A few weeks ago we published an article titled "The Great Mac OS X 10.4 Wish List", detailing a few personal wishes for the next version of OSX. Later I learned that quite a few Apple engineers read the article and so it felt good that the time spent writing the article was not just a voice in the void. A reader emailed me a few days ago asking me to do the same for other OSes and DEs. So here is my personal wish-list for a future version of Gnome. Please tell us about your own Gnome wish list in the comment section provided.
Post a Comment
Excellent article! Very comprehensive and any desktop will need all that you suggested in order to succeed. All the pieces of the puzzle are now in place, it just has to be put together.
Maybe not most the important item from the list, but something that not many others have suggested is Quick Lounge's funcitonality. I would definitely love to see it in the panel!
I really miss network transparancy like efs from Emacs or KDE. I can open a file on FTP server and edit it transparently on both Emacs and kate or gideon, but this is lacking on Gnome.
Also while I am at it, I could not belive that I had to put my FTP account password on the Nautilus URL bar for everybody to see in order to open a FTP site !!!
Can we expecta KDE wishlist soon? I think that however, several things on the GNOME wishlist would apply to KDE. Please make one, you usually know very well what areas are lacking in UI especially.
As much as I love gaim, it lacks some important features:
first of all: file transfers. As of now - they don't exist at all! Why not use some code from the licq/amsn apps, and just implement it under the gaim UI?
other things, which I consider minor, but still worth mentioning: contacts in ICQ, and Images in MSN.
On a totally different term, I think GNOME should also consider making more microsoft things work - I don't mean Internet Explorer, but it'd be nice if for instance one could use GnomeMeeting to connect to users of Windows Messenger, and so on. Even though the codecs are proprietary, people that own them should be allowed to use them in an easy way.
oh, and BUGS BUGS BUGS!
I think a Gnome Dialer for modems is really missing. Even if not many of us or almost everybody hasn´t a modem connection to the internet, such a feature is needed just in case of ...
Pretty good article Eugenia -- interesting points raised.
While GNOME's work on their HIG and related stuff has been good to see, they've missed some really elementary features that have been present in other desktops for years.
For instance, alphabetical keybindings on the main menu; this means you pop up the menu, then hit "P" for Programs, "I" for internet and "E" for Epiphany etc. It's an excellent time-saver for power users, and saves messing around with the mouse. Currently, you have to use the cursor keys in GNOME, or move your hands from the keyboard to the mouse, which takes up so much more time.
So, IceWM has this. KDE has this. Heck, Windows 95 had this EIGHT years ago! For proper usability, a decent environment for power-users, this is essential. Working efficiently and productively means good keybindings.
(Oh, and being able to move windows above the top of the screen would be great too on lower resolutions. GNOME is unusable with many apps without this option.)
M
Yes, you are right, a modem dialer is really needed. I didn't think of it because I am on broadband, but you are absolutely right it is needed. Thx.
Viewports...what are those? I've used linux since 1994 and have heard of viewports, but never knew what they were for.
I have to agree that the beta GTK File Selector is bad; it's almost as though it was made as a joke. Who in their right makes "Frobnicate the file" an option?? Apart from the fact that it is a very uncommon word that means "tweak", the option itself doesn't make any sense to me; what exactly does it do? I think I see where they are going with the "bookmarks" idea, but it seems really out of place. Lastly, I hope that they don't have a single arrow and all of that blank space at the top in the final version.
Re: What gnome does better than KDE.. and Eugenia's Wishlist
You can have a window open half on one workspace and half on the other workspace. Apparently pretty handy for developers, but bad for overall casual user usability.
Excellent article, you raise many agreeable points.
Aside from a couple things that both kde _and_ gnome could
use improvements on, and aside from a couple specific apps
you mentioned - it sounds like you want gnome to be more
like kde.
I'm guessing that if kde finally got around to creating a
theme or two and an icon-set or two that looked a little
more "gnomeish", a _lot_ of people would become happy
converts. One other thing that I'm pretty certain of, is
that much of the gnome-ism comes simply from the fact that
it appears to have stronger backing from large companies -
so people think it's going to some how be/become more
"legit".
I'm using KDE 3.2(beta2/cvs) and it honestly rocks. I was
using gnome quite extensively since 2.4 was released, but
it just never satisfied, for alot reasons. KDE merely
fails to satisfy on a few reasons.
There certainly is more than a couple things that kde
definitely can improve on (and it is), but with the
release of 3.2, I really think that it just leaped miles
ahead gnome, gnome 2.6 will still be behind. And just wait
in a couple months when all the kde PIM features that
couldn't get released on time w/ the official 3.2 dist are
delivered.
One thing KDE could use is a distro that concentrates
entirely on pure KDE, one that provides an out-of-the-box
default kde experience which attends to some of the
vanilla kde issues.
Hopefully you will do another article for a KDE Wish List,
if so, please be sure to use the latest 3.2 available to
base your wishes.
Cheers! And I hope all your gnome wishes come true!
Eugenia,
Well, since you've now done a wish list for GNOME, can you please do a wish list for KDE...?
Don't see how this is off topic!
Michael Lauzon, Founder
The Quill Society
http://www.quillsociety.org/
mlauzon@quillsociety.org
> it sounds like you want gnome to be more like kde.
Definately not. I want the developer tools and some of the architecture of KDE, but definately not the UI and its bloat.
>Hopefully you will do another article for a KDE Wish List,
As I replied earlier, I am preparing a KDE 3.2b2 preview article.
I can accomplish pretty much everything I want to using Gnome 2.4.1 and whilst I would like new features to play with an enhance my experience, the one key thing they should focus on that will make everybody happy is efficiency.
Mozilla starts up is less than 2 seconds, if that. Epiphany can take up to 5 or 6 seconds yet it is supposed to be native, leaner, and meaner.
Many of the smaller applications, like the calculator or most of the games, take more than 5 seconds to start when they should be nearly instantaneous.
There needs to be a concerted effort on the speed and efficiency of Gtk2 and the apps that use it. Then people will have a snappier, quicker desktop and implicitly a better Gnome experience.
This is just the pit of an iceberg that Eugenia has mentioned here. There are even more trivial things such as copying stuff from FTP to your harddisk that doesn't work properly. Try connecting to ftp://ftp.gnome.org and copy a subdir with mixed entries (files and directories) to your system and you get something like (copying file 87 of 12) and wonder why it's copying more files than existing. Or if you copy large dirs which are for example 15mb then you will only get 6mb of it and the rest will be copied as 0byte files.
I am not a friend of adding all these bindings to GNOME, on the one hand it offers the possibility for authors to write programs in the language they prefer but on the otherhand it makes things more complex. She has mentioned that things as GTK are really slow but please thing in the future, as soon as you add more libraries, more bindings, more languages into the middle of GNOME as soon you make the general operation become slower (noticable during execution and startup time).
A bunch of things inside GNOME can be fixed or cleaned up but what benefits does fixing stuff and cleaning up things give one when the supplied framework of the broken thing is not optimal in first case. It's like fixing a bit here, a bit there and something on another place but it's no optimal solution. A better solution would be to to trash that framework and go for a better one.
Nautilus is an example itself here - it feels nothing complete but nothing half. Same for Evolution which is NOT following the GNOME HIG here (which raises the question what purpose the HIG has at the end) the Toolbar (to give an example) is still not following the Menu&Toolbar preferences system (Text only, Icons only, Text besides Icons, Text below Icons) and things like this.
Eugenia brought up the user visible/annoying problems which she detected over the months using GNOME, the real technical problems can only be seen by someone who knows the underlaying framework of GNOME.
E.g. it's hard comparing Anjuta/Glade with KDevelop for example because to get an application with the quality of KDevelop you first need to screw (technically/engeneering) at the framework.
> >> it sounds like you want gnome to be more like kde.
>Definately not. I want the developer tools and some of the architecture of KDE, but definately not the UI and its bloat.
Is this a result of of KDE or Qt?
Both. Wait for my article in 1-2 days.
Please let's stay focused on the article here, which is a Gnome article, not a KDE one.
Very nice article, Eugenia. I think you are right about a new cd/dvd recorg program for gnome, since it must be integrated with the file manager. But in general, I don't feel there is a need to have apps made with a specific toolkit. There is a need for more standards and integration.
BTW, did you check Craig Drummond excellent work on GtkQt?
http://www.kde-look.org/content/show.php?content=6954
It gives you theme integration between desktops: fonts and themes mathing, bluecurve, keramik, and others. Unfortunately, it only works from the kde panel to gtk, and not the other way around. It is a hack, not a standard. But, hey, in this sense, bluecurve is a hack too (making different stuff look the same).
Eugenia, I think you still could influence KDE 3.2 usability. Some changes that are considered usability bigfixes can still be accepted. So please, give us a well argumented and carefull analysis of KDE 3.2 beta2. (As you usually do.)
Cheers,
Carlos Woelz
I made the switch from being a big KDE fan to being a bigger Gnome fan with Slackware 9.0. It was there, I tried it out, and I feel like Gnome is much more slick and professional. (to give an idea of where I'm coming from, I'm a huge OSX fan) The problem is that KDE is in a lot of ways more mature. There are lots more default apps, and many more tools. Gnome is picking these parts up at a fairly rapid pace, and one reason I prefer Gnome is I feel like these apps are of better quality.
At any rate, one of the biggest problems I have with Gnome is the menu. Editing it can be very cumbersome.
I'm also really interested in seeing Totem and Rhythmbox integrated into Gnome. KDE does a much better job of out of the box click on the app and it runs.
And as mentioned in the article, Samba browsing is just broken. Badly. Sometimes I can put in a user and pass on a freely shared directory, and it might let me in. Sometimes it just says it can't find it. And if I click on an mp3 on a remote drive to play it, that never will happen. These issues really need to get fixed. People are relying on these network transparency features more and more, and they are non-existant in Gnome.
In terms of the included apps comparision with KDE, the biggest hole is in the control panel. I really like the mentioned idea of making it similar to OSX's control panel, and it needs to control more. Sometimes when X won't behave I just hop into KDE and use the cumbersome control center to fix it.
Great artilce Eugenia, and I hope they pick up some of your ideas!
One thing I like about Windows that I haven't seen in any distro is that the close button can be clicked on when your mouse cursor is in the absolute top right hand corner of the screen. I like the ability to haphazardly move my mouse there with a flick of the wrist, click the left button, and the window is gone. In all the distros I have tried (RedHat, Slackware etc), there is a gap of a few pixels which means you actually have to point at the button, rather than being able to just flick your wrist in its general direction.
I don't really know much about this sort of thing - is this something to do with the theme, the window manager or perhaps the DE? Would I be able to modify an existing theme to enable me to do this?
Matt
This depends on the theme, but I do agree that the default theme should have this feature. It might not look as nice as another design, but it can be practical.
Same goes for the Gnome menu button.
>> it sounds like you want gnome to be more like kde.
> Definately not. I want the developer tools and some of
> the architecture of KDE, but definately not the UI and
> its bloat.
[note: the following is _not_ in any way a troll]
( and I just read in a newer post that you request we keep
the discussion to gnome... well, I'm gonna send this
anyhow 'cuz I already wrote it; guess you'll mod it down )
Lets break your statement down a bit:
1 - "... but definately not the UI and its bloat"
A truly arguable assertment... I have to guess that when
you say "bloat", you're talking the backed architecture
and libs - bloated as in, there's _lots_ of it, not
bloated as in slow and sluggish. I can confirm with my
own direct experiences using both extensively, that KDE
is defintely _not_ slower than gnome, and in fact often
seems faster - *especialy* when built from source, and
after being prelinked. So let's be fair and say that
speedwise, they're comparable.
So if you mean "bloat", as in "tons of architecture,
libs, and dependencies - then I'm curious in knowing
which parts/components of kde you find expendable? As
quoted from one of the berlin/fresco guys, on the
subject of using corba - something along the lines of:
"it's not bloat if you're using all the features".
You also mention the UI - you don't like it? The look?
Something a mere theme and a new iconset could rectify?
2 - "I want the developer tools and some of the
architecture of KDE"
Hmm... well, I know that you must understand that the
reason why that architecture, those tools, and those
apps are even available, let alone that they're so far
advanced and extensible and feature-rich, is _strictly_
due to all that "bloat" you have issue with. This does
not compute.
What I can see, is that it would be far easier, quicker
and realistic to solve a couple of kde's issues, that it
could ever possibly be to get these things you like about
kde into gnome.
Here's what kde could use to address your concerns:
A focused distribution that takes the standard vanilla KDE
and simply delivers it in a more asthetic and approachable
package. This would kill the issue people have with the
fact that they often need to spend alot of time getting
kde to look/act the way they want before they can start
working.
A new theme and icon set - something more "professional",
or "sharp". I would love to see this myself - I have a
really hard time with kde's general "fluffy"/"colorfull"
look and feel. And as we can all plainly see - looks are
everything. I bet if kde stayed _exactly_ the way it was,
but looked just like gnome... there'd be *alot* more kde
users...
Finally, to address your "bloat" problem, kde could use a
much better and refined/configurable build/install routine
that allows the user/distro to control exactly which bits
get installed and which dont. Though, initialy, the
focused distro I mentioned first could take care of this
easy enough.
Of course, continuing the effort toward the usability
guidlines are also important... but thats already happening.
These few things mentioned above are quite simple and
totally in the realm of possibility within a limited time
frame. To get gnome to change in much of the manner
mentioned in your wish list would take _much_ more effort
and time... at least to do right.
Cheers
Dang, talk about not reading it right.
Forget about it.
Ignore my comment.
Boy, is it embarrassing....
Anyway, my wishlist
- New file dialogs. Can't stand the old ones.
- One unified HTML component instead of 2 (or 3 when Evolution is put in the mix), like KHTML
- A bit far off, but, handwritting recogniction, for Tablet users....
- Something like iMovie, iPhoto, iTunes. I still mention iTunes cause Rhtymnbox acts more like WinAMP 1.0
She did not mean architectural bloat, she meant UI (User Interface) bloat as in too many often irrelevant options or poor UI design in general. To some extent I agree that KDE has a few useless options for 85% of the people and the interface is sometimes too cluttered, especially the toolbars and contextmenus. Thankfully, KDE developers are aware of the problems and are working on them.
Well I must admit that your reply does make a lot of sense. Looking at a neutral point of view then I do share all your points here. Unfortunately it's not always easy to make comparison between two Desktop Environments while you are not allowed to at least do the comparison in a sane manner. And to get my word on it. It's no troll it's simply a fact. Sure both Desktop Environments needs to improve. The question is which one has the easier way to go.
Good article.
I agreed with the comment about window sizing and moving. That one thing prevents me from using Gnome at present. I like the way it looks, but I CANNOT use it if some of the windows are going to be larger than the screen with no way to move it around to get to the buttons (as with ALT-mouse in KDE).
This isn't a biggie, but here's another annoyance: I really, really wish Gnome would let you set different background images for the different desktops. Not everyone would prefer that, I know, but it should be an option. People like me who power-use different desktops for different tasks LIKE being able to set up each differently.
Finally, I agree that Gnome needs a dialer. But those who say that most people don't use dialup are mistaken; it depends on the region!! In some parts of the world (even here in the USA), most people -- the VAST majority -- still use a dialup PPP connection to get onto the internet. Many people CAN'T get DSL in the USA, where there are large, rural regions too far from a local exchange.
> She did not mean architectural bloat, she meant UI (User Interface) bloat
Ah, I see... I read that wierd, thought her 'its' was referring
to KDE, not the GUI. My mistake.
And you both are right, there is too much unused buttons,
and menu items and whatnot that are an annoyance to most
people, while only being utilized by a small amount of
folks.
Which kinda brings back my original point... it's actualy
fairly easy to get kde to look more like gnome, on a UI
level - but very dificult at this point in gnome's life
cycle to get those features that make kde so advanced and
impressive.
nice nice, very nice.
On the top of my wish list:
Give me back the rounded corners! Please!!
Nautilus authentication is a must-have. And I wish the nautilus-cd-burner would be improved to be able to burn multi-session CD's, and all that stuff.
Victor.
bit hard to distinguish between a linux-distro (in my case, fc1) and gnome as such, but anyways:
1.(by a wide margin) being able to stream content from windows-pcs (mp3, divx etc.) just as easy i can with when booted into my windows xp. smb-integration doesn't get better, it even gets worse (yes, i know about linneighbourhood)!
2. a fully functional wmp-equivalent (_no_, xine, totem and mplayer aren't up to the task!). but it seems that helixplayer will take care of that (a shame that the open-source-developers seem not to be able to deliver a complete solution since years).
3. cleaning-up and consolidation of configuration-menues (much too shattered)
4. fasterloading apps
5. the nice svg-themes installed by default (as well as reg. fedora, a more elegant bluecurve!)
6. "align after type" for desktop. also, the font-size of the desktop-icons seems to be bigger than the ones within nautilus
7. being able to choose icons of different themes for all instances of the same filetype (folders, .pdf. .doc etc.)
8. better printing (especially in color), but this is likely up to the printer-vendors
that are my major points...i also like the ideas of the gnome bounty-hunt-in general, more integration of all kind!
"I think a Gnome Dialer for modems is really missing."
Actually, there is a simple dialer for Gnome that comes in the form of an applet called "Modem Lights" I use it all the time to dial in to my ISP.
Eugenia, the file selector that you show there its not the one that is being developed in the new GTK version.
That, AFAIK, was an older proposal... but I saw recently(cant remember where, sorry 
) a screenshot of the gtk file selector and it is a LOT different.
Looks like the anjuta 1.2.0 file selector. I'll try to find a link to it..
No, the fileselector that is there is the current one. The one you are reffering to is a third party GTK+ hack that is not part of the codeline.
forgot:
- automount and -unmount of removable media, like cds, floppys, usb-sticks with a rep. symbol popping-up on desktop
- combining the small volume-fader and the big volume-menue, means "one click on icon = main volume - doubleclick = big volume menue
Can I turn off the animation when minimizing a window in GNOME? This is GNOME 2.2 that comes with RH9.
I have been looking for a way to do this for a long time.
Thanks !
I totally agree with you about the lack of dev tools. There will be mon and C# and GTK# and pygnome and gtkmm blah, blah, but no integrated, easy to get setup solution to actually create applications.
Why didn't Borland use GTK+, etc in Kylix?
I think when people talk about bloat and KDE, they're talking about a general problem that plagues any free software desktop. Though KDE seems to be more prone to do so. It's called "reinventing the wheel". The KDE people frequently seem to feel it necessary to have a 'K' version of a game. Even though the only difference are the menu dialogs many times. Many games already had a gnome version, etc. But instead of trying to come up with a standard set of applications that could be maintained easier, they instead decide to maintain their own 'K' version all in the name of "it uses Qt, therefore it must be cool!".
Seriously. More and more people are wanting their favorite applications to integrate tightly with both desktops, but at the moment that's near impossible. What's a developer to do who has created a popular program only to discover that half his users are Gnome users and half his users KDE users? He himself is probably a Gnome user as well and as such probably doesn't want to have to maintain a KDE version, but at the same time he understands people's desire to have a 'K' version, but believes probably like I do that it's silly to maintain two seperate versions of the same thing that only differ in their GUI toolkit esentially.
What's a developer to do?
When I use Nautilus I am afraid it may overwrite something or crash while I am doing a large file(s) transfer.
My main problem with the way copy and paste works is the following scenario.
---
If I am copying/moving a folder from one location to another and the transfer craps out. I now have two folders with the same name the original that still contains some folders and files, and the new one that has the files and folders that were successfully copied. I click on the original folder and do cut and paste it to where the new folder is.
Instead of preserving the data that was already copied it overwrites it!!!!
I have lost so much data because of this one thing, I don't think Konqueror does this, and I know that this does not happen in XP.
I now pretty much use an xterm or konqueror for such an operation.
The other thing is when copying to a folder with nautilus I have to refresh before I can paste.
Both things I have mentioned about Nautilus are reproducable.
Also I don't want Nautilus drawing my desktop. If I disable it with GCONF editor, then Nautilus takes longer to load. I can't win.
After all these complaints I still love GNOME more than other environments or WM, weird.
Also I think kwin is faster than metacity and that is why we sometimes get those crappy redraws.
Yep, a good installer just for the apps that can or should come with Gnome (or even KDE).
A place where you can not only choose what you want, but where it will be installed and how it will be inserted on menu and/or desktop with an option to uninstall.
Of course, it must has some kind of dependence checking.
This can save us of the current options: "almost everything" or "vanilla broken".
1. yes, a GOOD gnome ppp dialer is really needed.. a GOOD equivalent of kppp has been missing for what? 5 years? Gnome-ppp was getting there in Gnome 1.4, but it still hasn't been ported to gtk2. =(
2. a gtk port of plastik, hell even make it the default theme with a few tweaks. Plastik not only looks modern, but is quite usable. Plastik has a chance to become KDE's default (hopefully!) in KDE 3.3, so maybe we'll see both desktops have the same theme!
3. better integration architecturally - one thing that KDE excells at is that nearly all of it's components are highly integrated in terms of code.
4. make nautilus fully spatial or not spatial at all (keep it as it is). Don't have seperate modes for each. The classic MacOS finder was great, so I hope they stick with #1.
5. share icon themes between GNOME and KDE.. They use the same icon theme format even, but they haven't standardized the names of icons!
GNOME 2.4 is definatly kickass, but upon checking out the new kde 3.2 beta, I yearn for a desktop with KDE's features and architecture and GNOME's simplicity and feel. I think that's definatly possible with further development of GNOME, or general polishing of KDE. Overall, I think KDE 3.3 (or 4.0) and GNOME 2.8 (or 3.0) are probably going to be more alike than GNOME 2.4 and KDE 3.1 are.
I think there is no way to turn off the minimizing animation in Metacity except patching a source file. You can compile from a tarball or source RPM. Edit src/window.c and comment out or delete this:
/* Draw a nice cool animation */
meta_effects_draw_box_animation (window->screen,
&window_rect,
&icon_rect,
META_MINIMIZE_ANIMATION_LENGTH,
META_BOX_ANIM_SCALE);
I do the same thing to get rid of the beep in GDM. It's too bad the developers can't give us a simple preference to disable annoyances like these.
Great post Anony, thanks. Your last paragraph hit the nail on the head.
My two wishes.
1). That more people contribute to make the GNOME experience better. You see, it's open source for a reason.
2). Ximian, RH, SUN, IBM, Novell, SGI, HP and community work together to create a kick-a$$ development toolkit based on GTK+. No, not mono!!! It should be Unix-based and have no relationship to M$, Apple or other Proprietary related formats.
But man can dream, can't he? There's are reason there are called wishes.
Eliminate the programs menu, replace it with an "Applications" folder.
Better icons (less fisher price looking).
Expose
Eliminate the programs menu, replace it with an "Applications" folder.
You can do that. Just remove the Applications menu from the panel (you can do that with Gnome 2.4), and create a link on your desktop to the "applications:///" uri on Nautilus.
Victor.
I hope gThumb supercedes EOG too.
The "reinvent the wheel" argument is a bit lame, given that GNOME was started as a project to reinvent the wheel. Still, its a good thing that the KDE developers reinvent user-visible components:
1) The KDE versions are often better in many ways. How many HTML renderers does GNOME have? KDE has one, because KHTML was designed from the beginning to be lightweight and easily pluggable. How stable is gstreamer right now? KDE has had a stable media framework (aRts) for awhile know, and the heavily componentized architecture meant that video support was added easily through aRts. How many GNOME apps use components? Almost all KDE apps are componentized, because the KDE component architecture is so easy and lightweight. Why did KDE roll their own configuration (KConfig) and integrated menubar/toolbar (KAction) system? Because gconf didn't exist, and GNOME still has nothing comparable to KDE's fully-configurable toolbars. Why did KDE roll its own MDI mechanism? So all the major KDE development apps (Quanta, KDevelop, Kate) could use it. Why did the roll their own I/O mechanism? So all KDE apps could transparently access remote resources over a huge range of protocols. How come Abiword, which is part of GNOME Office, doesn't appear to use GNOME-VFS? And I still think Qt is a better toolkit in most respects (except maybe for internationalization and accessibility) than GTK+.
2) The KDE versions integrate better with KDE. Why would a KDE user want to use an app that doesn't respond to DCOP calls? Or embed itself in Konqueror? Or transparently support centralized configuration mechanisms (hotkeys, etc). Why would a KDE user want to use an app that didn't automatically use the system-wide wallet manager? Or the integrated KDE spellchecker? Or the advanced features of the Klipboard?
3) The KDE versions have a nice, integrated, KDE-style API. For developers, KDE is a complete application framework, like .NET or Java.
Its a cost/benefit thing. If the benefits are greater than the costs, a KDE version will get built. If they are not, then it won't. That's why programs like Konqueror, Kontact, JuK, etc, were built from the ground up; why non-user-visible libraries like libxml are used as-is; and why specialized programs like the Gimp don't get KDE versions at all.
Look, I'm not trying to begrudge GNOME. GNOME has its own development style, its own developer community, and its own user community. I'm just saying that you shouldn't begrudge KDE's right to have the same.
...also, the font-size of the desktop-icons seems to be bigger than the ones within nautilus...
Look in font preferences. The desktop font is set larger than the file browser [application] font by default in rh9 and fc1.
Eugena, sometimes you'll write things that make so much sense, and sometimes you'll write things that completely blow my mind. This article has a healthy mix of both. Here goes...
Nautilus scripts/addons: Couldn't agree more that these need to be exploited. This could be extremely powerful, and this is something that no other file manager has, to my knowledge. However, I couldn't agree less with your ideas about a generic context menu. Think about it this way...when you right-click on a tarball, what's the most likely thing you'll want to do with it? If you wanted to open it, you would have double-clicked. No, you probably want to extract it. I'd go so far as to say this is the most common action performed on tarballs on my machine. Why would I possibly want to move this action from the top level? Submenus in the context menu annoy the crap out of me. They're ugly as sin and awkward to use. I do NOT want to be running through a list of "add to music library" or "open with gthumb" to find my extract option. The same thing applies to music (being enqueued in rhythmbox), photos (open with EOG/Gthumb/whatever) and all sorts of other filetypes on which you invoke one common action.
A new modern theme: Wasn't this supposed to have happened by now? Gnome is clearly in dire need of this.
Themeing engine: I think you might be a little mixed up about how Gnome's "themeing engine" works. Things like the obscure for-Eugenia-only feature of smoothing the corners of the buttons are handled by each individual theme engine. You would have to get every one of them to implement this. I have no idea what you're talking about with "the height of the theme".
File selector: This irks me. If you're going to publish a journalistic piece on a well-reputed websight with relatively high traffic, PLEASE take the time to do a little research! The gui shown in these screenshots is a temporary UI to test features. When people write the code for things like the implementation of custom widgets, they need a way to test it don't they? This is a design by engineers, for engineers and is not intended to be the final product. The only reason you're seeing it at all is because the development is open. This is not the final file selector. AFAIK it's being designed as I write this. I mean seriously, it has a lone checkbox at the bottom that says "Frobnicate the file"! Please do some reading before you publish this kind of stuff. It damages not only the reputation of yourself and OSNews, but the reputation of Gnome itself. As far as usability goes, it gives you the option to set your own bookmarks and define custom widgets. That's pretty good. That's basically a scaled-down nautilus right there. What more do you want?
Development tools: I agree that Gnome is in need of a really good IDE. Anjuta's cool, and it has some neat features, but it's C/C++ only, and the pace of development is agonizing (I believe it's one guy). Scaffolding looked really promising, but appears to have been abandoned. An IDE that could let me choose whether I want to write in C, C++, C#, Java, Python, or Perl (Gnome projects AND non-Gnome projects for all said languages) would be most appreciated. Until that time, it's all about gVim. I would probably vote against the inclusion of Eclipse. It's very Java-centric, a massive beast of a memory hog, and terribly cluttered. While it's quite powerful for Java development, it leaves quite a bit to be desired when it comes to C++.
Faster GTK: Actually I had a friend over the other day who commented with a disapproving frown "I'm seeing a lot of lag in your widget rendering." He's a hardcore windows-advocate, and I had to give him this one. I'm not sure whether it's GTK, X, the kernel, or a little of all, but it's true that GTK2 is noticably slow.
Easier to use API: What the hell. You didn't say what's wrong with it! I don't know what's wrong with it. Have you tried any of the bindings? GTKmm is pretty straightforward, pyGTK is a dream, as is GTK#. I've not yet tried GTK-Perl (perl seems so poorly suited for such a thing), so I can't comment there.
Glade: Might I ask what's wrong with it (besides it being "junk")? I use Glade2 all the time and I have no qualms.
Copy/Paste: This is getting to be embarrassing. Someone has to do this before the next release or I'm switching desktops.
Burner app: I agree that Gnome needs one, but this is another area where you should really do more research before you open your mouth. Coaster is not abandoned. The coaster guy has been working on libburn for a year now, and it's getting quite functional. Coaster is really no more than a gui that exposes libburn's functionality. The dropline guy is writing another frontend to libburn. This is exactly the same thing that we're hoping to have happen with gaim once the core/ui split is complete.
Gnome Office: Gnome office (unlike koffice, which is crapola) shows real promise. What's needed now is integration with the Gnome desktop left right and center.
MS does it with Word. You type your email in a stripped-down Word and send as normal. If we could do this with Abi it would really rock, especially since the Evo 1.4 editor leaves a lot to be desired.
It has already been done.
AbiWord provides a bonobofied version of itself (if you compile it) and with that - evolution can use it.
At least the 1.4 could.
http://www.ph.unimelb.edu.au/~msevior/abiword/evolution-abi.png
This is for display only though - i think.
Ok... This is my biggest pet peeve. I want to be able to make a hotkey do any command I want. The predefined list is fine, but I would like to be able to do something like.
WinKey+E = Execute (evolution)
Furthermore, it should be made so you can do more fancy things like
Win+p = Execute (xterm -e irssi-text -c irc.freenode.net)
or something similar. Fluxbox allows you to do anything and it's the one reason I still use Sawfish if I'm using gnome.
I ma still using redhat 8 with ximian desktop, and this is what i use.
1) Use the network configuration (neat i think) to set up a dialup connection. The connections will usually be ppp0, ppp1 etc.
2)Add the modem lights applet. In preferences/General type this as ur connect command textbox,
/sbin/ifup ppp0
disconnect command
/sbin/ifdown ppp0
Now if u want to connect/disconnect, just press the button on the applet. It will turn yellow while dialing and then green on connection/red when disconnected also gives modem connection speed etc.
Of course the problem is changing telephone no.s . But my ISP has a roaming number , so it doesn't really bother me.
Alternatively, u could create seperate connections for different no.s and then change the pppx option appropriately.
Also make sure that in
Preferences/Advanced
Modem lock file :/var/lock/LCK..ttyS0
(could differ depending on which port ur modem is connected)
Check Verify owner of lock file as true and give the lock file option as ppp0 (usually the default).
Hope this helps.
And of course the same commands can be used from the command line as well. Modem lights will appropriately change colour.
If the anyone was really bothered about usability they could have just set it up by default and added the applet to the panel, once u set up ur connection.
It would also be nice to see an app like "ACDSee" or "irfanview" for windows within gnome. KDE has kuickshow which is my favorite linux app so far for this, but gtksee seems to get the job done as well, it could just be updated and included in gnome.
It wouldn't take too much work to create a gnome dialer using wvdial, libwvstreams gtk event bindings and gtkmm to create a gnome ppp dialer that is quite magical. See http://open.nit.ca/WvDial or http://open.nit.ca/WvStreams for information on wvdial and wvstreams.
What Gnome needs is to be a lot more like KDE, and what KDE needs is to be a lot more like Gnome.
Gnome wins on design and beauty, while KDE have the functionality that most users need. The Gnome default theeme is much more graphically distinct than the default KDE theame. Icons in Gnome are more distinct in shape, this makes them much faster to identify etc. Gnome have better feedback for e.g drag & drop, but KDE is more network transparant. Just look at the KDE file dialog, and compare it to the one in Gnome. Even the new beta 2.6 version is far behind that of KDE. Look at the search functioality in Nautilus compared to that in KDE/konqueror. In KDE you can not only find files based on their names, but also according to their contents. You can even ask for image that looks similar to certain given image.
I agree with Eugenia that there is a strong need to coordinate HIGs for the free desktop. Certain things should be very similar regardless of what platform you use. I'm thinking of file/print dialogs, the ordering of "Yes", "No" "cancel" buttons. I don't care if is the Gnoem way, the KDE way or some brand new standard. As long as both Gnome and KDE have a shortage of good applications written for each environment we need to make it possible to create consisten looking and acting desktops from a mix of applications.
... all the stuff Eugenia said, plus ...
Non-regular shaped theme support in Metacity like I used to have in Sawfish on Gnome 1.4! At the moment it's:
"how would you like your window border shaped sir? Plain, or plain with a slight rounding of the edges?!!?"
And before anyone says it, I can't be ar*ed to install Sawfish even though it apparently works.
Great review BTW
"the ordering of Yes, No, Cancel buttons. I don't care if it is the Gnome way, the KDE way, or some brand new standard."
As for Yes/No buttons, there is pretty much two ways to do it. Yes/No or No/Yes. If you find out another one, I vote for that!
so much for the kde vs gnome stuff.
i use kde because i simply like it.
konqueror the best file manager to date.
when it comes to apps, usability/functionality first.
the idea of kde + qt apps / gnome + gtk apps is simply pathetic
"The "reinvent the wheel" argument is a bit lame, given that GNOME was started as a project to reinvent the wheel. Still, its a good thing that the KDE developers reinvent user-visible components:
1) The KDE versions are often better in many ways. How many HTML renderers does GNOME have? KDE has one, because KHTML was designed from the beginning to be lightweight and easily pluggable. How stable is gstreamer right now? KDE has had a stable media framework (aRts) for awhile know, and the heavily componentized architecture meant that video support was added easily through aRts. How many GNOME apps use components? Almost all KDE apps are componentized, because the KDE component architecture is so easy and lightweight. Why did KDE roll their own configuration (KConfig) and integrated menubar/toolbar (KAction) system? Because gconf didn't exist, and GNOME still has nothing comparable to KDE's fully-configurable toolbars. Why did KDE roll its own MDI mechanism? So all the major KDE development apps (Quanta, KDevelop, Kate) could use it. Why did the roll their own I/O mechanism? So all KDE apps could transparently access remote resources over a huge range of protocols. How come Abiword, which is part of GNOME Office, doesn't appear to use GNOME-VFS? And I still think Qt is a better toolkit in most respects (except maybe for internationalization and accessibility) than GTK+.
2) The KDE versions integrate better with KDE. Why would a KDE user want to use an app that doesn't respond to DCOP calls? Or embed itself in Konqueror? Or transparently support centralized configuration mechanisms (hotkeys, etc). Why would a KDE user want to use an app that didn't automatically use the system-wide wallet manager? Or the integrated KDE spellchecker? Or the advanced features of the Klipboard?
3) The KDE versions have a nice, integrated, KDE-style API. For developers, KDE is a complete application framework, like .NET or Java.
Its a cost/benefit thing. If the benefits are greater than the costs, a KDE version will get built. If they are not, then it won't. That's why programs like Konqueror, Kontact, JuK, etc, were built from the ground up; why non-user-visible libraries like libxml are used as-is; and why specialized programs like the Gimp don't get KDE versions at all.
Look, I'm not trying to begrudge GNOME. GNOME has its own development style, its own developer community, and its own user community. I'm just saying that you shouldn't begrudge KDE's right to have the same."
That's true, there are important benefits to having a KDE version of an application, like the ones you mentioned and also if the GNOME version dies, the KDE version might still live.
However, the REAL solution is to make the desktops adopt a comon standard so pretty much the same benefits will be available regardless of what desktop it was intented for. I am glad to see that this is realized by more and more people and freedesktop.org is the first step in making this a reality. Finally people are realizing that the best way for Linux to get on the average user's desktop is not by more intense competition, rather more cooperation, a more unified seamless platform. This base platform will of course have variations like KDE and GNOME< but they should essentially use the same base technology.
Hopefully with DBUS, the freedesktop.org standard slated for inclusion in KDE 4.0, we will see a Linux desktop that is seamless, easier to use, interoperable and easy to develop for. I only hope GNOME will go forward with this too, it is a lot work to change to DBUS, more so than for KDE, but GNOME's COBRA really sucks anyway, unless it changed a lot since 2.2 when I last used it to build a minesweepers program, which I also created in Qt (Qt 3.1 is really MUCH easier).
Right now KDE has a great architecture, but several problems with the UI, however the backend is great and KDE's problems are really a lot easier to fix. Unfortunately there are powerful forces that GNOME has that KDE does not yet have. GNOME has FAR more paid developers, corporate backing from SUN< IBM< Novell etc. KDE doesn't really. I really hope this will change, because I believe one without the other won't get very far.
"As for Yes/No buttons, there is pretty much two ways to do it. Yes/No or No/Yes. If you find out another one, I vote for that!"
ROFL I hope all UI designer will just make it Yes/No, from my experience with others, users rarely read the dialogs anyway =p Just make sure that Yes is usually the best option.
"when it comes to apps, usability/functionality first. "
You'd be surprised how many people feel otherwise.
There are some valid reasons for trying to keep to one kind of app - memory constraints, for example. On a desktop with 512mb-1gb RAM, this seems silly, but on a 96mb laptop, it's a real concern.
Perhaps I'm just slow, but a cursory examination of k3b doesn't reveal any particularly amazing features beyond an unusually good sense of user friendliness. gToaster seems roughly equivalent, although I admit, it needs more work.
I was surprised that Eugenia didn't bring up Dashboard, which is making some relatively fast progress now that the first release approaches. The functionality is amazing at this point - it works extremely well with Epiphany and Gaim, as well as XChat.
-Erwos
Another point that bug me with GNOME is the permanent change of stuff. KDE does indeed take better care of this. In GNOME once something becomes halfway usable and stable developers trash the stuff and start over with huge changes in the GUI and in the behaviour of the program. A good example is Evolution here. People got used to Evolution 1.x, they had Addressbook, Mail, Tasks, Calendar and things like this selected from the left Tree and now with the 'entourage' look people now need to press extra buttons on the bottom only to get to another place of the program. This is not just happening with Evolution but it's one of the bigger examples. E.g. Nautilus. The Hierarchic way was quite ok the only thing it needed was more stabilisation, fixes of gnome-vfs (ftp, samba and so on). Make all the things like 'all-applications:// applications:// http:// ftp:// start-here:// burn:// fonts:// <anothersecret>://' shown somewhere or in a manual so people do not need to play trivia with them. But no, the entire concept of Nautilus (once it got halfway usable and a bit stable) was trashed over for something complete different.
I know software development is a permanent process. But the process doesn't mean that you must trash entire concepts over for different things. What we need are stable applications, applications that keep their look over a longer period. Once a program becomes stable the developer could head over to other things where real work is required and fix the stuff there. But this permanent changes of things before it even gets stable or usable is really annoying and will never end. End as in 'freeze' stay that way. New stuff should be added or old things altered, but please in a way that doesn't annoy people.
Most of the time it's the 'mood' of the developer that makes such changes. You know the sentence 'the one who supplies the code leads the programs direction' or 'the maintainer decides' ... The argument may be valid in a 10 people project but GNOME as well as KDE became too important for more than just 10 people thus it requires well thought things and more diplomatic way of how changes are being dealt with.
See KDE for example, a lot of stuff stayed the way it was since KDE 2.0 they smoothly changed Icons and some Themes, added new features and cleaned up, more stabilisation and better integration. Changes are being made but the bottom layer of the apps still behave the same. While Konqueror got new features and new icons it still acts, behaves as it was since 2.0 they haven't touched or trashed the entire concept to go for yet another way.
Well, you can have:
Yes No (KDE I guess?)
and
No Yes (Gnome)
as you have identified, but also:
Yes
No
and
No
Yes
I just think there are some irreconcilable differences. The option is to either use the GNOME technology, or the KDE technology. Standards are nice to a point, but often, standards that aren't concieved as a result of an implementation (TCP/IP, POSIX), and are instead made by committee (CORBA), often suck hard. Integrating technology wouldn't be such a bad thing if the GNOME developers were willing to budge a bit more (they seem rather disinclined to include anything with C++). However, as it is, it seems to me that there is a strong possibility that the GNOME developers, for various reasons including corporate clout and the lowest-common-denominator status of C, have the possibility of pushing technologies at KDE, but not adopting technologies from KDE. That leaves KDE using foreign APIs and tons of wrappers. We're already starting to see this with D-BUS --- its a C library that the KDE developers need to write Qt wrappers for.
Frankly, I'm not really interested in GNOME technology. They've made some huge advances in usability and user interface polish, but the guts really aren't very interesting. Its a nice evolution over classic APIs, but its not competitive with .NET or Cocoa. That's why the GNOME folks need Mono so badly. If the direction of technology flow in a standardized system is from GNOME -> KDE, instead of vice versa, KDE users would be stuck with a less capable platform.
In the same fashion as fobricate, how about Aye {yes} or Nay {no}?
"However, as it is, it seems to me that there is a strong possibility that the GNOME developers, for various reasons including corporate clout and the lowest-common-denominator status of C, have the possibility of pushing technologies at KDE, but not adopting technologies from KDE. That leaves KDE using foreign APIs and tons of wrappers. We're already starting to see this with D-BUS --- its a C library that the KDE developers need to write Qt wrappers for. "
This is true, but not the part about D-BUS. D-BUS first of all is NOT a GNOME technology, rather a freedesktop.org one. D-BUS also has C++ bindings for it, not just C. The main problems with it are regarding secuirity, everything else is already pretty much where it needs to be. You should also keep in mind that D-BUS is based on the powerful and flexible DCOP KDE technology. This makes it significantly easier for KDE to adopt this than GNOME and as I mentioned it yields many advantages for both desktops and other desktops since D-BUS is desktop agnostic, Havoc even said that adopting D-BUS is even more difficult for the GNOMErs. So it is actually more KDE technology than GNOME and also more likely to be adopted first by KDE than GNOME.
"Another point that bug me with GNOME is the permanent change of stuff. ... In GNOME once something becomes halfway usable and stable developers trash the stuff and start over with huge changes in the GUI and in the behaviour of the program. A good example is Evolution here. ... Nautilus."
8< ... etc
I would tend to agree with you in most cases but I think both of the examples you mention were more fundamental changes in policy and understanding of the whole Gnome project - an Epiphany if you will!
I am completely in favour of the move to a spatial Nautilus. The reasoning behind it is sound and the justification of decisions seems to have been thoroughly and satisfactorily explained. Additionally, the ArsTechnica article which appears to have been a contributor to the move was truly inspirational. If Gnome can incorporate everything that article was proposing (written for the Mac OSX finder) then there will be a lot of happy Gnome users out there.
Similarly with Evolution, Ximian have made a move to separate the various components which I consider to be a well-considered move. This will facilitate its (long overdue IMHO) inclusion in the main Gnome release, will make it launch faster, and will enable various backends to be used by other apps such as the Gaim/contacts and clock applet/calendar work currently being done as part of the Gnome bounties.
What I don't agree with is the sudden changes in applications. Galeon's in, Galeon's out, GGV's in, GGV's out etc., etc. I understand the reasoning with many of the decisions made in the past and things seem to have stabilised a bit in more recent releases so it may just have been part of the 1.x -> 2.x transition but product stability is key if major consumer desktop distributions are to be based on the Gnome DE.
Perhaps they should create stable Gnome code-forks (read on before speaking up on this one!) of all the apps (eg: Gnome-browser instead of Epiphany/Galeon/whatever etc.). These would ideally just be mirrors of the underlying applications unless the app developers decide to change track for some reason. Taking the example of the Galeon split, the "Gnome-browser" fork of galeon could have been developed separately from the point of the split and slowly transitioned to the Epiphany codebase rather than what happened where a perfectly good and stable browser was ripped out and replaced by a semi-functional early release which has only recently arrived at a similar level of quality.
I'm sure there are many other ways to achieve a similar result but something more is needed than the current attitude of "if you don't meet our requirements we'll rip your app out and replace it with something which is either completely different or alpha quality". I can't imagine Sun or Novell being too pleased with that kind of dramatic change happening once their software is on millions of corporate and consumer desktops around the world!
I agree with what you said about the new GtkFileSelector. It is very poorly designed and it needs to be rethought. We are better off with the current one.
Later on in the article when you stated "Glade is junk, end of story", I stopped reading. Glade is IMHO the most elegant app for designing user interfaces on OSS desktops. It generates pure C, not the "meta object" bloat that QT Designer churns out. Add to this the fact that you have access to all GTK 2.x and GNOME 2.x widgets. Not to mention that Glade conforms to GNOME's HIG so it is impeccably easy to use. Have you ever used Glade for any length of time? I doubt you know C very well either.
When or if you come to the conclusion any product is "junk", never voice it aloud or on a website. Voice your individual concearns in a diplomatic way. If glade truly is "junk", do you really think so many people would use it?
- Proper url-wrapping from terminal to clipboard/open URL. I read e-mail in mutt, and still use terminal-based IRC clients -- this is a feature I like, but it doesn't work very well. When there's a URL that wraps, it doesn't open properly, nor does it copy properly.
- Kitchen sink.
Or it could be,
Ναί η Όχι
or
Όχι η Ναί
LOL!!
Damn that was meant to be in Greek.
Nai or Oxi
Oxi or Nai
:b
Greek works in the phorums.
I belive the most promising gnome cd-burning application is ECLiPt Roaster, the stable version still uses the old GTK, but check out the development version too, it rocks.
http://freshmeat.net/projects/ecliptroaster/
"Glade is IMHO the most elegant app for designing user interfaces on OSS desktops. It generates pure C, not the "meta object" bloat that QT Designer churns out. Add to this the fact that you have access to all GTK 2.x and GNOME 2.x widgets. Not to mention that Glade conforms to GNOME's HIG so it is impeccably easy to use. Have you ever used Glade for any length of time? I doubt you know C very well either. "
I'm sorry, but I've spent about a week with Glade and if anything it isn't easy to use and it's documentation isn't all that great either. I don't care that it complies with GNome's HIG, maybe when I tried it it didn't, last try was 4 months ago but it is way too cluttered and the overall feel does not seem integrated like Qt Designer's. I am also suprised that you call the files that Qt provides "meta object bloat". They actually reduce bloat, are easier to use and maintain, they are considered a significant advantage by some users. XML is really great for interface files. QtDesigner also provides acess to the KDE and QtWidgets, there is a special KDE tab.
I hope that you aren't the same person saying about how glade has a scalable interface and you lay them out instead of painting them. As if this was something special. Qt/KDE have had this since the 1.x days, in fact even before GTK+ existed.
This may be a feature that would be nice to see on Microsoft development platforms like Visual Basic .NET which STILL does not have a layout manager (scalable interface, one ui design for all monitors and resolutions). At least they are starting to incorporate some Qt-like features in the new MS Visual Studio.
Anyway the GLADE vs QtDesigner issue is a dead one, GLADE really doesen't even stand comparrison and QtDesigner is a much more powerful tool.
concerning the video editing apps for gnome, kino has been fully ported to gtk2 (in CVS, but works)
only downside : they won't use GStreamer in its current state
--
for the burning tools, you are right, we need a good one for gnome. and staying away from cdrecord front-end crap.
burning tools need to use and extend libburn (see freedesktop software) so to share the real burning code
Personally, I don't really have a whishlist for GNOME, because everything that would be specific to GNOME is already beeing worked on or at least considered. 
I'd like to comment on a few things though.
@Eugenia's List
Nautilus Spatial Mode
I agree that it can be hard to know what exactly a folder is showing without the path. I'm not sure though I'd enjoy it constantly showing the Unix path... Also I think that I'd rather use the navigational window whenever this would really matter (for example when browsing the filesystem or a complex tree, like a source tree). Still it might be nice to have some kind of indicator where I am, maybe the path could be abstracted in some way... like "Desktop > Music > Classical" and "Desktop > Poems > Classical" instead of "/home/daniel/Desktop/Music/Classical" and "/home/daniel/Desktop/Poems/Classical". I don't know, just some brainstorming. Nevertheless, I love spatial Nautilus already. It feels just right to me, as if it always had been the plan for it. I also love that the browser interface won't be gone at all.
A New Modern Theme
True, as a Red Hat user it's easy to forget that Bluecurve isn't the default. I still enjoy Bluecurve, even after a few years I have always come back to it. I know that you don't like square widgets, but I have a feeling that this is going to be changed soon. On the Todo List of Red Hat Artwork ( http://fedora.redhat.com/projects/artwork/ ) is "Modify widget style" and I believe that Garrett intends it to look a bit more like the window border buttons (thus rounded). I also hope for a bit of color separation between buttons and background and that the new style will look as clean and professional as the current one. It's certainly not trivial to make rounded widgets look good (there isn't a single one besides Aqua and maybe even Luna which I could look at for a longer time without getting sick of it).
ATM I enjoy Bluecurve Berry and Cream together with the "Simple" border style, which looks very square, but also very slick. 
File Selector
As others have pointed out, the current one really is just a testbed for the features of the API. It is wrong to say that it would be difficult to change it again. The new API should make that trivial. So if there is anything about the current API you don't like, then you should speak up. The interface can and will change.
Nautilus Authentication (local files)
I couldn't agree more, I absolutely think that this is one of those rather minor but particulary tricky problems that need to be solved for the home desktop (and for everyone else it would be nice too). In fact, I'd have to correct my statement from above for this one, as this particular feature doesn't seem to be planned yet.
Preference Panels, System Integration
Oh I also agree with that, the Ximian preferences are much more convenient IMO than the preferences menu. It's not bugging me a lot but this should be reconsidered. I think there was some talk about it once, but not anymore lately.
Video Editor
That would be great, but why should Ximian do it. This probably won't help them making money on the corporate desktop while still consuming lots of ressources. A community project might have better chances. In the meantime, there might be some proprietary solutions once Linux becomes more viable for the home desktop.
Epiphany
Marco mentioned a web developer addon once, I believe that's a great idea, at least better than adding all the little details to the interface that only developers care for or making the compromise of only adding the most useful developer features (I want them all ).
Images, Cameras, PDAs
I don't think gthump should ever replace EoG. If anything, it should supplyment it. EoG is much better IMO for viewing a single image after clicking on it in the filemanager.
@Charlie
Mozilla starts up is less than 2 seconds, if that. Epiphany can take up to 5 or 6 seconds yet it is supposed to be native, leaner, and meaner.
Many of the smaller applications, like the calculator or most of the games, take more than 5 seconds to start when they should be nearly instantaneous.
There really must be something wrong on your system. Maybe my system is just WAY faster than yours, but then Mozilla couldn't start in less than two seconds for you.
I just booted the computer and all the smaller applications like calculator and "most of the games" instead launch nearly instantaneous (less than a second).
Maybe this is because of something done in Fedora (like prelinking), I don't know. It was much worse for me with RH9 indeed, but RH9 was probably the slowest distro I have ever used for a longer time. I also noticed that sometimes application load time increases when my system is under heavy load, but that's probably to be expected.
@stu
[/i]What I don't agree with is the sudden changes in applications. Galeon's in, Galeon's out[/i]
That's not true, Galeon was never in. The release team is _very_ picky about accepting new modules, especially for the reason you mention. Epiphany is the first official GNOME browser and even this one had a hard time to get accepted at all. With the main issue beeing accessability considerations IIRC.
My vertical option for yes/no buttons was meant to be sensible, though I don't think it should be used Earlier versions of Macintosh used them in some places (filepickers for example). Using 0 and 1 is just silly.
The new filepicker is better than the old one. You mightn't think the interface is all that good, but the reason GTK's been stuck with the same old interface is because once they've changed the API, the interface can change too. I think for example that with the new API, ROX users (for instance) will be able to use a rox dnd savebox-style thing instead of a minifiler.
As for Glade, if it's HIG-based, then it must be some special secret HIG I haven't heard about. 'Are you sure you want to quit?/Cancel/Quit'? Whatever happened to 'Do you want to save the changes you made to the project "Foo"? Your changes will be lost if you don't save them./Don't save/Cancel/Save'? Glade is about as easy-to-use as a left-handed canopener for a quadraplegic.
Sorry, i don't quiet get why you would need a 'more professional clean' theme, like plastic off KDE. Don't get me wrong i like plastic, but Ximian's industrial in my books is nicer, it is clean and in fact looks very similar to plastic (or the other way round) - i actually used plastic so i can get some more 'unified' theming for my KDE-Apps on my Gnome-Desktop, having said that I have ditched KDE and its Apps completely now and are using solely Gnome. But really what is wrong with Industrial - it does look more polished and professional than plastic does (the highlights on plastics buttons are not very nice)
It is not easy at all to write a good column, cuz' I like writing sometimes. Well, this is overall a quite good column. However, If May I have a bit idea on it, it should be avoided to use active sentences such as "I want this" , "I need that" and so on. Unfortunately, GNOME is being developed by community and in GNU, only You want or you need that is not enough and that is very personal flavours. That is my personal idea only. About article, it is agreed up to 80% by me :-). Thanks
Eugenia, have you actually filed feature request bugs on GNOME's bugzilla for the less adventurous of your suggestions?
'cause i'm sure there are a few trivial issues, such as nautilus authentication(isn't it like using a GUI su when trying to run system config tools from a non root user?), which i'm sure could get dealt with fairly rapidly.
* File manager (Nautilus) :
- A better list and tree view. Icons are eye candy but often I want to see as much files as possible to navigate/copy/move/find.
- Good search functions so I can stop running a terminal to launch "locate" or "find" or "grep"
- Improved speed when viewing a folder with large number of files (try opening a folder with hundred of svg tests file in nautilus..)
- Better display for files with long name (it really looks bad and take space when nautilus starts a new line only to draw the last "e" of a name..). I think explorer cut long names so they re always drawn on only 2 lines
* Desktop :
- Fast switching of users
- Extend the concept of virtual desktop so we can have different icons & wallpaper on each
* Text editor (Gedit) :
- A way to edit files through ssh (or ftp, etc..)
- I'd like to have gedit warn me when a file is changed by another application
- A list of favorite files to quickly open logs, todo list file, or current project files
* Mail :
- It would be nice to have a mail only app with all the cool features of evolution mail (I don't need weather, calendar, rss agregation or what else, I thought gnome was aiming for simplicity ?)
* Web browser (Epiphany) :
- Autoscroll when clicking with middle button as in IE (or mozilla with some javascript extension)
- Grouped bookmark a la mozilla, very usefull to open your favorite sites in one click
* Multimedia :
- I'd love a small applet to play audio files. Just this. Play. Audio. An unobstructive one with no useless visualization plugins, stupid skins, or database behind, or burning or conversion feature or .. well you get it.
- For video totem is nice, keep it like this !
- As other have mentionned an acdsee like tool to see images would be nice. gthumb is not bad though I had some crashes
I think that the PPL from GNOME and KDE should get together and create a theme system that works on both DE, so for example I can create a theme and it will be able to be used on both? Possible? Yes it is, we just need code generators for both, so when it comes to building themes in c/c++ the generator could build sources from XML, for both DE and compile them... Can be hard but will work 
With the vast amount of support that Gnome is getting from Sun (500,000 installed desktops in China, a possible 800,000 more in the UK), it now has to be considered the premier Linux desktop. You can pretty much guarantee that most of the things on your wishlist are already being implemented as we speak, and I estimate that within a year from now, it will be far and away the best DE for Linux.
> the ordering of Yes, No, Cancel buttons...
Actually, there are six ways of ordering Yes, No & Cancel buttons horizontally
1. Yes, No, Cancel (aka KDEish order)
2. Yes, Cancel, No
3. Cancel, Yes, No
4. Cancel, No, Yes
5. No, Yes, Cancel
6. No, Cancel, Yes (aka Gnomish order)
Consider there is 30,000,000 desktops in China and still M$ couldn't make money there, do you really think JDS will make a dent there ?
I personally don't agree about the need to make icon size determine by default the window title height.
If you want to design nifty icons, please drop the antique pixel-based design and begin using SVG icons. These render nicely at any size.
Pixmap graphics for anything else than 2D games should be treated as obsolete if we actually want to get an open desktop architecture (so that different constructing blocks, configurations and resolutions may be truly interchangeable).
I agree with everything (some of the reasons made me switch to KDE, and am very happy with it), except the part about Glade. It's mostly the only thing in Gnome that attracts me to develop in it. That and Gtk#.
The ability to force applications to start in their specified workspace, á la Window Maker. An added bonus would be workspace specific icons/launchers.
I use gnome as my main desktop and I don't have any real gripes exept :
- I would really like to be able to set the homedir (startdir) in Nautilus (and the file selector), so I don't have look at all the configuration dirs
- An easy way to edit icon themes would be nice
I'm not a DE zealot. However, I've tried all the releases of GNOME up to 2.4 and each time, I find it wanting. The problems are numerous and I'm not going to go into them here. The point is, I believe the majority of Linux users do not use GNOME, so their wishlist is basically for GNOME apps to work more seamlessly with their DE of choice (usually KDE.)
The freedesktop folks are sorting out the nitty gritty stuff like drag and drop, and that is basically not a problem, or won't be soon. A bigger problem is in theming and fonts. How can I make my GTK apps look like my Qt apps? The don't have to look identical, but if I set my default fonts, then I need *all* my apps to follow those defaults. I don't know if a unified theme manager is feasible, but it is needed. GNOME theming is a joke, as it is. Why am i unable to change colour schemes without having to install a whole new theme? I've never yet found a GNOME theme that looks good.
There is no point talking about "Yes, No, Cancel" ordering in GNOME, because our goal is to *not use* these non-specific terms. Read the HIG - dialogue box buttons need to describe what you're going to do, not answer a question. Once you grok that, you'll grok the rationale behind the ordering.
You should reassess your comment that "the majority of Linux users do not use GNOME", considering the number of massively large scale deployments we've seen in Brazil, Spain, China, and more on the way. :-)
The majority of Linux desktop users do not post to OSNews. :-)
I just hope the gnome team don't ever remove the ablity to tab complete filenames from within the gtk file selector dialog, that feature alone makes up for the ugliness of the beast.
>> Give me back the rounded corners! Please!!
Easy one. Just make the right panel-backbround with gnome.
Here is a little instruction (in German):
http://www.osnanet.de/c.schroeder/linux/gnome.html
Or download a panel-background;
http://www.osnanet.de/c.schroeder/download/panel.png
Screenshot:
http://www.osnanet.de/c.schroeder/linux/gnome2-shozt.png
"GnomeMeeting is not a casual chat application and primarily is a PC-2-Phone application following its maintainers"
I'm the main maintainer and I can tell it is completely wrong.
Reaction here :
http://mail.gnome.org/archives//gnomemeeting-list/2003-December/msg...
Don't take it bad, that's just wrong, and I needed to clarify things so that people do not have bad info in mind.
1. _Good_ fonts in gnome-terminal. These days, I have to
use either rxvt or xterm while running Gnome.
I do really miss gnome-terminal that was available in Gnome
1.x. (IMO, the image viewer in G1.4 was also better than
the current one).
2. In Alt+F2, let arrow keys work as in bash (and KDE's
Alt+F2), not vice versa.
3. _Much more_ evident way to take icons away from the
desktop.
4. A possibility to have the panel shifted to the left or
right edge of the screen.
5. More nice clock built-in into the panel. :-)
TIA!
:-)
speaking of integration - any chance one might see some sort of "konqueror for gnome" some time in the future, means file- and webbrowser combined?
i'm used to use that (and, of course, many others, too...) in windows (a lot), and it's the only major reason that makes me sometimes think about switching to kde.
> 1. _Good_ fonts in gnome-terminal. These days, I have to
> use either rxvt or xterm while running Gnome.
Right click in the terminal, select "Edit current profile" and you'll see a dialog appear with a "Font:" option. So in fact, you CAN select what font(s) you want to use in gnome-terminal. You can hardly blame GNOME for what fonts your distro ships with.
> 4. A possibility to have the panel shifted to the left
> or right edge of the screen.
This is possible and very accessible. I bet you've never even used GNOME.
Has anybody filled a bug on bugzilla about this nautilus authentication thing? Seems everybody agrees nautilus should have that (and as far is i know, they aren't implementing that yet, which is a shame).
Victor.
There is no point talking about "Yes, No, Cancel" ordering in GNOME, because our goal is to *not use* these non-specific terms.
Well, read Yes as 'doing the default and confirming the implied action' and no as rejecting it and such. We're just simplying language for the sake of conversation.
"Simplifying the language for the sake of the conversation" doesn't help when it actually changes the debate, particularly for onlookers. You won't believe the number of times we've had to re-explain this because people thought they could describe it with "Yes, No, Cancel". We're *very specifically* moving away from those kinds of incomprehensible dialogues. :-)
See gnome-keyring and the vfs-daemon. It will be fixed in 2.6, as well as giving us a number of el-neat-o features. :-)
More and more people are wanting their favorite applications to integrate tightly with both desktops, but at the moment that's near impossible. What's a developer to do who has created a popular program only to discover that half his users are Gnome users and half his users KDE users?
An unconventional thought just came to me, which could save the Linux world... Having followed user arguments about Gnome vs. KDE for quite a while now, it seems conclusive: KDE has more advanced underlying technologies, while Gnome has a better-developed UI. Thus, my idea is compromise on both sides: Gnome (lets say, in version 3.0) switches from GTK+ to Qt as its developing environment, thus solving the problem of developer tools, network transparency, extensible programming language support, and speed. KDE, then, should adopt Gnome's HIG. Sure, it can keep its desktop as unique as possible, but where applications are concerned there should be a standard.
This would lead to a consolidation of Linux applications, an easier time for developers and a better experience for end users.
I should also mention that in an interview, Novell's 2nd-in-command (I forget his name) targeted unified development tools for Gnome and KDE as a main goal for the near future.
And PLEASE ! include a "kepp on top" feature in metacity.
If u think, that's tooo geaky (and I really DON'T think so), than include an window for that in the control panel. Such a thing :
WindowManager To use :
Metacity
Other : ...... (put there what u want)
(if Metacity hasn't been selected what is written under that is in grey : else u can configure it)
[X] Option "Keep On Top" available in metacity menu
[X] Option "Keep On Bottom" available in metacity menu
[X] Option "Omipresent" available in metacity menu
[12] size fonts for metacty
... and other option. Keep it simple by default, but I WANT my keep on top feature. And dont say me to use the kay combinition for that because
1) It's not documented. Nobody know from that
2) It's buggy !!!!
3) they are also people who likes to use the mouse. If i would'nt do that I would stick with mc instead of nautilus.
Hi, two people has requested this wishes.
Ok, you have them! 
For Metacity shortcuts you only need to edit gconf configuration via gconf-editor and then go to "apps"->"metacity" and there you can set you're preferred shortcuts, I have: Win+E -> Evolution, Win + B -> Epiphany (browser), ...
For Starting Apps in predefined Workspaces, you should use a program called Devilspie, which uses a XML configuration to "hook" windows when they're are launched and then do actions about it like, maximizing, moving to a workspace, ...
http://alts.homelinux.net/
"Read the HIG - dialogue box buttons need to describe what you're going to do, not answer a question."
Well, they do answer a question. Seriously, I don't know how buttons can avoid answering a question, when the dialog does ask a qustion. The word "dialog" means "conversation": the application asks questions via dialogs and the user answers them via clicking buttons.
Some examples:
| File "testme.pl" is an executable text file.
| Do you want to run it or view it?
| [Run in a terminal][View][Cancel][Run]
Does it ask a question? Yes. Do the buttons [Run] and [View] provide answers for that question? Yes.
| File "testme.txt" has been changed.
| Do you want to save it before quitting?
| [Don't save][Cancel][Save]
Do the buttons [Save] and [Don't save] mean exactly the same as [Yes] and [No]? Yes. They just have more descriptive labels. The sense is not changed by renaming buttons from "Yes" to "Yes I wanna save the file testme.txt" or "Save".
Please compare these two vending machines in terms of usability:
Koffee TM: "How do you like your coffee?"
[Black]
[White]
[Mocha]
[Cappuccino]
[Cancel]
[No suggar]
[Extra suggar]
Coffee Gnome TM: "Do you like black coffee?"
[No suggar]
[Extra suggar]
[Cappuccino]
[Mocha]
[White]
[Cancel]
[Black]
From the second one I can't even tell wheather I can have cappuchino with extra suggar. Well, there is definitely more chance I will hit the cancel button to collect my money back...
See gnome-keyring and the vfs-daemon. It will be fixed in 2.6, as well as giving us a number of el-neat-o features. :-)
Are we talking about the thing? Getting the right permissions when trying to access a local file (like /etc/foobar)?
Spark, I'm talking about auth caching in gnome-vfs for remote filesystems. There shouldn't be anything wrong with reading the correct permissions, local or remote.
Last time I played with gnome, metacity could not center windows as they display when first created, and I always found myself moving the window to the center, I hate looking at a window that is up in the left hand corner of the screen. So I use KDE basicly because of that.
But I also use KDE because of Konq's incredible integeration with SSH and SFT using fish.
I can browse around other servers and workstations that I have access to just like it was a local file system, and all the mime types work. Gnome and KDE should join forces and combine the best of both worlds into a new Desktop. Do we really need two major ones?
Yes reading the permissions is not the problem. The problem is that you can't do anything with those files unless you have started Nautilus or even GNOME as root. I do believe that this is a really critical problem if you have to tell every home user that he has to use the terminal to access some files on his very own computer. While it usually shouldn't be neccessary to edit such files, it certainly shouldn't be impossible either.
That's what this paragraph of the article was about and what I agree with. There is no Linux desktop yet which can do this, but OS X does.
Thats the only thing that keeps me using Konqueror (from gnome). Also add scroll wheel virtual desktop switching like in Fluxbox and KDE 3.2.
I wish gnome-vfs didn't suck. I mean in kde I can open files right out of konqueror over ftp,smb,sftp edit them click save and bam life is good. Also it auto asks for the proper authentication credentials every time. Does this happen in gnome? HELL NO!!!! If I even want to access a remote resource I have to type in urls lie : user:pass@host
then I have to copy the file to the desktop then copy it back. What a pain in the ass. If nautilus worked liked konqueror I would never use kde again.
BTW, the html renderer example is entirely bogus.
GNOME has two HTML rendering engines because one of them is capable of editing and one is not. It's unfortunate that they forked in this way, but that's the way it happened.
KDE has one HTML rendering engine. It is not capable of editing. As a result KMail is not able to do HTML emails, whereas Evolution can.
So, while Gnome has a roundabout way of getting these features, it's still a lot better than not having them at all.
The others are also mostly bogus. You're a smart guy Rayiner, but please choose your examples more carefully in future. aRts is not a "powerful, componentized media framework", it's a glorified sound server. Adding video is/was a painful hack - there's a reason KDE are thinking of dropping it in favour of GStreamer in KDE 4.
KDE doesn't have anything that would resemble a traditional "component architecture" - typically such things are language neutral for instance. KParts is merely a glorified C++ class loader.
Finally, much of its architecture is going to be reinvented anyway, but in a KDE and GNOME neutral fashion. For instance Waldo Bastien wants to do a new VFS system.
Hi Eugenia,
nice article. Many of the things I share.
I'd add this too:
Maybe it's my fault but I have got an 800x600 15" monitor.
Sometimes I cannot use PEARL applications because they are thought for an higher resolution ie DVDRip. It would be nice if gnome would allow the user to move the window outside the screen borders, to get rid of the OK and Apply buttons that are always out of sight and click....
- Regards
- Pashs
True network transparancy (see KDE).
Better printing support (see KDE).
More "efficient" nautilus modes (see konqueror).
Mouse Gestures and more (see KDE-3.2/khotkeys).
Ability to easily change ui colours (see KDE).
More tools for configuring system related settings that might need root password.
Rhythmbox with support for changing songs by scrolling with the mousewheel on the system tray icon (see KDE-3.2/juk).
"More tools for configuring system related settings that might need root password."
See gnome-system-tools
Hi,
I just read your article and as the UI designer and coder of Coaster, I would like to let everyone know that Coaster IS NOT DEAD. Harshy has been busy with school and recently got very ill, so updates to the website have been few and far between. Right now, we are waiting for libburn ( http://icculus.org/burn ) to stabalize so we can get a file format down. Anyway, if any of you would like to know more, join #coaster on irc.gimp.net and your questions can be answered there. Thanks for the good article .
-Bryan
On PlanetGnome http://www.gnome.org/~jdub/planetgnome/
So the every amusing OSNews is running a GNOME Wishlist. Sigh. Where do I start?
Nautilus Scripts/Addons
Nautilus 2.6 will support funky new plugins, with a clean API and decent menu merging. I though Eugenia was keeping up with Nautilus development, she certainly posts on nautilus-list now and again.
Spatial Mode
It's not finished -- either help make it totally rock by commenting/fixing/patching, or wait until is it finished.
Metacity Features
The standard rant about "viewports" and "workspaces". Again. Jesus people, ask for a feature and not the name for a subset of available features that you have used before. Please. I'm about to ask God to extend his kitten-killing to Metacity... For glueing windows to the corner of the screen... well, press shift once you are close to the corner and it will magically glue itself, without snapping to the windows it sees en route.
File Selector
Eugenia's comments are pretty useless: "...is pretty bad". Not really a comprehensive UI review, but thanks anyway. But that's not what I'm confused about -- I'm confused about the number of people who think that the "Frobnicate this file" check box in Frederico's example screenshot is part of the default UI! Please engage brain before posting. What would "frobnicating" do to a document I opened in gedit? In galeon? Did you ever consider the possibility that this widget is an example of an extra widget the developer can add to the file selector?
Volume and Showdesktop Icons
Honestly, compared against some of the bugs in GNOME this is laughable. Extract the patch from Red Hat, and file it as a bug upstream. It will probably be applied. Not Hard Work.
Development Tools
"Glade is junk, end of story" -- Eugenia. Right. Personally I consider Glade to be a wonderful interface designer, and makes coding GTK+ interfaces trivial. I hope this isn't referring to the "Generate Source" button in Glade, which is generally considered to be A Bad Idea when using C.
Personally, I hope that Eclipse's C support will mature and someone integrates Glade somehow, even if it is just a button to launch the binary. But I'm happy with Glade + libglade + xemacs.
Copy/Paste still misbehaves after all these years
File bugs! There is no fundamental reason why copy and paste shouldn't work, as is shown by the recent gaim hacking to copy right text to/from the gaim chat window into the Evolution composer.
GConf Editor
A search button for gconf-editor could be handy, but generally tried looking in /apps/[app-name]?
Samba on Nautilus
GNOME 2.4 I believe had a new smb: implementation, and in GNOME 2.6 it will rock even more.
Rhythmbox and Totem
"Use the XMMS visualization plugins" -- not possible. It is impossible to link GTK+ 1.2 and GTK+ 2.x code in the same binary.
"Totem ... use either Gstreamer or Xine on the fly" -- why? I'd say that everything GStreamer can play, Xine can play. If you want a player for everything, use Xine. In the future when GStreamer has the required features, we'll all be able to switch over to that instead.
"I would like Totem to recognize the file format and show an alert to the user "would you like to download from the web these formats and install them?" -- Totem already does this when it can, and has done so for a long time.
Epiphany
The usual minor issues which get blown out of all proportion as show-stopping bugs for the entire desktop. File a bug, create a patch, do something!
Text and Video Messaging Integration
I think the gnomemeeting maintainer covered this one...
Burning Application
GNOME the desktop is going towards tools to help end-users. Thus we have nautilus-cd-burner which is wonderful for the very common task of "burn these files". I have a patch (honest, I do) for Rhythmbox which lets that burn audio CDs from playlists. I don't see the need for a fully-featured 100% coverage CD burning tool in the GNOME desktop.
Most important for me is the topic with the legal/not legal codecs for eg Totem and the topic about copy and paste.
My full agreement to this !!!!!
For Videoediting the program transcode (swiss army knife for videocoding) will be a beautiful codebase
- I was referring to the seperate HTML engines inside Nautilus (GTK-KHTML) and Epiphany (Gecko). I don't really care if an app like Evolution uses it if the main engine isn't capable of a feature it needs, but there is no reason for the GTK-KHTML and Gecko split.
- aRts is a powerful multimedia framework. A lot of its capabilities are not documented as such, but they are there. It brings multimedia capabilities to every KDE app. I'm not going to argue that its as good as GStreamer. However, GStreamer really isn't a competitor to aRts yet --- its not yet fully stable, and almost no GNOME apps use it. As for GStreamer in KDE, we'll see. GStreamer is up for consideration, but along with other options like MAS. MAS would probably be a better choice, because it is desktop neutral and fills the whole range of required capability (sound server, network transparency, multimedia framework) instead of just part of it like GStreamer.
--- You do realize that, in this regard, KParts isn't any different from COM? COM is just a glorified C++ class loader too. They achieve cross-language compatibility by having other languages understand the Visual C++ v-table layout. KPart's language bindings (they've got ones for Java and Python, among others) operate on the same principle. On top of that, KParts provides the UI-merging and document handling features of OLE. So if COM/OLE is a component framework, so is KParts. Its not as overengineered as Bonobo/CORBA, but that's a good thing. CORBA is a first-class pain in the ass. Its one of those crappy designed-by-committee APIs I was referring to. The C++ mapping is really and truely horrid, and its alien to both classic C++ and modern C++ design styles. I know this from first-hand experience --- my current project at work is CORBA from the ground-up. A lot of the FUD about CORBA isn't true. Some ORBs, particularly OmniORB and ORBit, are really quite fast. But it is a hassle to use. In the end, the proof is in the pudding. KParts are used everywhere in KDE. They're so easy and well integrated, there is a very low opportunity cost in using them. In comparison, there are just a handful of Bonobo-enabled applications. And does GNOME have anything like DCOP scripting?
--- Actually, the developers have stated that much of the architecture would *not* be reinvented, at least not in the time-frame of KDE4. Its Havoc Pennington that wants the new VFS layer, not Waldo Bastien. Again, I would not have a problem with it, as long as the result is closer to KIO (every app uses it transparently), than to GNOME-VFS (some apps do, some apps don't).
The nice things about the KDE architecture aren't just embodied in each seperate component or API. Its about the unity. Unity is everywhere in the KDE framework, and it is strangely absent in the GNOME one.
I might be getting a little defensive about KDE, but I've got good reason to be. There is a principle known as "worse is better." It states that worse technology gets more widely used than better technology, because worse technology often has some advantages, like backwards compatibility, or ease of implementation, or corporate backing. Any alternative computer language user is painfully aware of this fact. GNOME has the potential of being that worse technology that still gets more widespread use. Its got backwards compatibility on its side --- there are many more C developers, especially in FOSS, than C++ ones. Its got ease of implementation on its side --- GNOME is structured as a loosely coupled collection of libraries, that plays nice with external technologies (Mozilla, OpenOffice, etc), while KDE is an integrated whole. Now, GNOME has major corporate support too.
Just because I'm paranoid doesn't mean that they're not out to get me!
What is gnome-keyring?
I have the impression you misunderstood what i was talking about...
Victor.
I didn't mean to come off so harsh on GNOME. It really is a nice desktop environment, and I'll be the first to admit that their insistence on following the HIG have produced some extremely polished applications. I'm just saying that the underlying guts are nothing special compared to the competition. Users coming from other environments might not notice the difference, but its painfully obvious to a KDE user.
I'd love to see fast image thumbnailing for nautilus.
I've yet to see any app running on xfree that can generate thunbnails on the fly quickly like windows does.
There's obviously something wrong with the way developers are approaching this issue on xfree. The same task on windows seems to be about 15-20x faster.
What exactly is windows doing that's so different?
Some of you here with lots of photos(3000+) will know exactly what i'm talking about here
Gnomemeeting needs to be able to communicate with the p2p phone program SKYPE so we Linux users can get into contact with the "millions" that are claimed to have downloaded SKYPE and are running it on their Windows boxes... I can't believe it's just me who would like to p2p with the fam. for christmas ;o)
And on our wish list this year is a NICE FAT CHECK!
There you have it Santa! If you did that you'd help us put quite a few people into full time jobs developing free (as in beer), software for the disabled of the world, and support our move to open source in the new year as well :o) - Does it get much better?
If anyone else wants to chip in, visit
globability.org/signature/sig.htm
ANY HELP will be appreciated!
Best Regards - And a Merry Christmas to everyone from The Development Team!
>Nautilus 2.6 will support funky new plugins, with a clean >API and decent menu merging. I though Eugenia was keeping >up with Nautilus development, she certainly posts on >nautilus-list now and again.
-->Nice !
>Metacity Features
-->Again : A Nice Control Panel for metacity, where u can also add a "I want my keep on top feature" . There is a bug sumbitted for that in bugzilla, but only the possibility to assign a keystroke for that has been implemented. Seriously, that's NOT a geeky feature
Development Tools
>"Glade is junk, end of story" -- Eugenia. Right. Personally I consider Glade to be a wonderful interface designer, and makes coding GTK+ interfaces trivial. I hope this isn't referring to the "Generate Source" button in Glade, which is generally considered to be A Bad Idea when using C.
>Personally, I hope that Eclipse's C support will mature and someone integrates Glade somehow, even if it is just a button to launch the binary. But I'm happy with Glade +
libglade + xemacs.
-->I either don't like Glade, I think QtDesigner is better, but ok.
doing a gtk-interface-builder plugin for eclipse would be very nice, but I think, it's very hard. But note that then eclipse3.0 come's out, CDT will also come out - so we will have a real nice C/C++ developper platform. Perhaps then, time has come to implement a gtk-plugin for eclipse. The fact is, we should not underestimate eclipse : I think, it will soon be somethink like "The Standart IDE"
>GConf Editor
>A search button for gconf-editor could be handy, but generally tried looking in /apps/[app-name]?
-->I think that should be a nice feature
>Samba on Nautilus
>GNOME 2.4 I believe had a new smb: implementation, and in >GNOME 2.6 it will rock even more.
-->No, sorry, but Samba implementation in gnome 2.4 is really really bad. Nautilus allways takes so much time to navigate throw my samba network (konqueror, and linNeighborhood are so much faster!). Worser : it so often happens that nautilus hangs up.
>Rhythmbox and Totem
>"Use the XMMS visualization plugins" -- not possible. It is impossible to link GTK+ 1.2 and GTK+ 2.x code in the same binary.
-->They are plugins from xmms ported to beep-media-player, which is a gtk2 app. look at :
http://sourceforge.net/mailarchive/forum.php?thread_id=3521242&foru...
But is it enought top port them to rhythmbox/Totem ? Do they have the same plugin interface ? (Does Rhythmbox has a plugin interface ???)
>"Totem ... use either Gstreamer or Xine on the fly" -- why? I'd say that everything GStreamer can play, Xine can play. If you want a player for everything, use Xine. In the future when GStreamer has the required features, we'll all be able to switch over to that instead.
--> I also though that this is a good idea. It would be nice to switch to what ever multimedia framework I want - gstreamer, nmm, Xine... But I think, it's not really possible. But gstreamer and Xine would be nice. I think that u only can play file with xine, but with gstreamer-cvs , u can now edit tags .
>Burning Application
>GNOME the desktop is going towards tools to help end-users. Thus we have nautilus-cd-burner which is wonderful for the very common task of "burn these files". I have a patch (honest, I do) for Rhythmbox which lets that burn audio CDs from playlists. I don't see the need for a fully-featured 100% coverage CD burning tool in the GNOME desktop.
-->Windows also has a CDBurning tool, that is quite good. I think, it's an important application
An unconventional thought just came to me, which could save the Linux world... Having followed user arguments about Gnome vs. KDE for quite a while now, it seems conclusive: KDE has more advanced underlying technologies, while Gnome has a better-developed UI. Thus, my idea is compromise on both sides: Gnome (lets say, in version 3.0) switches from GTK+ to Qt as its developing environment, thus solving the problem of developer tools, network transparency, extensible programming language support, and speed. KDE, then, should adopt Gnome's HIG. Sure, it can keep its desktop as unique as possible, but where applications are concerned there should be a standard.
This would lead to a consolidation of Linux applications, an easier time for developers and a better experience for end users.
I should also mention that in an interview, Novell's 2nd-in-command (I forget his name) targeted unified development tools for Gnome and KDE as a main goal for the near future.
Excuse but that's really dumb, specially cause the better applications are made of GTK, evolution, gimp, etc.., I still haven found that killer appication on KDE (qt) that will made an enterprise use it, Suse is a KDE based distro yet includes killer GTK applications like evolution and GIMP, maybe KDE should move to GTK.
While I'll agree that most of the better applications are made with GTK+, there are several killer apps for Qt:
Qt Designer
KDevelop
Quanta
K3b
And of course, Qt/KDE has pretty close equivalents for other things, like Kontact, Konqueror, etc.
However, Qt is still the better toolkit. Its got native versions on Mac, Windows, and UNIX. Its got better development tools (Designer vs Glade). Its faster. Its got an integrated API for stuff like XML parsing, DB handling, etc. Its got better documentation. Its got built-in scripting support (QSA).
From a programming perspective what is dumb is to use a c lib to do something that is a classic example of what oop shoud be used for and anyone who thinks otherwise should be shot in head (just kidding). That said i must note that from an end-user perspective it only matters if it works or not.
And PLEASE ! include a "kepp on top" feature in metacity.
If u think, that's tooo geaky (and I really DON'T think so), than include an window for that in the control panel.
May I make a suggestion? If your window manager is configured to not pop the window to front when you click in it then the need for 'keep on top' and 'keep on bottom' options disappear. Most window managers can be configured to bring the window to front or send to back (if already at front) when clicking the title bar. This way the windows keep their position unless you intentionally move them, and this feels so much more intuitive than specific 'keep' options for the windows. Saying that, I'm not sure how well metacity handles this (last I tried it didn't handle it well with sloppy focus, but things may have changed since then.) Anyway, I'm sure it isn't for everyone but if you try it you might find you like it. Personally I think it should be the default behavior instead of copying Windows.
Hi
I want that gnome runs with the gdk-backend directfb
Yeah A Bit Like Be. Especially the navigation part of it.
Just Right click and select the app/Folder!!
If I have to click on ten folders to find a file, then that is a pain in the neck
Not everyone uses Gnome (despite its corporate backing by RH, Sun, etc). But a KDE wishlist would be nice as well... perhaps for 3.3? My first one would be new icons for KDE's APM. KDE has some nice eye-candy and some great themes. Even Lindows made some improvements in their latest release (Laptop Edition), but the APM icons look like a 7 year old with MS Paint making 8 bit bitmaps. It needs some polishing.
Also, and this is sorely lacking in all Linux GUIs (I know, I know... window manager). All Linux GUIs need a working, quality, graphical VPN client.
Actually, significantly more Linux users use KDE than GNOME. Although, that number might have changed with the Sun/China deal.
While I'll agree that most of the better applications are made with GTK+, there are several killer apps for Qt:
Qt Designer
KDevelop
Quanta
K3b
And of course, Qt/KDE has pretty close equivalents for other things, like Kontact, Konqueror, etc.
However, Qt is still the better toolkit. Its got native versions on Mac, Windows, and UNIX. Its got better development tools (Designer vs Glade). Its faster. Its got an integrated API for stuff like XML parsing, DB handling, etc. Its got better documentation. Its got built-in scripting support (QSA).
Qt Designer and KDevelop are not common user applications, those are developers applications I would count it like a killer app, about Quanta I dont yhink on quanta like a killer application specially having blue fish in GNOME, Kontac is an Evolucion wannabe and wont come with the exchange plugin that wont help and yet Evolution is and will be better than Kontact.
Konqueror is to bloated and I have the same efect witk Epiphany, no a killer application eather and btw, I dont spend 90% of my time in the open file dialog so Nautilus work fine.
And Glade, well, Glade can create the graphic interface dor C, C++, PERL, Python, etc. something qt designer can't.
K3B I've never used I do all my burning typing burn:/// in nautilus and droping the files there, simple.
Oh, and about qt being better enviromet to export ro mac o Windows, met me tell you GTK applications are already there wit mac-gimp, abiword, sodipodi etc.
KDE still have a long way to go in applications development.
There is really no way how you could say that with confidence. Website polls don't count and vary greatly in results anyway.
Will be great that Mozilla, Galeon, Epiphany and may be Konkeror have the same BookMark database, may be with the use of a subsystem to share o common bookmark management.
There is really no way how you could say that with confidence. Website polls don't count and vary greatly in results anyway.
Actually is a well known fact that there are around 200 Linux distros and from those, only a small number sets GNOME as default DE. Remember that some of the most successful commercial distros like Mandrake, Lindows, Xandros and the likes have chosen KDE as default DE.
However, being RedHat the major player in the Linux world, should be perfectly understandable the industry support that GNOME have been gaining. But even with the recent deal between Sun and the chinese government, I don´t think that this "status quo" has changed, although it can change in the foreseeable future.
-- DeadFish Man
I see, I thought you were referring to the two versions of gtkhtml in use.
Yes, it would be nice if Evolution used Gecko, however given how few people use Konqueror (at least according to all the statistics I've seen) as their web browser, you could equally argue that KDE has two web rendering engines.
COM and KParts are entirely different. The COM vTable format might be the same as what the VC++ compiler spits out, however COM is definately language neutral. From the variants API, to IDispatch based late binding, COM is designed for language neutrality. The COM ABI is very simple. The C++ ABI is the most complex ever seen and is basically impossible to use unless you are a C++ compiler. Now, you could say that the cost of making it KDE and C++ specific justifies the benefit of having it widely used within KDE, I don't know. Personally I don't dig KDE or GNOME specific technologies, they are not usable by somebody like me who wishes to remain desktop neutral.
Finally, I'd ask why MAS is more "desktop neutral" than GStreamer. Is it because GStreamer has a G in the name? Can you give concrete examples of why MAS would be a better choice than GStreamer that don't involve politics?
KHTML + a slim GTK2 interface (similar to epiphany) would be quite nice IMO.
Trev
Actually is a well known fact that there are around 200 Linux distros and from those, only a small number sets GNOME as default DE. Remember that some of the most successful commercial distros like Mandrake, Lindows, Xandros and the likes have chosen KDE as default DE.
However, being RedHat the major player in the Linux world, should be perfectly understandable the industry support that GNOME have been gaining. But even with the recent deal between Sun and the chinese government, I don´t think that this "status quo" has changed, although it can change in the foreseeable future.
-- DeadFish Man
Well, you thing wrong, certainly most distros come with KDE but the GNOME is on the big leagues with RH, Sun Java and Novell using XIMIAN, the true is GNOME number of user is growing everyday and is gettin more support from the big companies.
The COM ABI *is* the C++ ABI. The COM ABI is defined as the structure of the v-tables spit out by the Visual C++ compiler. The C++ ABI is complex at points, but both KParts and COM only use the v-table portion of the ABI. I don't see how it is "completely different" from KParts. And its not impossible to use, because a number of different language bindings are available for KParts. PyQt and PyKDE, KJSEmbed, and the KDE Java binding offer pretty much full access to the API. More generally, libraries like Boost.Python provide nearly transparent for two-way integration with C++ code.
There is a cost/benefit trade-off associated with KDE or GNOME specific technologies, but ultimately, they are better for the user. If you can live with Mozilla not integrating with OpenOffice, or bundling disparate apps like Gnumeric and AbiWord and calling it "GNOME Office", and the main messaging client (GAIM) not even being a GNOME application, that's fine. That's why GNOME and KDE exit, anyway, to provide choice. But some people feel that integration is important, and KDE offers users that choice.
As for MAS, its very simple. GStreamer depends heavily on glib. aRts (optionally) uses some glib utility routines, but GStreamer is fundamentally tied to glib and its object model. And say what you want about politics, the fact of the matter is that politics can get a project killed. Hell, it was politics that started GNOME, so don't talk about politics like you can separate them from technical reality.
...how does it produce such high quality applications like The GIMP, Totem, Gaim and Evolution to mention a few? Some might go as far as stating that the looks, polish and finesse of many of these GTK+ applications supersede anything found in the qt sphere. Yet, my opponents still argue that GTK+ sucks. Oh, and glade does too.
I am sometimes tempted to request that these individuals who say Glade and GTK+ suck show me an application or example code they have written, and demonstrate to me how either of these quality tools have contributed to making their applications or code suck. Perhaps, I'm asking too much.
Then when the naysayers say that the language, C, was just not meant to be an object oriented framework, I point to them the GNOME platform and numerous GNOME applications successfully written with the language, C, GTK+ and sometimes Glade.
If the naysayers were right, GNOME , several GNOME applications and GTK+ applications would be failures. But, alas, GNOME, several GNOME applications as well as numerous GTK+ applications are being adopted on a large scaled by eminent entities around the globe. Yet, GTK+ and Glade sucks.
I have used both Qt and GTK+. Qt is arguably better and easier to use. But it is clearly over hyped. At least when compared to GTK+. In the end, it all boils down to where you like using C or C++ to design your graphical user interface. The fad is that object oriented languages are more powerful in producing graphical user interfaces.(I don't believe that either.)
So among the new age developers or win32 developers, C++/Qt is more attractive, hence GTK+/Glade just sucks.
I see a bunch of people simply writing 'there are better apps written in GTK'. Now I would like to know on what technical aspects their opinion is based on ? Ok The GIMP is a very good example sure but Evolution ?
Let's view the things from a different point. Say Evolution for example, the Menu and Toolbar ain't following the GNOME HIG at all while it is a GNOME related program. Interoperability is not existing e.g. Nautilus has it's own Bookmarks system, Galeon has it's own system, Epiphany has it's own system, GThumb ... and so on. There is no central repository for bookmarks. Addressbooks are not exchangable with other applications on the GNOME framework. I can't take an Evolution address and use it in other applications. I know GAIM has the possibility to do this since a few days (due the bounty) but the implementation is only limited to GAIM here and thus not adaptable to other programs. While the programs do show some aesthetical pro's their underlaying layer is not extendable or re-usable. They are simply standalone programs.
See KDE for example. There is one core repository for Bookmarks which were shared between Konqueror and other applications. They link FTP, Normal URI's, Pictures and whatever and keep it all in one common tree. A lot of simple re-use of components make it possible due to Objects that I can re-use.
The sentence that there are absolutely no good applications for KDE is plain wrong which I can not share. If we look close then we see more good applications showing up for KDE than for GNOME. Applications that seamless embedd into the Desktop due to great framework (It's impossible to code wrong here because of the framework). Applications for 2d/3d modelling which is not available for GNOME, great Astronomy/Astrology programs for students and scientists. A really usable UML program for the same userbase. It's easy to write new applications in a short amount of time depending on the needs. I for my own think that KParts is far better from the system than what GNOME has to offer. While every new app installed on KDE offers a good plugin (which can be used by other applications e.g. extends Konqueror with new features) the GNOME people play with shitty Views which you mostly not know whether they exist or not until someone tells you.
People simply say Abiword is better than KWord and Gnumeric is better than KSpread. Maybe this is true, maybe not. It all depends on proper testing and verification imo. Say even if Abiword or Gnumeric are better then you need to think further here. Abiword is written by a company and the technical things usually done in it that makes it GNOME conform are only GNOME and Unix specific the other parts are written by other people. A team of 5-6 people in it's best. KWord, KSpread and things like that should be seen as KOffice (as one suite) where only 2-3 people work on. But they on the otherhand work on 6-7 apps at a time.
I think that people should really spent into reading some tutorials and documentations around the KDE and GNOME framework and should so some minor coding to understand all this.
Look GStreamer for example, it exist for 4 years now but it's still not stable enough and not well implemented enough in GNOME everything around GNOME is half, the framework is half, partially broken.
1.- I think that people should really spent into reading some tutorials and documentations around the KDE and GNOME framework and should so some minor coding to understand all this.
2.- Look GStreamer for example, it exist for 4 years now but it's still not stable enough and not well implemented enough in GNOME everything around GNOME is half, the framework is half, partially broken.
1.- I've read the tutorials, yet qt applicatiosn are not to functional.
2- Sorry, but you are wrong, I don't have any problems with gstreamer and your words about evolution, qhy don't you ask the people who is using it and are productive with it in a office?
You have no base in your words.
3.-Applications for 2d/3d modelling which is not available for GNOME, great Astronomy/Astrology programs for students and scientists.
3.- So does GTK, www.gnome.org and clik on the software map link, surprised?
You write big words here buddy. Can you explain 'technically' why you think GNOME is better than KDE ?
Well your corporate interests is an interesting aspect. RH, Ximian, Novell but these companies are not doing the majority of work around GNOME (at least it was so). When I think 4 years back then GNOME was a funny community with people who spent all their free time into GNOME until Ximian and SUN made some mess out of it by going into the commercial directions. Companies like Ximian and SUN show up on all place and make lousy decisions where it will go in the next couple of years obviously leaving big questionmarks of the volunteers working on it. Announcements like 'MONO soon in GNOME' or 'we will make GNOME a corporate Desktop' doesn't sound motivating to people either. Even after Novell has bought Ximian it doesn't mean that Novell automatically became the legitimate owner of GNOME. It's not like they come up throw all the people out and take over leadership and maintainance. They have their own interests and it should stay theirs.
Making GNOME commerical is more the interests of 2-3 companies but not the interests of the people. Many people I know about just want a cool Desktop that operates well enough, where they can hack the one or other application where they can feel like they were part of it. When I want to work on a GNOME project then I want to feel like I'm part of that project, like I'm someone who contribute to it like equal partners. When I listen to announcements made by Miguel de Icaza or Nat Friedman then I get the feeling like being the dumb developer working for their moneybags. Is the community really growing ? I can't say - I usually do not count the people working on it and the people who join/leave it.
Some weeks ago I had an interesting conversation with Kandalf (Ralf Nolden) who joined us in the #gnome-de IRC channel and he told me that KDE tries to offer total freedom to the user and developer and thus they are trying to keep companies outside of the project. They do accept money as well as moral support from them e.g. SuSE or The Kompany but they are usually threatened as equal partners and don't give people the feeling like being ripped off. At least they do not need to show up on all places and make promises they can not keep.
Let's look at Novell who bought Ximian. Can anyone (maybe you) explain what Novell actually bought ? Did they buy GNOME ? NO !, did they buy MONO ? NO !. They only bought the brand Ximian, the employees but no properties. I also belive that people shouldn't expect to much from SUN, Ximian/Novell here since they are just contributing. They contribute and do not own and I want it to be like this for the next upcoming 20 years. I want to have the feeling to work for a free project a project driven by the community, a project which should be professional - not a project where developers and users get the feeling that 'now it starts to get interesting' to be ripped off.
Im not talking about the desktop Im talking about GTK appications, can't you read?.
And why do you think those companies are spending money on GNOME and XIMIAN, that's right, cause they have a better licence and a better desktop.
Yeah, som guy trying Linux is making pretty his KDE desktop meanwhile big companies are developing GNOME to make it better even better.
FYI I am in the GNOME community. I know the apps. I know what exists. I know the framework. I know the people.
Your reply for me not having a base of argumentation is nothing more than ignorance. To help GNOME become a better plattform we need to discuss about it's problems and not ignore it.
FYI I'm receiving more mails written with MUTT or PINE from all sorts of gnome related mailinglists than mails written with Evolution. But ok. Please tell me how many companies seriously use Evolution. I would be thankful to get some numbers here.
Spreading the 'word' that a lot of companies use it until the last one belives it is creating opinions. Showing some numbers that verifies it is more realistic. Say from 10 mio Linux users there are 5 mio using Evolution which means 50% of the Linux community uses Evo and the other 50% uses Mutt, Balsa, Pine, Sylpheed, MozillaMail etc.
I now tell you that I do belive that more companies simply use Microsoft Outlook or Mozilla Mail for doing their Mails rather than using Evolution. I can't confirm it but so can't you with Evo.
BTW, your words have no base at all, you just bash GNOME w/o proves, let me guess, do you have GIMP in your KDE desktop?, Evolution maybe? some other GTK professional applications? that can be used in a enterprice?
Actually, Im just showing the features KDE user don't know about GNOME and keep bashin it.
Read my post, IM just answering their questions.
Yes, I remark how GNOME is getting lots of supprot, so does the KDE communiy remarking KDE succes, there's nothing worng with it.
Your reply for me not having a base of argumentation is nothing more than ignorance. To help GNOME become a better plattform we need to discuss about it's problems and not ignore it.
Actually that's true, so should do the KDE community instead looking GNOME bugs.
now tell you that I do belive that more companies simply use Microsoft Outlook or Mozilla Mail for doing their Mails rather than using Evolution. I can't confirm it but so can't you with Evo.
And I bet they use less KMAIL, IM talking about companies using Linux not Windows.
I'm not saying that GTK+ sucks and you can't make good apps with it. Its a very nice toolkit, and its way nicer than some of the other stuff that's available. But Qt just has the edge in a lot of ways. Qt doesn't totally crush GTK+, rather, its slightly ahead in a race of two well-matched competitors.
In general, there are more GTK+/GNOME developers than KDE developers. The powerful KDE framework gives KDE developers an edge, in that its easy to throw something functional together in a short time, but that's also the reason why a lot of KDE apps look and feel unpolished. I'd wager a lot more man-hours have gone into Gnumeric vs, say, KSpread.
I think we should increase the quality of exchanging replies to each other here without giving the readers the idea of being immature kids.
Let's not stick on one application here. See the entire whole of a Desktop as is. A good plugins system, good interoperability across applications, cleaness, documentation, api documentation, easy to write new apps, not stuffed full of language bindings in it's core (e.g. make it an requirement). Apps that simply work. Apps that people know how they look and feel from other plattforms (offering similar look and feel). Apps that can be extended easily due to good framework. Apps that stay as is for quite some longer time and being made stable and solid and not changed over and over again because the developer feel bored over the weekend.
There are a bunch of wrong approaches in GNOME e.g. glueing OOo and Mozilla (GRE) into the core and there are a bunch of applications which are left unmaintained for long. E.g. no Presentation app (Criawips as well as Agnubis) are basically unmaintained. Dia is far from even following GNOME or GTK+ (agreements, styles, HIG or whatever) not to mention that it's useless for professional work. Look during my time as student I urgently needed an open source UML creating programm or some program to do DIN 60001 (?) well standard IT diagramms and DIA was quite unusable for this task. Saved files couldn't be loaded anymore, Font rendering always broken, LibXSL path was hardcoded and serious help (which I gave them) was ignored. Now what about the apps that simply work ? I need programs that can be seriously trusted. Who somehow work reliable. Reliability is the keyword here. You should be able to trust the program blindly. When I open Nautilus and copy a directory from FTP to my system then I must trust this program blindly. When I browse my files and by mistake click up or another view then the program should not crash.. Things like this. I do not care whether it's KDE or GNOME. This is a general thing that I wish from KDE and GNOME.
I want when I do a bookmark with GThumb that this bookmark is being shown in Nautilus as well so I can use Nauti to browse into that directory. When I browse with Nauti and create a bookmark to a .html file that I often need to read then I want this bookmark to be re-usable in Epiphany or Galeon so when I start them that I can load the page from within there.
When I write some addresses into Evolution then I would really love to use them in Abiword when writting a letter to a friend (like Include address here.. but I must admit that I rarely use Abiword so I don't know whether it works or not).
Hope you understand what I am saying here.
Let's forget the childish behavior with KDE and GNOME for example but I do see a bunch of KDE applications offering these keyfeatures. But asking the question "why does it offer these features or how can it offer it" then the answer is "due to it's framework".
Let's not stick on one application here. See the entire whole of a Desktop as is. A good plugins system, good interoperability across applications, cleaness, documentation
I spend much time writing about gtk applications as you just talk about the KDE desktop, remember, the desktop aint nothing w/o the appications.
Killer applications: Is there something like 'a killer app' ? It also depends on the needs and requirements of the user.
But well I do prefer Konqueror over Nautilus here because it simply works and I do love the KParts system over the GNOME views.
I do like Umbrello more than DIA. At least it saved me a lot of time when I urgently needed it.
I do like The GIMP because it's a cool application.
But maybe you realize that I speak more about the framework rather than applications itself. What I try to say is that a good framework offers you easy application writing, better integration a more unified look.
Example: GNOME is cleaning up a lot of it's previous made mistakes e.g. DEPRECATING a lot of old junk from libgnomeui, libbonoboui, GTK and so on (to name all them will take a lot of time). Now think about the people who develop programs, it's overkill to change all the functions to suit the new needs. We are basically stuck in permanently adjusting applications to a new API rather than making programs rockstable and go on writing new apps once something became rockstable. More deep explaination: GtkCombo differs totally from GtkComboBox while the API is frozen for the 2.x series it still means that you need to change to the new API sooner or later. You are forced to this somehow not just because of new API because everyone is doing so. And I can tell you that the GtkCombo totally differs from API to GtkComboBox. While you were able to add a GList to GtkCombo you now need to create a Treemodell for GtkComboBox which means that under bad circumstances you need a lot of time only to adjust these things. Ok DEPRECATING old things is good and I am thankful that this is happening (somehow) it finally shows that GNOME is getting rid of previously made mistakes. But we need to protect new developers from using DEPRECATED things by telling them before what they should not use. Such a documentation/tutorial is totally missing. People who still use GNOME 2.4 and started writing new applications could by mistake use things which are marked as DEPRECATED in 2.5 we need to protect them from going the painful way of makeing huge changes.
Well this is only the pit of the Mountain from various examples. If you repeat that GNOME is so cool by denying the problems it has then how do you think such problems can be fixed ? I think speaking about GNOME adding wishes to it makes a lot of sense and Eugenia did it right here. But this shouldn't exclude the fact that we need to speak about the problems as well. The pro's and con's.
Againg you are talking as a developer, final users don't mess with APIS just developers, Im a user and not a developer I don't care about APIS, I just care to have good applications to do my work and I found that on GTK applicatios, so does the GTK developers that's why they KEEP working on GTK APIS.
Actually I could drop GNOME for XFACE and I would be happy using my GTK applications there. Dont talk about framerwork.
What about this do you not understand? How good the toolkits are has little to do with which gets the "killer applications." The OSS community has a lot more C developers than C++ developers, and there were a lot of concerns over licensing before Qt became GPL, so GTK+ got a lot of initial momentum. That doesn't have anything to do with how good the toolkits are, but concerns outside the realm of the technical. I've mentioned the features in which Qt is better than GTK+. GTK+ has the edge over Qt in some respects (accessibility via ATK and internationalization via Pango), but on the balance, Qt is a better toolkit. Not vastly so, but noticably so.
* To be fair, there are a lot of excellent Qt/KDE applications (Kate, Konq, Kopete, KDevelop, Quanta, KOffice, Kontact, etc). They may not all be best-of-class, but they're very good.
Ok I do accept your statement of just being a user and no need to deal with API's. But now please tell me on what facts you base your opinion that GTK+ apps are generally better ?
Example: to know whether an Apple Macintosh is better than an AMD/Intel based system you need to know the technical aspects about it.
to know whether a yellow Bananna tastes better than a green one you need to know the technical aspect of things being ripe.
to know whether you get drunk from drinking coke or drinking vodka you need to know the technical aspect of alcoholics.
Same with Desktops and applications. To know whether a Desktop is good or not you need to be at least skilled enough to understand some basics about it. Everything else is just an opinion and basing opinions on fancy icons or because everyone else told you is plain wrong.
Example: to know whether an Apple Macintosh is better than an AMD/Intel based system you need to know the technical aspects about it.
to know whether a yellow Bananna tastes better than a green one you need to know the technical aspect of things being ripe.
to know whether you get drunk from drinking coke or drinking vodka you need to know the technical aspect of alcoholics.
Sorry but... is that a joke? Since the prehistoric ages, your average caveman knew rotten meat was worse than fresh meat. Those statements make no sense.
The same way I know a car runs faster than another without reading its technical specs (just driving), I can know if an app is better than another just using it. Or do you think I need to know the implementation details of GIMP and Paint to know why one is better than the other?
At least you know it's a painting program that you like to use. The GIMP is good as well as bad together.
One person use The GIMP because he knows what he is doing due to having spent some time into that program, reading some documentations and stuff like that and then says that it's a great program.
Other people can say it's a bad program because it's full of things they don't know about and thus do not figure out howto paint a simple picture.
It all depends on the needs of one. Some people may find Paint better than The GIMP depending on their needs and their knowledge what they like to do with that program. But what does it technically tell us about both programs ? Nothing !
It all depends on the needs of one. Some people may find Paint better than The GIMP depending on their needs and their knowledge what they like to do with that program. But what does it technically tell us about both programs ? Nothing !
That's the point. An end user won't care about underlying structure and technicalities. Even if a program is a messy visual basic project, as long as it works and fits his needs, it's good for him and better than the rest. And that's what matters in desktops for your average user.
Of course, that's not an excuse to improve the underlying code to be better not losing features, but you get the point
The integrated framework and the power of the APIs isn't just a developer advantage, its a user-visible advantage. Let me give you some personal examples:
1) All KDE apps automagically use KIO. Recently, I needed to do a website for a project. I just opened up Kate (a regular programmer's editor) and edited the files directly on the remote server. No need to "upload" or "synchronize" or anything like that. I needed to make some screenshots to put on the page. Again, no problem: I saved them directly to the server from ksnapshot. When I need to print a file from my dorm desktop to our central campus printers, there is no need for me to upload it to my network account, I can save it directly from KWord. GNOME has GNOME-VFS, but not many apps use it.
2) All KDE apps use a unified configuration mechanism. If I need to make a configuration change directly, I know that all the rc files are in a known place. GNOME has this know, but a lot of GNOME apps still don't use gconf for everything.
3) The QT/KDE theming mechanism separates color schemes from widget styles. No editing gtkrc manually to set something simple like theme colors!
4) All KDE apps use a centralized toolbar mechanism This makes all KDE toolbars fully configurable, using the same configuration dialogs. All KDE apps also use a centralized acceleration handling mechanism, and there is a similarly standard configuration dialog for those.
5) KDE apps are heavily componentized. That means that Konqueror can embed pretty much embed any media type. Nautilus has this capability through Bonobo, but there are far fewer componentized apps on GNOME, so there are fewer things available for Nautilus to embed.
6) Most KDE apps are DCOP scriptable. KJS offers very powerful scripting that can access the whole KDE API. This is for more advanced users, but there are a lot of those on Linux. Even less advanced users can use scripts made by others.
7) There are lots of centralized services that leverage the framework, like kwallet, and khotkeys. Some GNOME apps have mouse gestures, but with khotkeys, *all* KDE apps get mouse gestures. And thanks to DCOP, there are some *extremely* powerful things you can do in response to a hotkey or gesture invocation.
Lastly, don't forget that the framework speeds up application development and makes for cooler applications. If KHTML wasn't so easy to embed, so many KDE apps wouldn't use it to display HTML. KDE doesn't have as much developer manpower as the the GNOME community, and so this advantage makes for applications that are a lot better than they would be.
Sorry You're right, though, Scribus is nifty.
Scroll down for my reply to Ross' blog regarding my article:
http://www.burtonini.com/cgi/pyblosxom.cgi/computers/gnome-wishlist...
You say that GTK+ is better than Qt simply because of "killer apps"?
1) Killer apps usually mean applications that cause people en masse to get out, buy a new computer just to use it. One of the first killer apps for the PC is spreadsheets. With Lotus 1-2-3 on the market, thousands of businesses went out to buy a new computer for their clerk/secretary/accountant to do spreadsheets. The most recent one is Internet - hundreds of thousands went out and bought new PCs just to surf the Net.
So far, I have yet to see a single killer app for Linux in general, no less GTK+.
2) As for better apps, it is very subjective. I much prefer KMail to Evolution, simply because KMail only does mail, yet it integrates so seemlessly with other apps, like the addressbook, etc. As for GIMP, it was GIMP that started GTK (thus the name GIMP Tool Kit), it is kinda obvious they would use it until today. As for GAIM - how is it better than Kopete? Using them both, I would prefer to use Kopete (i.e. I can take addresses from a central location, but in GAIM I can't take addresses from Evolution).
Windows has far more applications than any other platform, does it mean that Win32 is better that say, Cocoa, or Qt, or GTK+ or BeAPI, or heck, even .NET? Nope, it is crappy, hard to learn, buggy, full of holes, and at times unstable. Nobody with a sane mind would even suggest that Win32 is better than anything else competing with it. The same case could be applied between GTK+ and Qt.
Besides, you only mentioned open source applications. What about Opera? It is certainly better than Epiphany, Galeon. Photoshop Album (unavailable on Linux)? mokey Imagineer? Raindrop Geomagic? Is there any GTK+ equilevent to those? What about Mainconcept? While hardly being the best video editing application, it certainly blows any GTK+ app doing the same. Since you said you are just comparing applications using the toolkit, how many GTK+ applications do you know that can run out of the box on a Zaurus, or one of those Motorola smartphones?
7) There are lots of centralized services that leverage the framework, like kwallet, and khotkeys. Some GNOME apps have mouse gestures, but with khotkeys, *all* KDE apps get mouse gestures. And thanks to DCOP, there are some *extremely* powerful things you can do in response to a hotkey or gesture invocation.
Some Gnome applications have mouse gestures? Didn't know about that... do you have an example; how do i enable it?
Thanks
Victor
I feel the author did not do their homework on this article. If they would have asked the Coaster team what was going on, they would find out that we have moved libburn to freedesktop.org and now just need to stabilize it, then the gui can really take off. So...project dead?...Not even close. I do know the site is horribly out of date, and I have been working on the updated version for some time. Being in school, sometimes time becomes nil. Rest assured that linuxland will see a proper update on the project sometime this week as well as an updated website.
So how do you solve some of these and other problems mentioned in the article. There are two ways. File bugs and/or send patches/updates. There is no other way these will get solved in a timely manner. Same goes to any UI complaints. File a bug or send a patch with the new design that is better. This is the way OSS works.
> Copy/Paste still misbehaves after all these years
> Copy/paste still just doesn't work properly. I cut or copy
> files off my desktop and try to paste them on another
> Nautilus window and that would only work half the time.
Oh so true. The solution (as I learnt on bugzilla) is to press the update button, then you will be given the option to paste. The same goes for associating an application with a file type. Press update, first, then the newly associated app will be available. Stupid, isn't it?
I must honestly confess that the reason I keep using gnome is that it looks cooler than KDE and it has more native apps (gimp et al).
Although I agree with some of your points about kde/qt's architecture, I want to bring up my reasons for strongly preferring gtk as a toolkit.
The main reason why I consider gtk to be a better toolkit mainly relates to CORBA. I don't understand why developers want a reinvention of COBRA. I don't really think DCOP or DBUS seems like a great idea. Perhaps I'm just naive; I haven't used it a great deal, but I still think CORBA seems like the way to go, even if its a bit less practical.
Honestly, I don't think I've ever seen a toolkit *ever* before that used CORBA. Every other toolkit I've seen is similar to kde. Every other toolkit is monolithic, in that the toolkit is designed for use of only one language. Also,every other toolkit that I've seen uses a lightweight reinvention of CORBA.
Perhaps I'm wrong, but when I develop programs I want to use functional languages exclusively. In gtk+, the ocaml bindings are usable. Scheme bindings are available. To me that is the ultimate luxury.
I agree with you about gtk+ api being ugly. It should be integrated like qt is, and I'm not sure if using C was a good or bad idea. Gtk should have excellent documentation like qt does. These two advantages of qt are a result of trolltech's fantastic work, and are likely to remain an advantage over gtk well, for a long long time.
--Tim
What about Eclipse-GTK? I can argue it is the best general purpose development environment available on Linux.
What about Gnumeric? Definitely better than KSpread. Arguable better than oocalc nad Even the venerable Excel.
And people should stop saying things like GIMP is hard because you can not do a simple paint image on it. Heck, I think Photoshop is hard too.
I think things like "You can embed anything into anything" are not necessarily all that. Only geeks use those features. Most people I know do NOT embed spreadsheets in Word (on windows). They just copy and paste into a table.
Most people do not deal so much with ftp, and transparency over ftp does not help them. For web developers that would probably be a killer feature, but even then, maybe for a few only.
Heck, on Windows, Outlook express and outlook do not seem to even share an addressbook. Outlook actually imports the settings from there. As a user, I get turned off more by ugly apps like emacs (yes I am not a developer, so its not of much use to me) and I think for my purposes Gedit is better. It looks beter and works rather predictably too.
I have never seen the need to preview spreadsheet files in the file manager. Pictures, yes, text documents yes, but not spreadsheets, or whatever people decide to preview next. I prefer to have a multimedia player that just works, and works well. On Windows I really hate windows media player, and I prefer to use zoom player for my video, because it is simple and does what it does extremely well. I prefer the GNOME approach of having good apps that work well alone, and have reasonable interoperability with others.
Editing toolbars to me is something I reserve for apps like Gnumeric, or Excel is windows, where I use functions very different from the defaults. I think if your web browser forces you to change the toolbar, then it has a serious problem. The only 'editing' I do is to remove the text from under the buttons if its there, and maybe keep 'priority' text on the sides of the buttons. But that is me only.
I think what is needed is to provide a platform on which developers can bring what they want to bring to the table. I agree with some of Eugenia's list, but I think some of those things are not really as pertinent as she makes it seem at times.
For instance, the issue about plugins for totem. The real problem is that people are using wmv as a method of delivery, when Microsoft will not make a Linux codec. Blaming it on GNOME will not solve anything. Besides, Mplayer does it already, and Gstreamer did nto have these capabilities until recently, and its not for inclusion into GNOME until 2.6 anyway.
I see a theme in the responses here. Hackers take the critics to heart, and scream "ignorant!" for things that one shouldn't have to be obliged to know.
We users ARE bitches about the software, it's sad, but it's the true, the same as you hackers are bitches about your programs. We scream because we can't help because your programs are too difficult to hack on. Why wouldn't you take our criticism and improve the system? When you're finished with us, you'll have a system made for users that'll make it the preferred desktop ever. It probably won't be the perfect API, but at least your users will be happy until developers beautify the code and won't have to bitch again and again, until they give up and use another desktop (as I did).
The first thing that I do when I fire up GNOME (or KDE) is to look for the terminal window icon so that I can get some real work done.
If this gets moderated down, then the moderator really doesn't understand what I'm trying to say...
The biggest wish is to have better integration between GNOME/KDE. How about a common:
- Multimedia sound subsystem (GStreamer?)
- Sound Server (artsd or MAS?)
- MIME Types (freedesktop mime types in progress)
- Hardware Abstraction Layer (freedesktop HAL)
- Messaging interface (freedesktop DBUS)
- Menus!!!! (???)
- Themes!!!! (???)
Also, lets get some better low level integration in Gnome/KDE. A common way to handle IM's, contacts, etc. would allow integration between any Evolution & Gaim or Balsa and Jabber.. rather than hacking Evolution & Gaim.
GConf causes me more headaches than anything else in GNOME. It is a good idea, let down by bad implementation. It really hurts bad if your home directory is on NFS. The only way I get get it to work reliably is by forcably killing gconfd and removing its log files before logging in. Even this doesn't always work if, like me, you routinely log into several different machines at once.
I'd love a proper DB backend. It would be a massive boon for system administrators - changing gconf keys for a network of thousands of machines would be trvial, and no more lock file problems!
i agree with this article; i am a gnome fan, it seems more professional and i like it, but i notice that on my desktop there are more kde apps (kppp, k3b, kate...) than gnome ones...
So you're saying that having powerful features is useless because most users will never get beyond the lowest common denominator that they are used to in Windows? Great logic, just great...
Plus, I disagree with most of what you say. Take, for example, the transparency issue. In a heavily networked age, its not unconcievable that many people might want to make a personal web page. Heck, I know a lot of people who have a little family page to share stuff like wedding photos. Isn't it a whole lot easier if they could publish their web page just by saving a file to a place that looks just like a directory, rather than figuring out what the hell an FTP client is? What about the legions of people who need to transparently access network resources in an office or school environment. Isn't it easier for them if they just have to learn about files and directories, instead of thinking about the distinction between local and network resources?
Another example: my little brother and all his friends use AIM heavily to talk to each other. He's got about 100 AIM contacts to manage. Meanwhile, my dad has over 1000 business contacts. Wouldn't it be easier for them if they had contacts just in one place, instead of spread throughout the system?
Even conveniences such as toolbar editing and embedding a re used by more people than you'd think. The average home user might not use these features, but office workers often get at least a couple of days of Office training, and they definately use these features to make their life easier.
@Aeonsfx: Have you used CORBA? Using CORBA to do IPC is like using a buick to swat a fly. Total overkill. CORBA is an incredibly complex and cumbersome API, especially from C++. CORBA is a dinosaur that really needs to die.
one thing that imho needs big improvement
is color prefs management.
The default black on white background is
very bad for my eyes (and probably not only
mine) and changing it is -aehm- nontrivial.
No matter what I do to my .gtkrc* or gnome
control center many text areas will still
display black on white - rendering affected
programms essentially useless for me.
There may exist some theme that would provide
decent colors but downloading and trying
a few dozens themes is not really feasible
to solve such a simple problem.
Both KDE and old style X/Xtk .Xdefault files
work much better in my experience