OSNews: Gnome Guy Goes KDE

Christian Paratschek is a long term Gnome user who has has not looked at KDE for over two years. He has written about his experiences of testing out SimplyMepis for a week, comparing the two desktop environments and their applications and finding the areas in which KDE can tempt even a dyed-in-the-wool Gnome fan.


"Why do we have bookmarks in the desktops context-menu?"

Maybe because they're frequently accessed from there? If I want to go to kde-look.org, for example, I can quickly select it from the context-menu bookmarks. Otherwise I must focus, restore, display, or perhaps even start the execution of Konqueror. While some may say this is a nonsense scenario, I find myself doing it all the time. It's an incredibly useful shortcut.

By Brandybuck at Wed, 2005/01/05 - 6:00am

Or you click the K-Menu and choose "Bookmarks" from there.

I also don't understand why Bookmarks belong to the Desktop context menu. It has nothing to do with the desktop. I don't create item on the desktop with it nor do I change its properties with it. And the desktop is often not even visible, while kicker is usually visible.
We could just as well add the complete k-menu to the desktop menu, or did I miss a fundamental difference here?

By uddw at Thu, 2005/01/06 - 6:00am

I have used gnome and for me I got very annoyed with it very quickly. One of the major issues I had was no url bar in the open and save file dialog boxes. I know this may not be really common for other people but about half the files I work with on a daily basis are remote resources and in kde I can work with all of them transparently from ANY kde application. Even better is that the io slaves integrate seamlessly with kwallet so I can have it remember usernames and passwords for these resources that are on remote servers and it makes my life a lot easier. I suspect that io slaves and being able to use them so easily from the open and save dialog box saves me a few hours/week of time.

I ran into a similar issue when trying to use nautilus. My directories are mostly deep structures and spatial browsing was a major pain for me. I looked in the menus and could find no way to turn it off. I did look online and found a gconf key to turn it off but that was still a pain.

Gconf is also something I HATE from a usability standpoint. I was looking in there for some stuff and many fields had limited inputs but had a textbox to type it into. If there are only 3 valid values a field can have then give me a drop down box or something to select one of those choices. Don't make me type n a 20 character text string that must be exact for it to work.

I look through the web browser and could find no options anywhere to set the proxy settings. I never did figure out where that one once mostly by this point I had given up on gnome for being what I wanted to use to get my work done. Under kde it was easy for me to find proxy settings, it was even easy to figure that out in mozilla, firefox, opera, and ie.

Sure many people may not need network transparency for normal usage but some of us do and it saves a lot of time. Many people complain that an app is unusable when it has x feature that they don't use and it has a button. However those same people complain when it is missing feature y that they do want. Everyone wants something different from the system and I am sure the features I want in an audio ripping program are different then your features. The only way I rip cds is to 500kbps or so ogg files. I did not see a way to do that in gnome with the audio encoder I found.

I do get sick and tired of people talking about app x as unusable. I almost never hear them state WHY it is hard to use. Unless you state something very specific about why something makes an app hard to use, ways it could be improved etc then your comments are pretty much valueless. As you can see up above I stated very specific reasons of why I did not like using gnome.

Here is a screenshot of my desktop. http://aesaeion.com/mydesk

There are many things I think that could could do to improve things. In konqueror for instance I don't really see why cut copy and paste should be in the default bar. However just because I don't use it does not mean that others don't. What we need are real measurements. Some kind of process on kde that recorded every feature activated and how that feature was activated to see how often various features are used and how they are used. If almost nobody ever uses a given button then it probably does not have to be there but without real numbers there is no way to judge what makes something more useful for people.

By Kosh at Wed, 2005/01/05 - 6:00am

"In konqueror for instance I don't really see why cut copy and paste should be in the default bar. However just because I don't use it does not mean that others don't."

Right now, Copy/Paste is available through these:

- Menubar
- Keyboard-shortcuts
- Context-menu
- Toolbar
- Drag 'n drop

Do we really need them all? Why not remove the ones that clutter the UI the most? And I would guess that the buttons in the toolbar are the ones that do the most cluttering. If we removed the corresponding buttons from the toolbar, it would not reduce the functionality one bit. And, if the user really wants to have those buttons there, he's free to add them there.

Of course, some people probably support multiple levels of redundancy (in this case we have five levels), but I think enough is enough.

By Janne at Wed, 2005/01/05 - 6:00am

Very good idea!

- Menubar has to be kept - every (global) functionality should be present in the menubar (my opinion).
- Keyboard-shortcuts don't clutter the UI (just a short "string extension" at the menu items
- Context-menu. Ok clutters the menu, but make a lot of sense there
- Toolbar clutters UI a lot. But at least I didn't ever use it.
- D'n'D does not clutter UI at all

By birdy at Wed, 2005/01/05 - 6:00am

None of these ought to be removed.

DnD is not complex, and many advanced users appreciate it, as with the context menus (useful for fine granularity c&p) and keyboard shortcuts. To a beginning user, however, these facilities are invisible, and don't affect them adversely in any way.

Most basic users will use either the menu or the toolbar.

I think Janne is labouring under the misapprehension that cleanliness = functionality. This misunderstanding runs through many discussions of Gnome's current drive towards 'usability' (I use apostrophes to differentiate the actual word from the way some people use it). The toolbar is a very specific development in UI terms that has provided massive benefits to users for decades. It allows for fast access to core commands, and offers a nice small set of options to beginner users, who aren't able to take in/don't want all of the menu options. We use icons to make the purpose evident, and labels in some cases. Turn off the toolbar and address bar in your browser for a day, and see how much slower you operate when you have to search menus for the functions you need quickly. Many beginners don't know about Ctrl+X/C/V, context menus, or drag and drop, but they know what the icons and menu entries do. This is a good thing, and it's a major success of software that so many users have been introduced to a task that is not visually obvious.

Hence: Toolbars work. They work for a lot of people, and a lot, if not most users (by total users, skilled and unskilled) use the toolbar buttons as their primary means of operating applications; and also their primary means of copying and pasting.

If it doesn't suit you, then that is fair enough. KDE provides a very straightforward toolbar organisation tool, which you can get by right-clicking on any toolbar. Just remove the icons that you don't want. This is the power of C++ & QT which allows for functionality to be inherited like this; something that GTK doesn't do and every application has to have their own custom-coded solution. Same as the good old file selector problem in GTK.

Finally, it is also important to be aware that toolbar icons play a vital role in enabling visually-impaired people to use computers. Often, searching through menu items is far more difficult, frustrating and error-prone than for those with 20/20 (or corrected) vision. Being able to click on large, colourfully-distinctive icons allows the visually impaired to be able to use computers more readily and face one fewer barrier in the work place and in an increasingly computer-oriented world. Again, KDE provides single settings for large fonts throughout all applications, large menu fonts and large icons, which are a massive help and only need to be set once, for the entire system. Again, thanks to QT and the KDE base classes, which provide these settings to all programs.

Sidenote: It is also important to not say "Well, if it's on the toolbar, it doesn't need to be in the menu system", as screen readers for the blind are very frequently forced to only use menus, and are unable to describe toolbar icons accurately to the user.

By Luke Chatburn at Wed, 2005/01/05 - 6:00am

"DnD is not complex, and many advanced users appreciate it, as with the context menus (useful for fine granularity c&p) and keyboard shortcuts. To a beginning user, however, these facilities are invisible, and don't affect them adversely in any way."

I was suggesting that C/P-buttons could be removed from the toolbar, and the reason is that the functionality in question is already available through numerous other ways. And I have yet to see a user who doesn't know about Ctrl+x/c/v, and I work with alot of users.

"I think Janne is labouring under the misapprehension that cleanliness = functionality."

Yes and no. Clean UI makes the app more functional than cluttered UI would, in my opinion. I'm laboring with the current design-philosophy of "we must have all possible features, and we must present them in as visibly as possible". That results in UI that is cluttered and confusing. If you had to choose between cluttered UI and clean UI, which would you choose? I'm not advocating removal of features, far from it. I'm advocating cleaning up the UI. What Gnome-folks have done is to remove functionality altogether. Sure, they can make the system appear "cleaner" (I would use the word "dumber" thouhg) that way. I'm not suggesting that. I'm suggesting that we keep the existing functionality, but we emphasize the really _relevant_ functionality. if the user wants to use the more advanced functionality, he should be able to do so. But the UI should reflect the primary functionality by default. So, instead of designing the UI in such way that it prominently displays some advanced features that 99% of users never use, why not design the UI in such way that the REALLY useful features are prominently displayed, with the rarely used features are not visible by default, yet available?

A good example of this is the file-dialog (I had a discussion about this a while back in the mailinglist). Have you noticed that it doesn't have a place to enter a filename? It has an entry called "Location" (which IMO implies an URL or a directory), but no filename. Yes, you use "location" to type the filename. This thing was very confusing for my wife, and as I thought about it, I agreed with her. Do you know why it says "Location", and not "Filename" (or something similarly descriptive)? It's because you coulse use KIO-slaves to save in to remote filesystems, and it should reflect that fact. Well, that's fine and dandy, but 99.9% of users simply want to type a filename and not use some whiz-bang features. They just want to save (or load) their file! yet this thing is designed around a feature that gets rarely used, instead of the primary purpose of the app that gets used over and over again.

"KDE provides a very straightforward toolbar organisation tool, which you can get by right-clicking on any toolbar. Just remove the icons that you don't want."

So, new users should spend their time trimming down the UI, instead of using the app in question for the purpose it was designed for? This is what I meant by apps growing together with the user. Right now KDE pushes newbies right in the middle of humungous number of options, buttons, menus and features, and they should then just remove what they don't like. What it should do is to gradually guide them in it.

New users are not prepared to go through options and configuration-tools. Advanced users have no problems doing that. Advanced users can clean up the UI, but newbies are not so prepared to do the same. That is why it's better IMO to let the app grow with the user. As the user gains more knowledge, he can enable more features and change the UI. But as he first uses the system, he's presented with a clean and intuitive UI, instead of UI that has humungous context-menus, lots and lots of icons and insane number of options. Advanced users can sort it out, newbies cannot. Right now we expect newbies and advanced users alike to go through the options. With my suggestion, we would only expect advanced users to go through the options. And they are alot better prepared to do it than newbies are.

Form Follows Function. What is the primary purpose of the app? Then design the UI to reflect the primary purpose of the app. And keep the UI clean and functional, so that the primary functions are highlighted. I don't think that's unreasonable.

By Janne at Wed, 2005/01/05 - 6:00am

>So, instead of designing the UI in such way that it prominently displays some
>advanced features that 99% of users never use, why not design the UI in such
>way that the REALLY useful features are prominently displayed, with the rarely
>used features are not visible by default, yet available?

So how do you know what features 99% of people use? If you have some magical method for knowing that I would love to know it but without real numbers I have doubts that anyone has even a good clue on what people use. Until there is some way that kde can track that information and it can be submitted back for analysis the answer is that nobody really knows that.

Until that information is really known then removing visible choices in most cases is a very bad idea since you don't know the effect that has on people. The gnome people removed the text box from their file dialog box and so you can't type a url in there anymore or anything else for that matter. No matter what arguement is given it is removed for usability and they know best. It is all BS and I don't want to see that here. Until you have numbers to back up your viewpoint stuff should not be removed, renamed etc.

I have see how people have used some of the stuff I have written and it did not match how I thought it would be used at all many times and stuff I thought would be a usability issue was not and stuff that I hadn't even though of confused people and so I changed it. The point is that you don't make these decisions in an information vacuum.

By Kosh at Thu, 2005/01/06 - 6:00am

I think you are Janne are basically saying similar things, it's just that you don't agree on which features should be displayed prominently, like in the toolbars, and which should not be.
As for the copy/paste/cut buttons, I agree with Janne that they should not go on the toolbar. You can make a case, as you have, that they may be beneficial to certain users, but for the vast majority, they are either superfluous or distracting. For someone who is seriously visually impaired, just adding a few icons to the toolbars will not make the application any more usable. These people have specially designed environments or will use assistive tools so that regular menus are usable for them. (Seriously, my job is to design special interfaces like this).
Also, we can't ignore the example set by other systems and applications. I'm on Windows XP now but lets inspect the toolbars of a couple random apps I have installed on this computer.
Applications WITH the copy/paste/cut buttons in the default toolbar:
Openoffice Writer

Applications WITHOUT the copy/paste/cut buttons in the default toolbar:
Adobe Acrobat Reader
Internet Explorer
Movie Maker
Outlook Express

So of all these apps, in which copy/paste are useful functions, only one has them in the default toolbar. Only Openoffice Writer, whose primary purpose is text manipulation, has the buttons.
Why should KDE be so different? I've never heard of anyone complain of lack of toolbar buttons, the reason people don't like gnome (me included) is lack of features, not lack of buttons.

By Leo Spalteholz at Thu, 2005/01/06 - 6:00am

The problem you noted with GConf about having to type a text field where only 3 values are possible is NOT a GConf problem, but is a problem with the developer who put those options there. GConf has ways of making sure you can only select certain options, and in any case, gconf2-editor is NOT meant for editing apps by normal users, but by developers or normal users. It is just an interface which doesn't even have to be in the distro.

By Maynard at Thu, 2005/01/06 - 6:00am

Okay so if it is not meant to be used by normal users then how do I do stuff like turn off spatial mode by default? I found no gui option anywhere for it and only found instructions to use the gconf editor to do it. It seems if there are options that regular uses will need to change then there needs to be a way to change them. Saying that gconf editor is only for developers is a cop out around the crappy ui they have for setting stuff. I also did browse through gconf editor and I did not see even one case where it had a selection box instead of a text entry box for limited inputs. So while it may support that in theory it does not support that in practice.

It seems a lot of useful options where removed from the gui in gnome to make it "easier" and the only real way to change them is using the gconf editor so I don't see how that can even be an optional component.

By Kosh at Fri, 2005/01/07 - 6:00am

Bear in mind that I actually used the first Linux-Mandrake release because it had KDE bundled. As I recall, that was one of the only differences between it and Red Hat at the time.

Now, I'm using Ubuntu Linux for chiefly the same reason I installed that first Mandrake: I get a mostly mainstream distribution with the added benefit of a polished desktop already set up. I've been a big cheerleader for KDE in the past, and even admonished RMS a couple of times for his "ban", rather than working to get the licensing mess straightened out. I used to think of GNOME as a waste of bandwidth and storage, and a pointless project whose reason for existence was over.

Nowadays, they've got their HIG, their initiative to remove the cruft, and their drive to integrate with freedesktop.org standards and software. It's a nightmare under the hood, yeah, but it's sweet-looking and sexy. There's something really nice about being able to do those things that Windows is supposed to make effortless, like plugging in a camera or a USB drive, and just have them work, and it works on my current desktop. App integration is another nice thing. Keeping up with the Joneses in the form of a searchable desktop is another nice thing, and it's getting close to being ready for prime time. There's something nice about having system tools that work more or less the same on all supported systems, and I've got that now.

Now, if only it weren't GNOME...

I mean, once you get past the sexiness, things start getting annoying. Using gconf2-editor to change application settings because adding them to a preference panel was "too confusing"? Annoying. gThumbs insisting on printing at 72dpi, and giving no options to change that? Annoying. Sound-Juicer not having an option for quality settings? Annoying again. In fact, I can point to apps, point to features that were once in apps but have been ripped out because they're too confusing, and I can point to mailing lists where users were patronized for wanting the "confusing" feature back. This is progress? The cleaned-up interfaces are nice, but, c'mon, 72DPI OUTPUT WHEN PRINTING PHOTOS? DUMB, DUMB, DUMB!!! Heck, Ubuntu uses Mozilla Firefox instead of Epiphany as their default Web browser. I think that's pretty telling; I know I prefer the features to the, erm, elegant interface. Um, yeah, you have to open the bookmark editor to move things around on the Bookmarks bar. That makes sense. Uh-huh. Sure. Whatever.

*ahem* What I meant to say was that the simplified interfaces are nice at times, but sometimes they just get in the way.

I look forward to the future of KDE, when they've cleaned up interfaces, dumped cruft like aRts in favor of using gStreamer, and I suspect it will be awesome again by the time GNOME disappears under the mountain of lawsuits that are sure to come when they start using more Mono apps for the base system.

C'mon, guys; KDE used to be a nice, simple, easy-to-use *n?x system. It can be again. Show the GNOME folks how it should REALLY be done. ;-)

By regeya at Wed, 2005/01/05 - 6:00am

As it happens, I have for some time spent time writing a document where I go through the UI of KDE and various apps. I will point out weaknessess and features that make the UI seem more busy than it should be and I will mace suggestions as to how to fix the situation.

Now, I'm not a KDE-developer nor am I some uber-expert when it comes to UI's. I'm just a KDE-user that has used KDE since 1.x-days and who applied some common sense to things :). Hopefully I can spark some discussion when I'm finished with the document.

By Janne at Wed, 2005/01/05 - 6:00am

You surely will spark some discussion.
But this is not enough. If you really want to help, find weak points, suggest improvements and try to get in contact with the developer who works on this code. Work with him (i.e. exchange mails) until you have a result.
If you stop earlier, your work might not have much effect.


By aleXXX at Wed, 2005/01/05 - 6:00am

If the document is like your posts here, please drop it, it would not be very useful to have something saying: "it has to be so because I say it".

Don't get me wrong, there are definitely some things wrong with KDE and UI but, first of all, let's wait final KDE 3.4 to renew the debate, since lots of work has already be done and we are so close to get it finished.

By Davide Ferrari at Wed, 2005/01/05 - 6:00am

"If the document is like your posts here, please drop it, it would not be very useful to have something saying: "it has to be so because I say it"."

It's not. It's about some of the most common things people have criticized KDE for. It outlines the problems, it mentions why they are a problem, it offers potential solutions to the problem and rationale behind those solutions. It has before/after-screenshots and it has some comments from new users.

But hey, apparently it's not useful, so I'll just drop it. Forget I ever mentioned it at all, apparently I have just wasted my time.

By Janne at Wed, 2005/01/05 - 6:00am

I know I've been harsh on kmail/kicker thread, but your way to talk about it made me do that...
I think you have to finish your work on which you had spent time and publish it (the best place is kde-usability mailing list). Even if there is only one good, motivated and well explained criticism (and ideas to improve the situation, obviously), your document will be useful.

By Davide Ferrari at Wed, 2005/01/05 - 6:00am

"I know I've been harsh on kmail/kicker thread, but your way to talk about it made me do that..."

I honestly don't know how I should have talked then. I made no personal attacks. I told why I think the current system is less than ideal, and I told how it could be fixed, and I gave some rationale that explains my proposal. Seriously, I don't know what else I could have done.

But, if I have to choose between voicing my opinions and stirring some emotions in the process, or remain silent, I choose the former. That way I can spark some discussion, and maybe something good will come from it.

"I think you have to finish your work on which you had spent time and publish it (the best place is kde-usability mailing list)"

that is the place where I plan to publish it. We'll see.

"Even if there is only one good, motivated and well explained criticism (and ideas to improve the situation, obviously), your document will be useful."

Well, so far it has stuff regarding Kmenu, Kicker, Konqueror, context-menus and the like, together with screenshots, with more stuff being planned.

By Janne at Wed, 2005/01/05 - 6:00am

I don't think such a work should be dropped. KDE 3.4 may have some enhancements, but most "problems" still exist.
I would rather make first results public (just like open source code). The mailinglist "kde-usability" is a great place to discuss such results with the right people. And after some discussions one can see which kind of "enhancements" are accepted by KDE and work in this direction for further application.

By birdy at Wed, 2005/01/05 - 6:00am

Obviously we can start to talk about some issues before 3.4 but I think that, in primis for technical and time schedule reasons, they won't never be in 3.4

For example, long discussions have been mad on KControl on kde-usability even before 3.3 times, but the KControl will only change in KDE4, when more things could be changed without problems (I mean, there will be a major binary change, a major UI change could happen too)

By Davide Ferrari at Wed, 2005/01/05 - 6:00am

I have used KDE exclusively since KDE 1.1.1, so my opinions on Gnome are mostly based on heresay. But, if KDE is mildly overloaded, I must say that I think Gnome is severly crippled. To me that is much worse. Gnome has dumped usefulness for usability.

By Anonymous at Wed, 2005/01/05 - 6:00am

> GTK/Gnome has great designers UI policy thouhg so it makes up for the horror under the hood.

I always failed to understand why specifying pixel distances in the interface style guide makes up for the horror of the framework not supporting/enforcing them.

By Anonymous at Wed, 2005/01/05 - 6:00am

KDE has a UI policy, just like GNOME does. What's more, it's far more integrated, advanced, and complete, and has a useful set of widgets, including full database support, docking toolbars (ie, it doesn't suffer UI nightmares like GIMP).

I used to be a GNOME fan. Even a KDE hater. But basically, if you think GNOME and KDE are still competing, you're still thinking of GTK 1.x days. They're long gone, and anyone who's really sat down and USED KDE for a while will tell you it's lightyears ahead. At this point, I'm pretty much convinced that KDE has won. Checking out the difference in development progress in the two (not just desktops, but apps for them) will make that fairly clear. When Krita gets here, it'll start to become obvious to all just how much better KDE is.

By Jel at Wed, 2005/01/05 - 6:00am

"KDE has a UI policy, just like GNOME does."

I think the point that was made WRT the UI policies was that GNOME projects will try to adhere to the interface guidelines very strictly. I've read both the interface guides, and have seen that some almost official KDE apps have UI quirks that the interface guide wouldn't want. A prime example was amarok before 1.1 (iirc) where the menu could only be accessed by a button, not by a 'normal' menubar. Little things like that can make a huge difference to some people.

One other thing that's probably the only thing keeeping me on GNOME is the tendency of core KDE components to work almost as they should. Sometimes not having a feature can be nicer than an only semi-functional version of it. An example that shows this well is konqueror's tendancy to hold on to filehandles and fam locks it shouldn't. If I mount a device by going to devices:/ in konqueror and do something on it and hit 'back' to get back to devices:/, I should be able to right-click and unmount it properly. I've not gotten that to work ever. I've voted on various bugzilla entries but still no progress has been made since I first ran into this issue (3.1.?)

I think there's still a place for both the environments on the linux desktop, especially since competition tends to challenge both to do better.

By medgno at Thu, 2005/01/06 - 6:00am

I had the same problem with file handles, it was fixed in 3.3 though. Are you running the latest version? Also, don't install the fam daemon, its useless as far as I can tell and creates a ton of problems, all you need is libfam.

By Leo Spalteholz at Thu, 2005/01/06 - 6:00am

I uninstalled the fam daemon and started a kde session (first time on a new user) and tried mounting my /dev/sda1, creating a file, 'back'ing up and then unmounting it. konqueror complained there was an error unmounting. 'lsof | grep sda1' shows konqueror holding an open filehandle to it. This is on the most up-to-date KDE that Debian has (3.3.1) Is there some special option to put in fstab or somewhere else to tell konqueror to be careful about its handles?

By medgno at Thu, 2005/01/06 - 6:00am

> So so true ...

No, not true:

For example his claim:

"Experts are able to edit text-files or use the command line anyway, novices are just scared away by too many features."

is just complete nonsense.

What experts?

Will experts in spreadsheet-apps be able to edit text-files?
An expert has - by definition - great expertise in a field, but lacks in another.
Or more importantly, will experts be wanting to check out all configuration files? Where will the experts find out what configuration settings are valid anyway?

I can edit text files, but I sure don't want to go hunting for features in them.

Please, don't listen to such rants.

By Roland at Wed, 2005/01/05 - 6:00am

I certainly agree with this. I would qualify as an expert in python and zope but that does not make me an expert in apache, bind or anything else for that matter.

Editing some text files is okay depending on what they are an how easy it is to find them but I hate having to turn on stuff using those text files outside the areas I am an expert it. Trying to find features and turn them on in gconf SUCKED. I like kde since it has just about every option I need and I only have to set those settings once for every major version of kde. I have not had to reconfigure almost any setting since kde 3.0 came out. KDE also supports the way I work by having things be url transparent and having components shared.

I know others may not care but in kde you have essentially one advanced kind of text editor no matter how you open it. You can configure the kate component once and it doesn't matter if you open a file in kdevelop, kate, kwrite, konqueror etc it will be identical and that is a HUGE gain for me. I have become so used to that stuff being shared that I use the colors that code is highlighted in as part of how I read it since it just works everywhere I need it. It is one reason I can't stand using other systems since components are not shared. KDE goes beyond just that though, the address book, proxy settings, spellcheck, etc are all shared.

I can setup once how I want a feature to work and everywhere that feature is used it works the same way. That is a major gain in my mind and I hope kde goes ever farther in that direction.

By Kosh at Wed, 2005/01/05 - 6:00am

Go figure. Gnome's HIG may be the only reason I'll never consider using it. A matter of taste, I guess.

By ac at Wed, 2005/01/05 - 6:00am

If you read the article carefully, while he praises GNOME's simplistic interface over and over again, he still ranks most KDE applications above their GNOME equivalents because they are more featureful. He prefers more features over more simplicity. You can't have both.

By Brandybuck at Wed, 2005/01/05 - 6:00am

GNOME is programmable in C/C++/C#-Mono/Python, etc. The GTK GUI libs will soon be pretty complete and featureful and stable over binary releases. The decision zbout how much GNOME to build into GTK is being made over time (vfs and native versus GTK widgets like file choosers etc.)

After that whatever these very flexible and ever smaller and more efficient libraries can be bound to is whate GNOME will be programmed in ... my guess is that for OO it will be C#/Mono and Java moreso than C++.

By AtWork at Wed, 2005/01/05 - 6:00am

but they also do not make me happy.
Therefore, sticking users in an unhappy way of life prevents them from staying with Gnome.

However, this is not important. I cannot say why, but overall, I find Gnome unusable and I like KDE as is. Everyone has its tastes and smells, so I really hope Gnome and KDE will keep being different. If KDE would evolve towards Gnome, I won't upgrade anymore, so let KDE be as KDE folks want it to be, and use whatever you like.

By christian at Wed, 2005/01/05 - 6:00am

I think that the K menu shown in the screenshots... shows some of the problem.

slides 26-32

slides 52-65

KDE menus have the following:

More Programs >
K3b (CD & DVD Burning)
KMix (Sound Mixer)
KRec (Recording Tool)
KsCD (CD Player)
Kwave Sound Editor
RealPlayer 10 (Media Player)
XMMS (Multimedia Player)

While Gnome menu looks like this:

CD Player
Music Player
Sound Juicer CD Ripper
Sound Recorder
Totem Movie Player
Volume Control

Also Notice that the icons are more "meaningful"
in the sense that some icons represent more the task at hand.
Some icons in KDE are good, some are not obvious.

KDE added the (description) parenthesis in the menu
to add usability, which is some sort of a good idea
to a problem where the name is not obvious.

Now it may be disrupting to "rename" application
to hard core KDE fans just to please other people.

It would be nice to do, for applications that are
"way off track" or for which the name is meaningless.
In marketting, that's a common way of solving an issue
where customer don't buy the product
because the name is irrelevant, not appealing or not selling well.

The major problem I see is alignment and meaningfulness.
When someone reads this menu, the brain spin.
Don't worry the same problem exists on most bloated Windows menu.
The idea is to reduce the information, make it to the point
and make it easy to discriminate the task at end from a list.

The first thing your brain is doing is trying
to focus on too much information, then trying to discriminate
by grouping them and actually trying to read the information.
Since it cannot do that easily, what you end up doing is reading
what is on the left within parenthesis, the second problem
is that doing this task is actually hard since the parenthesis content
is not aligned.

So a first small improvement would be:

More Programs >
GtKpod........\t...(iPod Manager)
K3b...........\t...(CD & DVD Burning)
KMix..........\t...(Sound Mixer)
KRec..........\t...(Recording Tool)
KsCD..........\t...(CD Player)
Kwave.........\t...(Sound Editor)
RealPlayer10..\t...(Media Player)
XMMS..........\t...(Multimedia Player)

At least now, the element are aligned,
but since the content is far from the icon
it's still a bit hard to associate.

Let's try to read what we are looking for first.

More Programs >
iPod Manager........\t...(GtKpod)
CD & DVD Burning....\t...(K3b)
Sound Mixer.........\t...(KMix)
Recording Tool......\t...(KRec)
CD Player...........\t...(KsCD)
Sound Editor........\t...(Kwave)
Media Player........\t...(RealPlayer10)
Multimedia Player...\t...(XMMS)

Now let's sort it by description

More Programs >
CD & DVD Burning....\t...(K3b)
CD Player...........\t...(KsCD)
iPod Manager........\t...(GtKpod)
Media Player........\t...(RealPlayer10)
Multimedia Player...\t...(XMMS)
Recording Tool......\t...(KRec)
Sound Editor........\t...(Kwave)
Sound Mixer.........\t...(KMix)

Now, it starts looking like a Gnome/Apple menu.

Actually, there's nothing wrong with putting the names
instead of the description, if the names are meaningful,
such as this:

More Programs >
KCD Burning..........\t...(K3b)
KCD Player...........\t...(KsCD)
KiPod Manager........\t...(GtKpod)
KMultimedia Player...\t...(XMMS)
KRecording Tool......\t...(KRec)
KSound Editor........\t...(Kwave)
KSound Mixer.........\t...(KMix)

If everything starts with a K, then the sorting actually makes sense,
or just rip it off and assume that in console it takes a K prefix...

More Programs >
CD Burning..........
CD Player...........
iPod Manager........
Multimedia Player...
RealPlayer 10
Recording Tool......
Sound Editor........
Sound Mixer.........

Some grouping might also help:

More Programs >
CD & DVD Burning....\t...(K3b)
CD Player...........\t...(KsCD)
iPod Manager........\t...(GtKpod)
Recording Tool......\t...(KRec)
Sound Editor........\t...(Kwave)
Sound Mixer.........\t...(KMix)
Media Player........\t...(RealPlayer10)
Multimedia Player...\t...(XMMS)

* Mock-up anyone?

The best would be to omit (parenthesis)
or have names that are so meaningful that
it's just so easy to remember.

KEdit, KWrite, KPaint, KolourPaint, DigiKam, Kuickshow, KGhostView, KPDF

The major problem is that there is so many
competing app, that it may not be fully possible.

Also, don't make the menu too big and make it
visually and logically grouped by task whenever possible.

If the menu is too big, add submenu, but use proper folder names
not "More Programs", "All Programs", that's meaningfulness.
Basically, it doesn't provide any plus-value information.

Also, make the menu editable on the fly by drag-n-drop.

Hide menu bar features into an advance submenu
with a proper grouping and organize it
in an easy to find 3 clicks thing with
a keyboard shortcut.

If the user use it "often", the user can simply
drag'n'drop the menu item and edit it in place
[if you really want to provide editable menu]
or learn/redefine the keyboard shortcut ALT-CTRL-SHIFT-K (even better).

One thing, you might wanna do is add a "feature counter plugin" KPART
into some KDE apps and ask users to provide you with statistics
on frequently used features by sending their results
via some HTTP website when connected in XML format
by clicking "send KDE usability results".

That might help developpers "justify" the cut
the visibility of the kitchen sink,
by having proper statistic to support it.

Also, having icon contest with meaningful icons is a good idea.
Getting inspired from other graphical environment is also a good idea.

Long life to KDE!

Sincerely yours,

By fprog26 at Wed, 2005/01/05 - 6:00am

> I think that the K menu shown in the screenshots... shows some of the problem.
> GtKpod
> RealPlayer 10 (Media Player)
> xine
> XMMS (Multimedia Player)

A really good idea to blame KDE for non-KDE applications added by the distributor. ;-)

> KDE added the (description) parenthesis in the menu to add usability

KDE's default though is to display the description and in parenthesis the app name.

By Anonymous at Wed, 2005/01/05 - 6:00am

Hope somebody in the KDE inner circle took notice to Freds well articulated point

Perhaps instead of comparing KDE to what others do wrong, KDE designers should look at Apples UI, as I think most would agree that have set the standard for polish and ease of use.

Also it is time for KDE to get out of its hacker mode. And get companies like Novell and IBM to participate more.

KDE is great it just needs a little direction (and a name change)


By Barney at Wed, 2005/01/05 - 6:00am

> as I think most would agree that have set the standard for polish and ease of use

Really? I want to bet that "most" have never used an Apple computer.

By Anonymous at Wed, 2005/01/05 - 6:00am


Or have tried to and said "Okay... Now how on earth do I do task X? I don't see an obvious way of doing it." At which point their Mac tutor points out the way, or something similar with the statement that it is "More intuitive this way!".

One day, someone is going to tell me how dragging a disk icon to the trash bin to eject it is in the least bit intuitive. For a new user, I'd be deathly worried that all my work was going to be deleted, after spending half an hour wondering where the 'Eject' option was.

Anecdote: A very intelligent man came to my shared house a year ago to just check his webmail on my friend's OSX Mac while he was away from home. I head off to work and return for lunch three hours later. Turns out, he's playing games, because after two hours, he couldn't find the web browser. You see my friend had Mozilla installed, and the big cuddly 'M' on the dock, amidst 25 other icons didn't exactly shout "Hey! I'm a web browser!" at this chap. Poor choice of icon on the part of the Mozilla Foundation, perhaps, but then, OSX didn't provide any tooltip or program menu structure to guide him to the fact that it was a browser, either.

And similarly to Gnome, I've spent a great deal of time in the last two years trying to change settings on clients' Macs for networking and behaviour that Apple decided were too intricate for normal users and text files had to be hacked every which way to solve the problem.

There is a danger in discussions of usability in modern times that too many people assume that Apple has 'got it right'. If they had usability perfectly solved, working on a Mac would be so much more comfortable and improve work flow, that every company would be compelled to Switch in order to save the man hours in efficiency gains and reduced administration time. That hasn't happened, and even desktop market share isn't growing beyond the current 3.5-5% that Macs currently occupy. If Apple were really so good that we should pay attention to everything that they said and did, they'd have the 80%+ market share that Windows has, right now.

By Luke Chatburn at Wed, 2005/01/05 - 6:00am

Well, apart that he said wrong things (the default of clean KDE 3.3 installation in KMenu is "App description (App name)"), apart that he should give a look to the new KMenu and kicker of 3.4 (kudos to Aaron Seigo)...well, he could, yes. (if you don't get it this last sentence is ironic)

By Davide Ferrari at Wed, 2005/01/05 - 6:00am

Personally, I don't have a problem with app names. We are not a corporation and have no reason to emulate one. Having developers choose whatever names they want for their apps highlights the fact that this is software written by real, actual people, who are doing it because they like doing it; lends a sort of intimacy to it. Having a description next to the names solves the usability issue nicely, and lets us avoid those terrible corporatesque names like Internet Explorer.

By Gábor Lehel at Wed, 2005/01/05 - 6:00am

He doesn't even know how to write GNOME correctly.

By Anonymous at Wed, 2005/01/05 - 6:00am

I like tho control center although i don't like the way searching is done.
When i type a search word then there apear the topics, then i have to click on the topic further down to get the help page. The indirection in the middle could IMHO be removed.

Another cool thing could be the configuration dialogs based on skill level like xine does it. Although it requires a lot of work to get it right an imposes new problems. For example missing options.

One thing which is really crowded are the kde context menu's for drives. The most used item "unmount" lies inbetween not so much used functions. Especially for newbies who are no familiar with mounting /UNMOUNTING this makes it even harder.

By Tim at Wed, 2005/01/05 - 6:00am

>configuration dialogs based on skill level
This is a developers nightmare, on many levels. First of it adds lot of complications to testing, and getting usefull bugreports would become even worse. And the problems it creates for user support are huge. The Gnomes used to have skill setting in Nautilus, I don't think they do anymore.

By Morty at Wed, 2005/01/05 - 6:00am

i think the main usability criteria should be possibility to do all things only using keyboard [shortcuts].

for example, i've hidden all toolbars and menu in kmail and opera

By nick at Wed, 2005/01/05 - 6:00am

No, this is about power and flexibility of a DE, but not about "usability" in general.
Anyway, it's a great KDE feature, you can do almost everything by keyboard shortcuts

By Davide Ferrari at Wed, 2005/01/05 - 6:00am

I know that there was even some discussion on wether a 3.4 release should happen or not. But why not have a 3.5 release that focus on UI improvements only? It would not distract developers from 4.0 that much and it would allow them to get feedback on those changes and see how everything worked before committing them to 4.0.

By John Usability Freak at Wed, 2005/01/05 - 6:00am

Maybe I'm missing something, but why not put the big UI/useability changes in 4.0.. After all, that's what major releases are for, no?

By Mikhail Capone at Thu, 2005/01/06 - 6:00am

"Maybe I'm missing something, but why not put the big UI/useability changes in 4.0."

So many motives.
Remember 2.2->3.0->3.1->3.2->3.3? When developing 4.0 most of the focus be on technical issues, porting to a new Qt version, new sound server, and some new features, etc. And right after the release it will be much more important to fix bugs on the new foundations, and after that building new features that take advantage of the new foundations. Also the feedback on non-UI issues will be overwhelming. In the event of a 3.5, developers would have a stable framework to test UI changes (3.4), and feedback on that release would concern UI changes alone. Developers could then take that feedback into account as they work on 4.0. That's the only way I can see it working.

"After all, that's what major releases are for, no?"
3.1, 3.2, 3.3, 3.4 are all major releases.

By John Usability Freak at Thu, 2005/01/06 - 6:00am

I think KDE is fantastic. It looks great, is packed with features and has great technology behind it. However, I would really like it if the default settings were cleaned up as I am sick of all the Gnome zealots going on and on about how "unusable" it is because of things like cluttered toolbars. Whenever I install KDE on a new machine, I always spend 20 minutes going through all the options and toolbar configurations to make it less cluttered. It isn't terrible, I just hate it when people insult the effort that has went into KDE just because of some extra icons or something else that can be changed in 30 seconds.

Anyway, I've attached a screenshot of my regular Konqueror window as an example of the kind of cleanups I'd like to see. As you can see, my toolbar only contains: Back, Forward, Home, Up, Stop, Address Bar, Go/Enter. Regular users won't need anything else (I'm a programmer and that's all I need too) and advanced users can add more items if they want them. Very importantly, all buttons are on one horizontal toolbar to take up minimum space. My web fonts are set quite large and I have big friendly icons for file management. I use the same profile for both file management and web browsing, with my homepage set to my home directory. I've never actually used the sidebar except when exploring it, so I turn it off by default.

I think this setup looks very appealing, except perhaps for the status bar at the bottom. It has too many ugly looking frames, the little green circle serves no real purpose and I have never needed the "lock view" button".

Just in general though, the toolbars are very overloaded with things. This would take a minimum amount of time to fix and make things look a lot less cluttered. For example, in Kontact there is a "Print" icon in the toolbar for emails, address book contacts and news groups which I have never ever used.

I personally find Gnome very spartan of useful features and personally don't like the look of it (the default fonts and theme aren't attractive). However, I would very much appreciate if some of the visual complaints that Gnome users keep applying to KDE were addressed so that they would only be compared on features, and KDE is far superior here.

By ryukendo000 at Wed, 2005/01/05 - 6:00am

I see you have removed the "Clear Location Bar", I can't actually understand how you mangage without? It's the thing I miss the most each time I try Firefox, and the reason I stay with Konqueror as browser. Probably combined with the fact that there is no other gain using Firefox over Konqi.

I have 16 icons in my Konqueror toolbar, and regularly use 8-9 of them. I could spend the 30 seconds removing a few of them, but Konqueror would not become any more usable for me so why should I bother? Of course if the icons was not there to begin with, I would not add them. But it would not become more usable either.

By Morty at Wed, 2005/01/05 - 6:00am

I removed "Clear Location Bar" because I find it easier to 1) double-click on the address to select all the text 2) type in an address (which automatically clears the previous one). The clear button isn't used everywhere so I'm just not drawn to using it, whereas selecting text works with all text-entry boxes.

Hmm, 16 icons on your toolbar is far, far too many in my opinion. What are they all? I think the KDE designers should remove all the toolbar icons that are very rarely used by most users. The whole point of the toolbar is to allow certain features (like "Back") to be accessed with just one click. When you have too many icons, it not only looks cluttered, but you now have to scan along all the icons to find the one you want because there are so many. Having 20 odd icons looks very daunting also as it makes the application look far more complicated than it actually is.

Like I said in my first post though, I know I can change the toolbars myself and I do, but I just wish the default settings were a lot cleaner looking so that people couldn't use it as a lame reason not to use KDE.

By ryukendo000 at Wed, 2005/01/05 - 6:00am