-2.16 as well.
What’s the current state of the gnome codebase, architecture, whatever? The last time I heard anything about it–admittedly, quite some time ago–I was told that it was in terrible shape.
Can I have an update?
You can probably expect to see Mono everywhere for 2.14 at the latest. Mono of course would bring Beagle with it (its kind of the whole point).
>
>
You can also expect Commodore and Atari to release updated versions of the Amiga and Atari ST computers shortly before the 2005 Xmas shopping season…..
And you can’t improve speed if you add more features and technology to a plattform.
Sure you can. I suppose you’ve never coded before. Look at KDE. Every version after 3.1 has been faster than the previous one, while continually adding features.
You just need to improve the speed of existing features more than the cost of new features.
When the volume is connected, the device is scanned and the appropriate metadata is extracted from the files and inserted into the database of available files, and when a user hits “open” those files are available until the device is disconnected.
This is a bad idea. What if you insert a DVD? It has 4.8 GB of data on it, perhaps 10,000 files. I’m sure as hell not going to want to wait a few minutes for Gnome to extract the metadata out of all of those files before I can access them. Don’t even think about adding a 200GB external drive with a system like that.
Well, considering that Gnome is on a 6 month development cycle, and I know we’re talking about having it included for 2.12 in Dropline (although not certain right now). I’d say its a pretty good bet.
Also, considering that there are new Amigas kicking around out there now. I’d say I’m doing pretty well.
Right on. All of what Apple and MS is doing is convoluting the OS and grinding the system down for a little wow factor. The OS needs to goto back to basics where it is the interface between the apps and the hardware and user. No more and no less.
Keep it simple and lean and get it the f away from my system resouces so they can be devoted to things that count be it gaming or 3D or video production etc.
Least BeOS was onto something. Pity the others haven’t caught up. BeOS is dead, long live Haiku!
I would prefer to see enhancements to the GNOME office applications. Abiword and Gnumeric are really maturing nicely now but for me, DTP is a huge hole. A GTK tool equivalent to Scribus would be great, along with soemthing like KColorChooser (there was such a tool in GNOME 1.x which was removed.)
A GNOME Illustrator tool would be great also.
By the way, what happened to the GNOME Powerpoint-style presentation application? I know there was one but it seems to have progressed very slowly.
A GTK tool equivalent to Scribus would be great
Just use Scribus. I don’t understand why people are so adverse to running applications not done in their DE’s GUI toolkit. So it’s written with QT, is that really such a big deal? Do we need to clone every good application with every GUI toolkit? It’s a gigantic waste of effort.
“A GNOME Illustrator tool would be great also. ”
Inkscape anyone?
http://www.inkscape.org/
” I want a filemanager, media player etc all minimal and fast. ”
You want Rox, xine/xmms/mplayer, and flux/black/openbox/enlightenment/windowmaker.
It’s not so hard to install a few things yourself.
ALT-~ switches windows within an application! You are now officially my hero. Since I got this powerbook, I’ve been wondering why the heck alt-tab doesn’t switch between windows correctly, so this alt-~ thing is really great. Thanks
Re: gnome1 speed and features
By leo (IP: —.gv.shawcable.net) – Posted on 2005-04-12 01:05:08
And you can’t improve speed if you add more features and technology to a plattform.
Sure you can. I suppose you’ve never coded before. Look at KDE. Every version after 3.1 has been faster than the previous one, while continually adding features.
Leo is completely right about this. Gnome 2.0 was faster than the version of KDE at the time. But KDE caught up in a hurry.
In fact, the gnome folks are involved in memory optimization bounties right now.
The speed of gnome seems very, very dependant on the amount of RAM you have so I figure this to be a noble goal.
I am really surprised that this is not a major goal of the project. An interface as sleek and simple as gnome begs to be fast. From the looks you almost expect to be really fast. I use XFCE on a 128MB laptop because gnome is not fast.
XFCE uses X.org (of course duh) and GTK just like gnome and it is lightning fast. Don’t give me the feature creep stuff when samba browsing is better in XFCE and it has a xdg compliant menu editor. Both are things that gnome has to work on.
The sad part is I prefer the mix of mac/pc and even os/2 ideas present in the gnome interface.
Stuff I want from gnome:
1. menu editor (coming gnome 2.12)
2. launch feedback for desktop objects
3. Better network browsing (coming gnome 2.12)
4. Beagle (coming in gnome 2.12) – I prefer Storage actually just like the article writer
5. Speed (some improvements in gnome 2.12 promised) BUT … this really should be a major focus.
Before you start as Havoc proposed into moving to more of a object oriented focus for the desktop instead of apps and changing around paradigms of desktop use, it might be a good idea to round out the functionality and focus on speed optimization in a HUGE way.
But that is all IMHO.
I’m not saying I don’t use it just because it QT, but Gnome could use a good MSPAINT thingy too..
I recently put Ubuntu Hoary on my Dad’s computer, and here are some thoughts regarding Gnome:
– Most of the stuff he did in Windows is easily done in Ubuntu/Gnome – image editing from digital camera pics etc, document editing using Openoffice2 Beta etc. For average users, GIMP is great, so is Openoffice.
– I had to do a fair bit of fiddling around until it reached a fairly user-friendly state for my Dad – but some of that can be blamed on the way I had the PC set up in the first place (he inherited it from me).
– Some of the potentially best stuff is just coming up the pike – I’m thinking of F-stop here.
– Gnome is slow compared to Windows on the same hardware – I’m using a 600MHz Celeron II on an Abit BE6 MB with 384MB RAM.
All in all I’m very optimistic – but I’d concur with those who suggest the priority of Gnome coders should be optimisation for speed and memory usage. At work I use Gnome whereas most of my colleages use KDE, and although I feel I have the better desktop, I can’t deny theirs is faster and more responsive.
I still think that Gnome – as it currently stands – shows the way towards a truely user-friendly Linux desktop that makes migration from Windows easy.
I hope that Gtk+ 3.0 is a C++ widget library, and it should be more rich feature widgets libaray, it should include the docking libary, libglade and move all the widgets from gnome library to it.
and I don’t think Java, mono or gcj should become the mainline language for GNOME 3.0, the main language for GNOME 3 should be C++, because C++ is good for desktop applications.
The Fedora crowd think it is a good enough reason not to include Scribus onthe CDs by default for the personal desktop or the workstation profiles. It is only available on the repositories.
I agree though; they should not be so partisan.
Most of that is happening already (libraries are being deprecated, more is going into gtk). I don’t think you’ll see gtk go C++ though. For one thing, you would lose all the good bindings you get right now, including the excellent C++ bindings.
Am I really the only one who has read the wiki entry on the subject? http://live.gnome.org/ThreePointZero
So OSNews reader…do you use Ubuntu? Do you still have Warty? Are you scared to upgrade because you used the backported packages? Here is how to upgrade:
These are the directions to upgrade to Hoary on all computers with a
broadband connection.
First, edit your sources.list file. Put this line in the terminal:
sudo gedit /etc/apt/sources.list
Then when gedit comes up, replace Warty everywhere you see with Hoary. Comment out any un official repo’s for now (especially backports)
save gedit for sources.list and exit.
then in the console put this:
sudo apt-get update
then when its done, put in this:
sudo apt-get dist-upgrade
this is the big one. The full update. It should work, but it will take
a while to download the new Hoary files and install them. Expect to
download 500+ megs each time. It takes 26 minutes to do this on my
300mhz laptop, then than 10min on my P4.
NOTE: It is not advisable to do other things with the computer once
all the packages are finished downloading and they start installing
(after sudo apt-get dist-upgrade command) like web browsing and music
and such. The computer is replacing those programs, and if you use
them they will crash and hang. just stick to the console after all the
Hoary packages are downloaded.
then when its done, put in this:
sudo apt-get install ubuntu-desktop
If it says something like “already is current version”, then you are
upgraded. Reboot into Hoary.
I know you will kill me, but I don’t care 🙂
I love the clean look of spatial, is faster and clean, but I don’t like to see all window taking different places and sizes. I can’t remember all window positionn and some time, when I reduce the screen resolution, the window goes out of screen.
Gnome developers put two good options, “aways browse” and “every opens new window”, so I need another option, “don’t remember size and position”.
I don’t think this can be a retrocession…
Cya…
Apache is able to make an agreement with Sun to have free access to their TCK to certify their open source implementation of a J2EE server (Geronimo).
BTW, Sun is becoming more open with their licenses. I believe they are planning to license J2SE 6.0 (Mustang) under 3 different simple licenses. Although none of them are open source licenses, they are in deed more open (for example, I beleive one of the license will allow companies to made modifications to the J2SE as long as they keep it to themselves)
You’d have a list of all the files you created. How is that a problem?
The problem is that such a list makes no logical sense, while a directory hierarchy makes sense on every system.
How does a list like this not make sense
Music/
MusicFile1.mp3 ————- Meta-data
MusicFile2.mp3 ————- Meta-data
Pictures/
Picture1.jpg ——————- Meta-data
Picture2.jpg ——————- Meta-data
Documents/
Document1.doc ————- Meta-data
Document2.doc ————- Meta-data
I’m pretty sure a list like that makes much more sense to 90% of users than a folder full of folders full of folders.
Supporting multiple languages is a good idea.
Scripting languages such as python and ruby are good for small tools, plugins etc.
Managed languages are faster to develop compared to C++/C and should be used by the majority of apps.
C/C++ are probably more appropriate for the core stuff.
However, this would entail a lot of glue code and wrapper code. I think Mono offers the best solution. Again, I don’t think MS suing Novell will make Mono obsolete. They have made statements allowing free implementations of the C# language, the CLR etc. Suing Novell, at the worst, will make them remove the windows API compatibility layer, but the gnome APIs (GTK# etc.) will be intact.
Basically, Nautilus will determine using statistics whether or not you want a particular folder to stay open after you opened one of its sub folders. By default, they should keep the /, ~ and Computer folder open. This should reduce the window clutter and make the spatial view more effective.
is Obdev’s amazing LaunchBar <http://www.obdev.com/> – the original.
THIS is what Gnome needs – a way to immediately launch apps, find files, navigate the box, & cetera, without having to either reach for the mouse or launch the effing Terminal.
Without such a tool, any activity on Linux feels unspeakably cumbersome and, well, plain old-fashioned.
The Gnome equivalent would be a small app of huge consequence.
“The Fedora crowd think it is a good enough reason not to include Scribus onthe CDs by default for the personal desktop or the workstation profiles. It is only available on the repositories.
I agree though; they should not be so partisan.”
Being partisan has nothing to do with it. QT and all of KDE is already in Fedora core. The claim that Fedora doesnt include scribus is false since Fedora extras is very much part of the Fedora project. For FC4, fedora extras will be enabled by default and you can install it right away with “yum install scribus”
“Red Hat hasent written 1 desktop app in Java. They are a bunch of hypocrites. They push Java but they still arent using it themeselves.
”
The real reason why Red Hat has written GNOME applications using the Java bindings is the relative immaturity of both the bindings and GCJ so far and the fact the developers are already supporting PYGTK extensively for a long time. The free Java development stack is almost entirely supported by Red Hat and will be used to ship natively compiled Eclipse, Tomcat etc in Fedora Core 4. When its gains better momentum, desktop development based on Java can happen
“Being partisan has nothing to do with it. QT and all of KDE is already in Fedora core. The claim that Fedora doesnt include scribus is false since Fedora extras is very much part of the Fedora project.”
Rubbish. Not everyone has a fast network connection, even in this day and age. If Fedora were serious about the various installation classes, they would see the necessity about including a DTP application on the actual install CD-ROMs. At the moment, they are dropping applications such as KOffice from the CDs in the name of saving disk space. The rationale for their choices is to eliminate duplicate functionality. However, there is no DTP package at all on the install disks.
Whether you care to admit it or not, moving something to the Extras makes it a second-class citizen.
”
Rubbish. Not everyone has a fast network connection, even in this day and age. If Fedora were serious about the various installation classes, they would see the necessity about including a DTP application on the actual install CD-ROMs. At the moment, they are dropping applications such as KOffice from the CDs in the name of saving disk space. The rationale for their choices is to eliminate duplicate functionality. However, there is no DTP package at all on the install disks. ”
koffice *is* planned to dropped dropped for disk space. not in the “name of it”
fedora extras can be made into ISO images. Work is already being done but if you want it to be in core, propose that to the fedora devel list. complaining in osnews is not going to help really
“Whether you care to admit it or not, moving something to the Extras makes it a second-class citizen.”
not when fedora extras is enabled by default and when anaconda support and ISO images are planned
“fedora extras can be made into ISO images. Work is already being done but if you want it to be in core, propose that to the fedora devel list. complaining in osnews is not going to help really”
I’ve proposed it on Bugzilla.
If, as you say, the Fedora team are planning to release ISOs of the Extras, that will make a difference for me, though. I would be happy with that. I still think that applications like Scribus should be included as part of the Workstation or Desktop default installations though.
“I’ve proposed it on Bugzilla. ”
I believe proposing stuff that should be discussed prior to inclusion is better done through the development mailing list rather than bugzilla.
Its pretty late in the development cycle for FC4 to consider this but if FC4 core doesnt fall to 1 CD or something like that scribus can definitely be considered for inclusion in FC5 or above
How does a list like this not make sense
Music/
MusicFile1.mp3 ————- Meta-data
MusicFile2.mp3 ————- Meta-data
Pictures/
Picture1.jpg ——————- Meta-data
Picture2.jpg ——————- Meta-data
Documents/
Document1.doc ————- Meta-data
Document2.doc ————- Meta-data
I’m pretty sure a list like that makes much more sense to 90% of users than a folder full of folders full of folders.
I’ll tell you why it does not make sense.
First: not all system are/will be able to read metadata instantly, so my backup or what would be restricted to be usable on GNOME systems only.
Second: You don’t seem to understand standard metadata of a file can’t store anything. For example, I divide mp3’s on my system among the ones I already burned on CD and the ones I have to burn, plus the ones I know I will never burn on a CD. How do I do such a distinction with mp3 metadata? Or do you want to add a whole and fully extensible metadata layer? And how do you do it on top of the current filesystem?
(Have you ever browsed shared folders of people on p2p systems,for example? You would be amazed to see how many of these filesystems are elegant and rational and separate things that standard metadata can’t separate)
Third: Such a system makes sense if we rely on GNOME only. The one time the user will have (God we hope not!) to go to the shell or something similar because he’s experiencing some problem, he will be at least scared, after months/years spent with a system that hides the filesystem from him. We’re going to create even more clueless users.
Fourth: If such a list does not make sense to me, it means there are users that don’t feel sense in that list. Take care of these users too and let them use the standard folder mechanism.
You’re tied down to your old habits. Probably you have your music/ in your home folder and a tree of subdirs for every artist/album and so on. I was used to it too, when using ugly players like XMMS. Now I’m using amarok, my music/ folder is still organized in subdirs as before but I no more care about it, I use the collection browser (created on IDtag basis) to browse through my music, and I don’t care about dirs and filenames.
Remember that filenames and dirctories structures are a *view* on the real HD bits (01010011101011 and so on), this metadata view is simply a layer over, and you have only to get used to it. If you don’t want, create your personal DE by using different simple tools together, but don’t be a retrograde praying for “don’t touch my habits”.
Wish lists and requests are great, but they are useless if a project isn’t established to accomplish them. Unfortunately, the side effect of Linux’ rising popularity on the desktop is that it is beginning to be littered with more sayers and less doers.
I hope GNOME3 provides an efficient mechanism that will enable users contribute to the project. I hope the artificial barrier between the GNOME environment and GNOME applications is completely eliminated. I hope GNOME3 inspires the next generation of creative and talented contributors (i.e writers, artists, interface designers, promoters, testers, bug hunters and hackers). I hope GNOME3 provides tools to make these contributors exponentially productive.
Finally, I wish people started contributing more to free software projects and whined less about them. The only way you can make anything better is by contributing, not whining.
While I do like the “usabilty” of the gnome desktop.
What I do not like about gnome are the concepts of it.
Cleaner is better:
The libraries are messy. Not clean, nor mean. The hierarchy is a non sense.
The issue does not lies on the programming language. It never has been and never will be. It is all about the architecture. And the gnome architecture is … well you should know it. There is no architecture.
This lead to a non avoidable stack of issues on dependencies, interoperability and maintenability fronts.
The HIG is one way to fix things up. They ask you to comply to it which sort out the interoperabily and the user consistency problem but still we have the dependencies and maintenability ones.
I’d like to have first a clear definition of the architecture then a roadmap of it.
Less is better:
The gnome meta-data problem is not a coding problem it is an architectural one. Trying to reimplement it in the userland while it exists already in nearly all filesystems is a non-sense. Just build up a interface on it. It shall be faster with less bug and with more features. Also, It shall remind you that a desktop metaphor is *only* a GUI that expose to the user and *simplify* the lower layers concepts. It is not a place to create OS level features!!! (especially if the OS already have those) The features should not be implemented there! I took this exemple because it is very easy to understand.
But what gnomes lack the most is a devoted and skilled recognized by all leader. It does not have to be a single person it can be a committee similar to the JCP one. I know there is some lookalike stuff out there but it is not even close to JCP it looks like a draft (for years now)…
IMHO voila, c’est tout.
Heya,
anyone know if there is any progress on a clipboard that keeps content available when you close an application? I suppose preferably a standard accross the toolkits/desktops?
I haven’t really experienced this as a problem the most of the time since my application is still open when I paste, but maybe it would be handy …
Michel
Ahhh!,
lets finish the Parrot VM and make part of Gnome.
Ruby, Python, Perl6, Tcl, … Dynamic languages are the future.
>a clipboard that keeps content available when you close an application?
Klipper, and it has done it for ages. I have mine set up to save content on exit, so also I have everything avalibel between sessions.
The only thing I need to see in gnome to make me happy with it is adding the ability to manually type the path in the file save/open dialog. Who’s the flipping retard that thought it would be fun to make me click my way through ump-teen folders when I could just type the damn thing in?
The only thing I need to see in gnome to make me happy with it is adding the ability to manually type the path in the file save/open dialog. Who’s the flipping retard that thought it would be fun to make me click my way through ump-teen folders when I could just type the damn thing in?
press Ctrl+L to have the address dialog.
Apple didn’t develop Spotlight to be the Heuristic File finder for lost files.
Apple does prefer people to actually distinguish between case and where they put their files. They just have more failsafes when novice users think they can just dump files anywhere they want and then complain they can’t find their files. The novice user in OS X doesn’t know how to use sudo let alone do much more than assume the Installer.app puts applications in the /Applications folder and that their /Documents folder is actually /home/user/Documents.
Spotlight is interesting when one has large amounts of documents and you are looking for specific content you can’t recall is in one of them and you need that information quickly. I have hundreds of Postscript, PDF, LaTeX, etc. I want to recall a specific piece of writing where I read a certain passage, so I use Spotlight to dig for this info.
How this relates to Apple having designed their filesystem so that the user no longer has control of where files are located or their naming conventions is a stretch.
By the way. Tiger is out April 29th. http://www.apple.com has a lovely blurb on it.
For my Linux needs I’m quite happy the direction KDE is headed.
… So why not gnustep ? 🙂
Of course, the look’n’fell must be improved, but the framework is quite nice!
I’m not (so much) tied down to my habits. In fact I think a fully implemented metadata-powered smart-storage would be absolutely lovely.
But such a system would require consistency at all levels. It would have to be fully integrated in the filesystem and in the filesystem tools (I heard Reiser4 allows something of this kind). I would have to be able to tag every file with arbitrary tags (for example “pop music/burned/Janice likes it” or “to read/for article on JMB”). This way (and by using metadata-aware file managers, that would answer to search-engine-like queries and would create virtual folders, instead than relying on absolute paths) the system would be perfect, and I would throw away the current hierarchy without complaints. Too bad it would require to build a new Linux kernel (to say the least,GNOME runs on Solaris and *BSDs too), improve filesystems, rewrite/deeply patch filesystem tools, and so on.
The sad truth is that what GNOME (or amaroK, or iTunes on other platforms) wants to build is a crappy patch on filesystems that are not aware of metadata tags. They can only be aware of metadata in the files themselves. So you have to explain me: can amaroK know which mp3’s you have burned and which not? Which your girlfriend like and which not? Or does your word processor/PDF reader/browser tags documents you have to read tomorrow and which not? I think not.
Don’t get me wrong: Such a crappy patch would always be better than nothing. But it can’t substitute the flexibility the decades-old folder system achieves with only a little price for usability. I think GNOME (or other desktop enviroments) should simply offer to the user both worlds, until a full shift of paradigm will be achieved in the OS arena (and I hope free software will be leader on this…)
OS/2 — the failure that it was — did one thing right oh so many years ago;
* The file browser and all other associated applications had exactly the same view on the underlying data as the command line.
With OS/2, you could do everything in either CLI or GUI. No surprises would occur; if it worked in one, it would be available in the other. All properties were manipulatable in both.
While I am not suggesting that the GUI be limited to the structure of the file system, the Unix-way of having everything as a file is a great advantage even over much of what OS/2 offered. Why not take that advantage and expand on it on both the GUI and CLI sides?
Example: If you can browse in the file manager to a virtual resource, have that virtual resource available at the same place using the same syntax in the shell.
* Trash – Appears on the desktop, not in ~/Desktop.
* Same for any other icon such as Computer and user’s Home.
No window managers (including KDE) get this right. (KDE, BTW, is a bigger offender since it has many wonderful IO Slaves that are not available to view from the shell; you can’t cp from or cd into a virtual ogg or mp3 directory on an audio CD, for example.)
>KDE, BTW, is a bigger offender since it has many wonderful IO Slaves
>that are not available to view from the shell.
Are you sure? Have you ever tried KIO-Fuse?
From where I sit, deciding between Mono or Java is like deciding between being eaten by alligators or by lions. Sure, each person is going to have a preference, but the end result is the same… You become fodder, and after some processing, excrement.
Both Java and Mono are, at their core, proprietary technologies developed by companies with selfish interests. I’m certainly not against making money, or enhancing return on investment for the shareholders, or however you care to characterize the driving goal of a publicly traded entity — I just hope that most people in the “open source community” can see that hitching their wagon to such a beast might not result in the smoothest ride.
The idea that if Sun/Microsoft goes “evil” (or more evil), we can switch to Microsoft/Sun is pretty laughable. You’re still in the same kind of boat, but now you don’t have another ship to jump to. Great thinking!
There are many languages with rich libraries, proven track records, and open standards to pick from. There’s no need to run to the company-sponsored solutions other than to look like you’re doing the trendy thing. Personally, I really dig Python, partly because it’s governed by a non-profit foundation, but Ruby has many things going for it. And when you need low-level speed, what’s wrong with C, or even C++?
Choosing between Mono and Java is not a real choice, it’s just different pages from the same book.
Nick
Just for example the smb browsing of win2K server (who failed miserabily in nautilus once you get inside a directory) and work without nothing to say in xffm (xfce file manager)…
or making gnome-terminal as fast as multi-gnome-terminal would be great too…….
gnome is still used by the power of its artworks designer…
press Ctrl+L to have the address dialog.
yeah that’s not much better. 1) you have to type the path AND file name, you can’t just type in a path so you can see what files are available. 2) It’s trying to guess what directory you want and makes it impossible to type in the path. Not that it’s a bad feature, but they should list that in a drop down similar to how KDE does rather than just filling it in while you’re typing.
Oh, I can see some problems with the whole content-derived metadata issues from the start:
1: It assumes that people who are unwilling to use folders, are willing to enter metadata. This isn’t a big deal when you are dealing with a few hundred files, but when you start looking at a few thousand files, things get out of hand pretty darn quick. I work with people who dump everything on the desktop and use google desktop search to find stuff. They still waste lots of time trying to find the exact keywords. Folder heirarchies are a form of metadata.
2: It breaks when you start adding on new filetypes with different metadata formats.
3: Indexing new collections can suck down huge volumes of processor time.
By all means, full-text indexing is part of a solution to the problem of managing collections of data. But I don’t see that it will be a panacea.
I’m running Fedora 4 Test 2 and I cannot find the cron frontend that was supposed to be in GNOME 2.10. Was this not ready in time or have Fedora omitted it as part of the install or is it that I just don’t know where to look?